Joseph M. Juran Co-Editor-in-Chief
A. Blanton Godfrey Co-Editor-in-Chief
Robert E. Hoogstoel Associate Editor
Edward G. Schilling Associate Editor
Fifth Edition
JOSEPH M. JURAN, Co-Editor-in-Chief, has published the leading reference and training materials
on managing for quality, a field in which he has been the international leader for over 70 years. A
holder of degrees in engineering and law, Dr. Juran has pursued a varied career in management as
an engineer, industrial executive, government administrator, university professor, corporate director,
and management consultant. As a member of the Board of Overseers, he helped to create the
U.S. Malcolm Baldrige National Quality Award and has received over 50 medals and awards from
14 countries, including The Order of the Sacred Treasure from the Emperor of Japan for “. . . the
development of Quality Control in Japan and the facilitation of U.S. and Japanese friendship”;
and the National Medal of Technology from the President of the United States for “his lifetime
work of providing the key principles and methods by which enterprises manage the quality
of their products and processes, enhancing their ability to compete in the global marketplace.” He
is also founder of the consulting firm of Juran Institute, Inc., and founder of Juran Foundation, Inc.
The latter is now a part of the Juran Center for Leadership in Quality at the Carlson School of
Management in the University of Minnesota. Among his 20 books, the Handbook is the international
reference standard.
A. BLANTON GODFREY, PH.D., Co-Editor-in-Chief, is Chairman and Chief Executive Officer of
Juran Institute, Inc.. Under his leadership Juran Institute has expanded its consulting, training, and
research services to over 50 countries. Previously, Dr. Godfrey was head of the Quality Theory and
Technology Department of AT&T Bell Laboratories. He is a Fellow of the American Statistical
Association, a Fellow of the American Society for Quality, a Fellow of the World Academy of
Productivity Science, and an Academician of the International Academy for Quality. He is a coauthor
of Modern Methods for Quality Control and Improvement and Curing Health Care. He contributed
to the creation of the Malcolm Baldrige National Quality Award and served as a judge
from 1988 to 1990. In 1992 he received the Edwards Medal from ASQ for his outstanding contributions
to the science and practice of quality management. Dr. Godfrey holds an M.S. and a Ph.D.
in Statistics from Florida State University and a B.S. in Physics from Virginia Tech.
ROBERT E. HOOGSTOEL is Associate Editor for all but the statistical sections of the Handbook.
As Manager of Technical Education at Dana Corporation, he conducted training and consulting in
quality management throughout the corporation. For ten years he was a consultant with Juran
Institute, working with clients worldwide in the manufacturing and process industries. He coauthored
with Dr. Juran the Institute’s training course in quality planning, and appears in the
Institute’s video training series, “Juran on Quality Planning.” He has engineering degrees from
Cornell University and the University of California at Berkeley. He is a registered professional
engineer in New York State.
EDWARD G. SCHILLING, PH.D., is Associate Editor for the statistical sections of the Handbook.
He is Professor of Statistics in the Center for Quality and Applied Statistics at Rochester Institute
of Technology, where he has held the position of Director of the Center and Chair of the Graduate
Statistics Department. Prior to joining R.I.T. he was manager of the Lighting Quality Operation
for the Lighting Business Group of the General Electric Company. He received his B.A. and
M.B.A. degrees from SUNY Buffalo, and his M.S. and Ph.D. degrees in statistics from Rutgers
University, New Jersey. Dr. Schilling is a Fellow of the American Statistical Association and the
American Society for Quality, and a member of the Institute of Mathematical Statistics, the
American Economic Association, and the American Society for Testing and Materials. He received
the ASQ Shewhart Medal in 1983 and its Brumbaugh Award four times. He is registered as a professional
engineer in California and is certified by ASQ as a quality and reliability engineer. His
two books, Acceptance Sampling in Quality Control and Process Quality Control (with E. R. Ott),
are among the leading texts in the field.
Contributors ix
Preface to the Fifth Edition xiii
1. How to Use the Handbook J. M. Juran 1.1
2. How to Think about Quality J. M. Juran 2.1
3. The Quality Planning Process John F. Early and O. John Coletti 3.1
4. The Quality Control Process J. M. Juran and A. Blanton Godfrey 4.1
5. The Quality Improvement Process J. M. Juran 5.1
6. Process Management James F. Riley, Jr. 6.1
7. Quality and Income J. M. Juran 7.1
8. Quality and Costs Frank M. Gryna 8.1
9. Measurement, Information, and Decision-Making Thomas C. Redman 9.1
10. Computer Applications to Quality Systems Fredric I. Orkin and
Daniel P. Olivier 10.1
11. The ISO 9000 Family of International Standards Donald W. Marquardt 11.1
12. Benchmarking Robert C. Camp and Irving J. DeToro 12.1
13. Strategic Deployment Joseph A. DeFeo 13.1
14. Total Quality Management A. Blanton Godfrey 14.1
15. Human Resources and Quality W. R. Garwood and Gary L. Hallen 15.1
16. Training for Quality Gabriel A. Pall and Peter J. Robustelli 16.1
17. Project Management and Product Development Gerard T. Paul 17.1
18. Market Research and Marketing Frank M. Gryna 18.1
19. Quality in Research and Development Al C. Endres 19.1
20. Software Development Lawrence Bernstein and C. M. Yuhas 20.1
21. Supplier Relations John A. Donovan and Frank P. Maresca 21.1
22. Operations Frank M. Gryna 22.1
23. Inspection and Test E. F. “Bud” Gookins 23.1
24. Job Shop Industries Leonard A. Seder 24.1
25. Customer Service Edward Fuchs 25.1
26. Administrative and Support Operations James A. F. Stoner,
Charles B. Wankel, and Frank M. Werner 26.1
27. Process Industries Donald W. Marquardt 27.1
28. Quality in a High Tech Industry M. K. Detrano 28.1
29. Automotive Industry Yoshio Ishizaka 29.1
30. Travel and Hospitality Industries Patrick Mene 30.1
31. Government Services Al Gore 31.1
32. Health Care Services Donald M. Berwick and Maureen Bisognano 32.1
33. Financial Services Industries Charles A. Aubrey II and Robert E. Hoogstoel 33.1
34. Second-Generation Data Quality Systems Thomas C. Redman 34.1
35. Quality and Society J. M. Juran 35.1
36. Quality and the National Culture J. M. Juran 36.1
37. Quality in Developing Countries Lennart Sandholm 37.1
38. Quality in Western Europe Tito Conti 38.1
39. Quality in Central and Eastern Europe Robert E. Hoogstoel 39.1
40. Quality in the United States J. M. Juran 40.1
41. Quality in Japan Yoshio Kondo and Noriaki Kano 41.1
42. Quality in the People’s Republic of China Yuanzhang Liu 42.1
43. Quality in Latin America Marcos E. J. Bertin 43.1
44. Basic Statistical Methods Edward J. Dudewicz 44.1
45. Statistical Process Control Harrison M. Wadsworth 45.1
46. Acceptance Sampling Edward G. Schilling 46.1
47. Design and Analysis of Experiments J. Stuart Hunter 47.1
48. Reliability Concepts and Data Analysis William Q. Meeker, Luis A. Escobar,
Necip Doganaksoy, and Gerald J. Hahn 48.1
Appendix I. Glossary of Symbols and Abbreviations AI.1
Appendix II. Tables and Charts AII.1
Appendix III. Selected Quality Standards, Specifications,
and Related Documents AIII.1
Appendix IV. Quality Systems Terminology AIV.1
Appendix V. Quality Improvement Tools AV.1
Name Index and Subject Index follow Appendix V
Charles A. Aubrey II Vice President, American Express, New York, NY (SECTION 33, FINANCIAL SERVICES
Lawrence Bernstein Vice President, Operations Systems, AT&T Network Systems (retired); Have Laptop-
Marcos E. J. Bertin Chairman of the Board, Firmenich S. A., Buenos Aires, Argentina (SECTION 43, QUALITY
Donald M. Berwick, M.D. President and Chief Executive Officer, Institute for Healthcare Improvement,
Maureen Bisognano Executive Vice President and Chief Operating Officer, Institute for Healthcare
Improvement, Boston, MA (SECTION 32, HEALTH CARE SERVICES)
Robert C. Camp, Ph.D., PE Best Practice Institute™, Rochester, NY (SECTION 12, BENCHMARKING)
O. John Coletti Manager, Special Vehicle Engineering, Ford Motor Company, Danou Technical Center, Allen
Dr. Tito Conti President, OAM s.r.l., Ivrea, Italy (SECTION 38, QUALITY IN WESTERN EUROPE)
Joseph A. DeFeo Executive Vice President and Chief Operating Officer, Juran Institute, Inc., Wilton, CT
Irving J. DeToro The Quality Network, Palm Harbor, FL (SECTION 12, BENCHMARKING)
Ms. M. K. Detrano Quality Director, Network Systems Product Realization, Lucent Technologies (retired);
Detrano Consulting, Somerville, NJ (SECTION 28, QUALITY IN A HIGH TECH INDUSTRY)
Dr. Necip Doganaksoy Research and Development, General Electric Corporate, Schenectady, NY (SECTION
John A. Donovan Director PARIS Associates, Fairfax, VA (SECTION 21, SUPPLIER RELATIONS)
Dr. Edward J. Dudewicz Professor and Consultant, Department of Mathematics, Syracuse University,
John F. Early Vice President, Planning and Strategic Support, Empire Blue Cross and Blue Shield, New York,
Dr. Al C. Endres Director, Center for Quality, and Chair, Dept. of Management, University of Tampa, Tampa,
Professor Luis A. Escobar Dept. of Experimental Statistics, Louisiana State University, Baton Rouge, LA
Edward Fuchs Chapel Hill, NC (SECTION 25, CUSTOMER SERVICE)
William R. Garwood President, Eastman Chemical Co., Europe, Middle East, and Africa Region (retired),
Dr. A. Blanton Godfrey Chairman and Chief Executive Officer, Juran Institute, Inc.,Wilton, CT (CO-EDITORIN-
E. F. “Bud” Gookins Managing Partner, Strategic Quality Consulting, Elyria, OH (SECTION 23, INSPECTION
Vice President Al Gore Washington, DC (SECTION 31, GOVERNMENT SERVICES)
Dr. Frank M. Gryna Distinguished Univ. Professor of Management, The Center for Quality, College of
Business, The University of Tampa, Tampa, FL (SECTION 8, QUALITY AND COSTS; SECTION 18, MARKET RESEARCH
Dr. Gerald J. Hahn Research and Development, General Electric Corporate, Schenectady, NY (SECTION 48,
Gary L. Hallen Eastman Chemical Company, Kingsport, TN (SECTION 15, HUMAN RESOURCES AND QUALITY)
Robert E. Hoogstoel Vice President, Juran Institute (retired), Pittsboro, NC (ASSOCIATE EDITOR; SECTION 33,
Dr. J. Stuart Hunter Professor Emeritus, Princeton University, Princeton, NJ (SECTION 47, DESIGN AND
Yoshio Ishizaka President and Chief Executive Officer, Toyota Motor Sales USA, Torrance, CA (SECTION 29,
Dr. Joseph M. Juran Founder, Juran Institute, Wilton, CT (CO-EDITOR-IN-CHIEF; SECTION 1, HOW TO USE THE
Dr. Noriaki Kano Professor, Dept. of Management Science, Faculty of Engineering, Science University of
Tokyo, Tokyo, Japan (SECTION 41, QUALITY IN JAPAN)
Professor Yoshio Kondo Professor Emeritus, Kyoto University, Kyoto, Japan (SECTION 41, QUALITY IN
Professor Yuanzhang Liu Institute of Systems Science, Academia Sinica, Beijing, China (SECTION 42, QUALITY
Frank P. Maresca Manager Quality Improvement, Mobil Oil Corporation (retired), Fairfax, VA (SECTION 21,
Donald W. Marquardt Donald W. Marquardt and Associates, Wilmington, DE (SECTION 11, THE ISO 9000
days after submitting his final manuscripts for this handbook. He was the victim of a heart attack. The editors are
grateful for his contribution to this book. Consistent with his performance in a long career in quality, Don’s submissions
were crisp, complete, and timely.
Dr. William Q. Meeker Professor of Statistics, and Distinguished Professor of Liberal Arts and Sciences,
Department of Statistics, Iowa State University, Ames, IA (SECTION 48, RELIABILITY CONCEPTS AND DATA ANALYSIS)
Patrick Mene Vice President, Quality, The Ritz Carlton Hotel Company, Atlanta, GA (SECTION 30, TRAVEL AND
August Mundel Professional Engineer, White Plains, NY (APPENDIX III)
Daniel P. Olivier President, Certified Software Solutions, San Diego, CA (SECTION 10, COMPUTER APPLICATIONS
Fredric I. Orkin President, Frederic I. Orkin and Associates, Inc., Highland Park, IL (SECTION 10, COMPUTER
Gabriel A. Pall Senior Vice President, Juran Institute, Wilton, CT (SECTION 16, TRAINING FOR QUALITY)
Dr. Thomas C. Redman President, Navesink Consulting Group, Rumson, NJ (SECTION 9, MEASUREMENT,
James F. Riley, Jr. Vice President Emeritus, Juran Institute, Inc., University Park, FL (SECTION 6, PROCESS
Peter J. Robustelli Senior Vice President, Juran Institute, Inc., Wilton, CT (SECTION 16, TRAINING FOR QUALITY)
Professor Dr. Lennart Sandholm President, Sandholm Associates AB, Djursholm, Sweden (SECTION 37,
Dr. Edward G. Schilling Professor of Statistics, Rochester Institute of Technology, Center for Quality and
Leonard A. Seder Quality Consultant (retired), Lexington, MA (SECTION 24, JOB SHOP INDUSTRIES)
Professor James A. F. Stoner Graduate School of Business Administration, Fordham University, New York,
Dr. Harrison M. Wadsworth Professor Emeritus, School of Industrial and Systems Engineering, Georgia
Institute of Technology, Atlanta, GA (SECTION 45, STATISTICAL PROCESS CONTROL)
Professor Charles B. Wankel College of Business, St. John’s University, Jamaica, NY (SECTION 26, ADMINISTRATIVE
Professor Frank M. Werner Associate Dean, Graduate School of Business Administration, Fordham
Ms. Josette H. Williams Juran Institute, Inc., Wilton, CT (APPENDIX V)
Ms. C. M. Yuhas Have Laptop-Will Travel, Short Hills, NJ (SECTION 20, SOFTWARE DEVELOPMENT)

In the preface to the Fourth Edition of this handbook, Dr. Juran commented on the events of the four
decades between signing the contract for the First Edition of this handbook (1945) and the publication
of the Fourth Edition (1988). He noted the growth of the handbook itself—in circulation and in
status—and the parallel growth of importance of quality in society generally. The growth was attributable
to the increasing complexity of products and the systems in which they participate, and,
because of our increasing dependence on these systems, to the unprecedented potential for disruption
when these products fail. This threat (and its occasional frightening fulfillment) is what he long
ago identified as “life behind the quality dikes.”
In the decade that has passed since the Fourth Edition, the importance of quality has continued to
grow rapidly. To some extent, that growth is due in part to the continuing growth in complexity of
products and systems, society’s growing dependence on them, and, thus, society’s growing dependence
on those “quality dikes.” But the main impetus for the growing importance of quality in the
past decade has been the realization of the critical role quality plays as the key to competitive success
in the increasingly globalized business environment. Upper managers now understand much
more clearly the importance of quality—convinced by the threat of the consequences of product failure,
by the rapid shift of power to the buyers and by the demands of global competition in costs,
performance, and service.
As the importance of achieving quality has sunk in, the quest to learn how to achieve it has grown
also. The emergence in the United States of America of the Malcolm Baldrige National Quality
Award, and its many offspring at the state level, have promoted the development of quality by providing
a comprehensive, home-grown organizational model for the achievement of quality, and by
opening to view organizations that have applied this model successfully. It is difficult to overstate the
importance of these models of excellence in the promotion of quality practice over the past decade.
They have provided managers at all levels with evidence that “it can be done here,” and, more important,
they have provided in unusual detail, roadmaps of how it was done. In Europe, the European
Quality Award and its offspring have provided much the same motive power to the quality movement
that the Baldrige Award has provided in the United States.
The mounting success of quality in the industrial sector has caused recognition of the importance
of quality to spread throughout manufacturing industries, the traditional home ground of quality
ideas and applications, and beyond to the service sector, government, and non-profit enterprises. In
this regard, we are especially pleased to welcome the contribution on quality in government of Vice
President of the United States Al Gore.
In recognition of these changes, the editors have made some fundamental changes in this handbook.
1. We have changed the name from Juran’s Quality Control Handbook, to Juran’s Quality
Handbook. The new name signals the change in emphasis from quality control, traditionally the
concern of those working on the manufacturing floor, to an emphasis on the management
of quality generally, a concern of managers throughout an organization.
2. We have changed the structure to reflect the new emphasis on managing quality. The Fifth
Edition has 48 sections, arranged in five groups: Managerial, Functional, Industry,
International, and Statistical.

The revision has not consisted merely of rearrangement. Once again, as in the Fourth Edition, the
content of this edition has has undergone extensive editing and updating. There are many entirely
new sections on new subjects. There are total rewrites of other sections. And there are many new
additions of case studies, examples and other material even to the few “classic sections.” An editorial
undertaking of this scope and magnitude would be unthinkable without the help and support of a
number of our colleagues and friends.
The founding editor of the handbook, Joseph M. Juran, has placed his unmistakable stamp of
vision and clarity on this new edition—the fifth in which he has played a guiding role—by his contributions
to its planning and, more directly, in the six major sections that he authored. My association
with him since I joined Juran Institute in 1987 has provided a deep and rewarding exploration
of the evolving field of quality management. Sharing the position of Editor-in-Chief of the present
volume has been a part of that experience.
Our Associate Editors, Edward Schilling and Robert Hoogstoel, shared the major literary and
diplomatic burden of helping the contributors create handbook sections that would at once reveal
their individual subject-matter expertise and would mesh smoothly with the other sections to make
a coherent and useful desk reference, in the long tradition of this book. Ed Schilling edited
Sections 44 through 48, those concerned with mathematical statistics and related applications;
Bob Hoogstoel edited most of the remaining sections and provided overall coordination of the
editorial effort.
The grounding in practical experience which has characterized earlier editions of this book is
strengthened further in this edition by the examples provided by the numerous managers who have
shared their experiences on the quality journey through their presentations at Juran Institute’s annual
IMPRO conferences, workshops and seminars. We also wish to acknowledge the generous support of
Juran Institute, Inc. throughout this endeavor. Many of the figures and charts come straight from Juran
Institute publications and files, many others were created with support from people and facilities within
the Institute.
Among the many colleagues at Juran Institute who have made major exertions on behalf of this
book, Josette Williams stands out. Her own editorial and publishing experience have sharpened her
sense of what goes and what doesn’t, a sense she shared willingly. Jo provided a comforting presence
as she managed the flow of correspondence with the contributors, and helped the editors enormously
by performing calmly and expertly as liaison with the publisher astride the flow of manuscripts, the
counterflow of page proofs, and the publisher’s myriad last-minute questions of detail and the manuscript
tweakings by contributors. Jo went far beyond the usual bounds of the responsibilities of an
assistant editor. She worked closely with authors, editors, the publisher, and others in making this
edition happen. Her style and grasp of language and clarity of expression are present in almost every
section. This handbook owes much to her dedication, focus, and thousands of hours of hard work.
Fran Milberg played a major role in preparing the manuscript for submission. My Executive Assistant,
Jenny Edwards, frequently found her considerable workload in that job added to by the sudden, often
unpredictable demands associated with the preparation of the manuscript, answering authors’ questions,
and keeping me on track. It was too much to ask of a normal person, but Jenny, as always, rose
to the occasion, for which I am most grateful. Many others among the Juran Institute support staff
helped at various stages of manuscript preparation, including: Laura Sutherland, Jane Gallagher,
Marilyn Maher, and Carole Wesolowski. In the early stages of organizing for this effort we were grateful
for the assistance of Sharon Davis and Rosalie Kaye. Special thanks go to Hank Williams who
spent hours at the copier and many other hours helping Josette make sure manuscripts were sent on
time to all the right places.
It would be unfair (and unwise) to omit mention of those closest to the contributors and editors
of this book, the wives and husbands whose personal plans had occasionally to be put on hold in
favor of work on the book. Larry Bernstein and C.M.Yuhas sidestepped the problem by making
Section 20, Software Development, a family project, as is their joint consultancy. Other contributors
no doubt were faced with dealing with the inevitable impingement on family life in their own
ways. As for the editors, we unite to thank our wives for their support in this endeavor: Dr. Juran’s
wife of 73 years, known to him as “Babs,” and to the rest of us as a gracious inspiration and
Editorial Assistant Emerita of the first three editions of this book and numerous of his earlier
books and papers; Judy Godfrey, now a survivor of three books; Jean Schilling, a veteran editor of
her husband’s earlier publications and who has been patient and supportive in this effort; and Jewel
Hoogstoel, for whom the answer to her persistent question is “It is done.”We hope they will share
the editors’ mutual sense of accomplishment.

J. M. Juran
The Managerial Group 1.2
The Functional Group 1.2
The Industry Group 1.2
The International Group 1.2
The Statistical Group 1.3
Table of Contents 1.3
Use of the Index 1.3
Cross-References 1.4
Citations 1.4
Special Bibliographies 1.4
Literature Search 1.4
The Internet 1.4
Author Contact 1.5
Other Sources 1.5
This is a reference book for all who are involved with quality of products, services, and processes.
Experience with the first four editions has shown that “all who are involved” include:
 The various industries that make up the international economy: manufacture, construction, services
of all kinds—transportation, communication, utilities, financial, health care, hospitality, government,
and so on.
 The various functions engaged in producing products (goods and services) such as research and
development, market research, finance, operations, marketing and sales, human resources, supplier
relations, customer service, and the administration and support activities.
 The various levels in the hierarchy—from the chief executives to the work force. It is a mistake to
assume that the sole purpose of the book is to serve the needs of quality managers and quality specialists.
The purpose of the book is to serve the entire quality function, and this includes participation
from every department of the organization and people in all levels of the organization.
 The various staff specialists associated with the processes for planning, controlling, and improving
While there is a great deal of know-how in this book, it takes skill and a bit of determination to
learn how to find and make use of it. This first section of the handbook has therefore been designed
to help the reader to find and apply those contents which relate to the problem at hand.
The handbook is also an aid to certain “stakeholders” who, though not directly involved in producing
and marketing products and services, nevertheless have “a need to know” about the qualities
produced and the associated side effects. These stakeholders include the customer chain, the public,
the owners, the media, and government regulators.
Practitioners make a wide variety of uses of Juran’s Quality Handbook. Experience has shown that
usage is dominated by the following principal motives:
 To study the narrative material as an aid to solving problems
 To find structured answers in tables, charts, formulas, and so on
 To review for specific self-training
 To find special tools or methods needed for reliability engineering, design of experiments, or statistical
quality control
 To secure material for the teaching or training of others
 To use as the basic reference in quality training, quality management operations, and even in
designing and managing fundamental work functions
Beyond these most frequent uses, there is a longer list of less frequent uses such as:
 To review for personal briefing prior to attending a meeting
 To cross-check one’s approach to tackling a problem
 As a reference for instructors and students during training courses
 To indoctrinate the boss
 To train new employees
 To help sell ideas to others, based on: (1) the information in the handbook and (2) the authoritative
status of the handbook
Usage appears to be more frequent during times of change, for example, while developing new
initiatives, working on new contracts and projects, reassigning functions, or trying out new ideas.
Irrespective of intended use, the information is very likely available. The problem for the practitioner
becomes one of: (1) knowing where to find it and (2) adapting the information to his or her
specific needs.
Knowing “where to find it” starts with understanding how the handbook is structured. The handbook
consists of several broad groupings, outlined as follows.
The Managerial Group (Sections 2 through 17). This group deals with basic concepts
and with the processes by which quality is managed (planning, control, and improvement) along with
sections of a coordinating nature such as human resources and project management.
The Functional Group (Sections 18 through 26). This group roughly follows the
sequence of activities through which product concepts are converted into marketable goods and services.
The Industry Group (Sections 27 through 33). This group illustrates how quality is
attained and maintained in selected leading industries.
The International Group (Sections 35 through 43). This group deals with how quality
is managed in selected geographical areas throughout the world.
Collectively the previous four groups comprise about 70 percent of the handbook.
The Statistical Group (Sections 44 through 48). This group shows how to use the principal
statistical tools as aids to managing for quality. This group also includes the glossary of symbols
(Appendix I) as well as the supplemental tables and charts (Appendix II).
Collectively the statistical group comprises about 25 percent of the handbook.
There are three main roads for locating information in the handbook:
1. Table of contents
2. Index
3. Cross-references
In addition, there are supplemental devices to aid in securing elaboration. Note that the handbook
follows a system of dual numbering, consisting of the section number followed by page number, figure
number, or table number. For example, page number 16.7 is the seventh page in Section 16.
Figure or table number 12.4 is in Section 12, and is the fourth figure or table in that section.
Table of Contents. There are several tables of contents. At the beginning of the book is the list
of section and subsection headings, each of which describes, in the broadest terms, the contents of
that section.
Next there is the list of contents that appears on the first page of each section. Each item in any
section’s list of contents becomes a major heading within that section.
Next, under each of these major headings, there may be one or more minor headings, each
descriptive of its bundle of contents. Some of these bundles may be broken down still further by
alphabetic or numeric lists of subtopics.
In a good many cases, it will suffice merely to follow the hierarchy of tables of contents to find
the information sought. In many other cases it will not. For such cases, an alternative approach is to
use the Index.
Use of the Index. A great deal of effort has gone into preparing the index so that, through it,
the reader can locate all the handbook material bearing on a subject. For example, the topic “Pareto
analysis” is found in several sections. The index entry for “Pareto analysis” assembles all uses of the
term Pareto analysis and shows the numbers of the pages on which they may be found.
The fact that information about a single topic is found in more than one section (and even in many
sections) gives rise to criticisms of the organization of the handbook, that is, Why can’t all the information
on one topic be brought together in one place? The answer is that we require multiple and
interconnected uses of knowledge, and hence these multiple appearances cannot be avoided. In fact,
what must be done to minimize duplication is to make one and only one exhaustive explanation at
some logical place and then to use cross-referencing elsewhere. In a sense, all the information on one
topic is brought together—in the index.
Some key words and phrases may be explained in several places in the handbook. However, there
is always one passage which constitutes the major explanation or definition. In the index, the word
“defined” is used to identify this major definition, for example, “Evolutionary operation, defined.”
The index also serves to assemble all case examples or applications under one heading for easy
reference. For example, Section 5 deals with the general approach to quality improvement and
includes examples of the application of this approach. However, additional examples are found in
other sections. The index enables the reader to find these additional examples readily, since the page
numbers are given.
Cross-References. The handbook makes extensive use of cross-references in the text in order
to: (1) guide the reader to further information on a subject and (2) avoid duplicate explanations of
the same subject matter. The reader should regard these cross-references, wherever they occur, as
extensions of the text. Cross-referencing is to either: (1) specific major headings in various sections
or (2) specific figure numbers or table numbers. Study of the referenced material will provide further
A Note on Abbreviations. Abbreviations of names or organizations are usually used only after the
full name has previously been spelled out, for example, American Society for Quality (ASQ). In any
case, all such abbreviations are listed and defined in the index.
The text of the handbook emphasizes the “main road” of quality management know-how, that is, the
comparatively limited number of usual situations which nevertheless occupy the bulk of the time and
attention of practitioners. Beyond the main road are numerous “side roads,” that is, less usual situations
which are quite diverse and which require special solutions.
(The term “side road” is not used in any derogatory sense. The practitioner who faces an unusual
problem must nevertheless find a solution for it.)
As to these side roads, the handbook text, while not complete, nevertheless points the reader to
available solutions. This is done in several ways.
Citations. The handbook cites numerous papers, books, and other bibliographic references. In
most cases these citations also indicate the nature of the special contribution made by the work cited
in order to help the reader to decide whether to go to the original source for elaboration.
Special Bibliographies. Some sections provide supplemental lists of bibliographical
material for further reference. The editors have attempted to restrict the contents of these lists to
items which: (1) bear directly on the subject matter discussed in the text or (2) are of uncommon
interest to the practitioner. A special bibliography in Appendix III lists quality standards and
Literature Search. Papers, books, and other references cited in the handbook contain further
references which can be hunted up for further study. Use can be made of available abstracting and
indexing services. A broad abstracting service in the engineering field is Engineering Index, Inc., 345
E. 47 St., New York, NY 10017. A specialized abstracting service in quality control and applied statistics
is Executive Sciences Institute, Inc., 1005 Mississippi Ave., Davenport, IA 52803. In addition,
various other specialized abstracting services are available on such subjects as reliability, statistical
methods, research and development, and so on.
In searching the literature, the practitioner is well advised to make use of librarians. To an astonishing
degree, library specialists have devised tools for locating literature on any designated subject:
special bibliographies, abstracting services, indexes by subject and author, and so on. Librarians are
trained in the use of these tools, and they maintain an effective communication network among themselves.
The Internet. In the past few years an amazing new tool has appeared on the scene, the Internet.
It is now possible to find almost any book in print and many that are out of print in just a few minutes
using some of the well-designed web sites. Using the new search engines, one can find hundreds
(or even thousands) of articles on numerous topics or by selected authors. Many special sites have
been developed that focus on quality management in its broadest sense. A simple e-mail contact with
an author may bring forth even more unpublished works or research in progress. Sites developed by
university departments doing research in quality are especially useful for searching for specific
examples and new methods and tools.
Author Contact. The written book or paper is usually a condensation of the authors’s knowledge;
that is, what he or she wrote is derived from material which is one or two orders of magnitude
more voluminous than the published work. In some cases it is worthwhile to contact the author for
further elaboration. Most authors have no objection to being contacted, and some of these contacts
lead not only to more information but also to visits and enduring collaboration.
Other Sources. Resourceful people are able to find still other sources of information relating
to the problem at hand. They contact the editors of journals to discover which companies have faced
similar problems, so that they may contact these companies. They contact suppliers and customers
to learn if competitors have found solutions. They attend meetings—such as courses, seminars, and
conferences of professional societies—at which there is discussion of the problem. There is hardly
a problem faced by any practitioner which has not already been actively studied by others.
In many cases a practitioner is faced with adapting, to a special situation, knowledge derived from a
totally different technology, that is, industry, product, or process. Making this transition requires that
he or she identify the commonality, that is, the common principle to which both the special situation
and the derived knowledge correspond.
Often the commonality is managerial in nature and is comparatively easy to grasp. For example,
the concept of self-control is a management universal and is applicable to any person in the
Commonality of a statistical nature is even easier to grasp, since so much information is reduced
to formulas which are indifferent to the nature of the technology involved.
Even in technological matters, it is possible to identify commonalities despite great outward differences.
For example, concepts such as process capability apply not only to manufacturing
processes, but to health care, services and administrative, and support processes as well. In like
manner, the approaches used to make quality improvements by discovering the causes of defects
have been classified into specific categories which exhibit a great deal of commonality despite wide
differences in technology.
In all these situations, the challenge to practitioners is to establish a linkage between their own
situations and those from which the know-how was derived. This linkage is established by discovering
the commonality which makes them both members of one species.
J. M. Juran
The Meanings of “Quality” 2.1
Customer Needs and Conformance to
Specification 2.2
Definitions of Other Key Words 2.2
Satisfaction and Dissatisfaction Are Not
Opposites 2.2
Big Q and Little Q 2.2
The Effect on Income 2.3
The Effect on Costs 2.4
The Juran Trilogy Diagram 2.6
Allocation of Time within The Trilogy
Primitive Societies 2.8
Effects of the Growth of Commerce 2.9
Artisans and Guilds 2.10
Inspection and Inspectors 2.11
Government Involvement in Managing
for Quality 2.11
The Mark or Seal 2.12
The Industrial Revolution 2.12
The Rise of Quality Assurance 2.13
The Twentieth Century and Quality 2.15
Lessons Learned 2.16
Inventions Yet to Come 2.17
This section deals with the fundamental concepts that underlie the subject of managing for quality.
It defines key terms and makes critical distinctions. It identifies the key processes through which
quality is managed. It demonstrates that while managing for quality is a timeless concept, it has
undergone frequent revolution in response to the endless procession of changes and crises faced by
human societies.
The Meanings of “Quality.” Of the many meanings of the word “quality,” two are of critical
importance to managing for quality:
1. “Quality” means those features of products which meet customer needs and thereby provide
customer satisfaction. In this sense, the meaning of quality is oriented to income. The purpose of
such higher quality is to provide greater customer satisfaction and, one hopes, to increase income.
However, providing more and/or better quality features usually requires an investment and hence
usually involves increases in costs. Higher quality in this sense usually “costs more.”
Product features that meet
customer needs
Higher quality enables
companies to:
Increase customer
Make products salable
Meet competition
Increase market share
Provide sales income
Secure premium prices
The major effect is on
Usually, higher quality
costs more.
Freedom from deficiencies
Higher quality enables
companies to:
Reduce error rates
Reduce rework, waste
Reduce field failures,
warranty charges
Reduce customer dissatisfaction
Reduce inspection, test
Shorten time to put new
products on the market
Increase yields, capacity
Improve delivery performance
Major effect is on costs.
Usually, higher quality
costs less.
FIGURE 2.1 The meanings of quality. [Planning for Quality, 2d ed.
(1990). Juran Institute, Inc., Wilton, CT, pp. 1–10.]
2. “Quality” means freedom from deficiencies—freedom from errors that require doing work
over again (rework) or that result in field failures, customer dissatisfaction, customer claims, and so
on. In this sense, the meaning of quality is oriented to costs, and higher quality usually “costs less.”
Figure 2.1 elaborates on these two definitions.
Figure 2.1 helps to explain why some meetings on managing for quality end in confusion.
A meeting of managers is discussing, “Does higher quality cost more, or does it cost less?”
Seemingly they disagree, but in fact some of them literally do not know what the others are talking
about. The culprit is the word “quality,” spelled the same way and pronounced the same way,
but with two meanings.
At one bank the upper managers would not support a proposal to reduce waste because it had the
name “quality improvement.” In their view, higher quality also meant higher cost. The subordinates
were forced to relabel the proposal “productivity improvement” in order to secure approval.
Such confusion can be reduced if training programs and procedures manuals make clear the distinction
between the two meanings of the word “quality.” However, some confusion is inevitable as
long as we use a single word to convey two very different meanings. There have been efforts to clarify
matters by adding supplemental words, such as “positive” quality and “negative” quality. To date,
none of these efforts has gained broad acceptance.
There also have been efforts to coin a short phrase that would clearly and simultaneously define
both the major meanings of the word “quality.” A popular example is “fitness for use.” However, it
is unlikely that any short phrase can provide the depth of meaning needed by managers who are faced
with choosing a course of action. The need is to understand the distinctions set out in Figure 2.1.
Customer Needs and Conformance to Specification. For most quality departments,
the long-standing definition of quality was “conformance to specification.” In effect, they assumed
that products that conformed to specifications also would meet customer needs. This assumption was
logical, since these departments seldom had direct contact with customers. However, the assumption
can be seriously in error. Customer needs include many things not found in product specifications:
service explanations in simple language, confidentiality, freedom from burdensome paperwork,
“one-stop shopping,” and so on. (For elaboration and discussion, see AT&T 1990.)
The new emphasis on customer focus has caused the quality departments to revise their definition
of “quality” to include customer needs that are not a part of the product specification.
Definitions of Other Key Words. The definitions of “quality” include certain key words
that themselves require definition.
Product: The output of any process. To many economists, products include both goods and services.
However, under popular usage, “product” often means goods only.
Product feature: A property possessed by goods or services that is intended to meet customer
Customer: Anyone who is affected by the product or by the process used to produce the product.
Customers may be external or internal.
Customer satisfaction: A state of affairs in which customers feel that their expectations have
been met by the product features.
Deficiency: Any fault (defect or error) that impairs a product’s fitness for use. Deficiencies take
such forms as office errors, factory scrap, power outages, failures to meet delivery dates, and
inoperable goods.
Customer dissatisfaction: A state of affairs in which deficiencies (in goods or services) result in
customer annoyance, complaints, claims, and so on.
In the world of managing for quality, there is still a notable lack of standardization of the meanings
of key words. However, any organization can do much to minimize internal confusion by standardizing
the definitions of key words and phrases. The basic tool for this purpose is a glossary. The glossary then
becomes a reference source for communication of all sorts: reports, manuals, training texts, and so on.
Satisfaction and Dissatisfaction Are Not Opposites. Customer satisfaction comes
from those features which induce customers to buy the product. Dissatisfaction has its origin in deficiencies
and is why customers complain. Some products give little or no dissatisfaction; they do what
the producer said they would do. Yet they are not salable because some competing product has features
that provide greater customer satisfaction.
The early automated telephone exchanges employed electromagnetic analog switching methods.
Recently, there was a shift to digital switching methods, owing to their superior product features.
As a result, analog switching systems, even if absolutely free from product deficiencies, were no
longer salable.
Big Q And Little Q. Definitions of words do not remain static. Sometimes they undergo extensive
change. Such a change emerged during the 1980s. It originated in the growing quality crisis and
is called the concept of “Big Q.”
Table 2.1 shows how the quality “umbrella” has been broadening dramatically. In turn, this
broadening has changed the meanings of some key words. Adoption of Big Q grew during the 1980s,
and the trend is probably irreversible. Those most willing to accept the concept of Big Q have been
the quality managers and the upper managers. Those most reluctant have been managers in the technological
areas and in certain staff functions.
The Effect on Income. Income may consist of sales of an industrial company, taxes collected
by a government body, appropriations received by a government agency, tuitions received by a
school, and donations received by a charity. Whatever the source, the amount of the income relates
in varying degrees to the features of the product produced by the recipient. In many markets, products
with superior features are able to secure superior income, whether through higher share of market
or through premium prices. Products that are not competitive in features often must be sold at
below-market prices.
Product deficiencies also can have an effect on income. The customer who encounters a deficiency
may take action of a cost-related nature: file a complaint, return the product, make a claim,
or file a lawsuit. The customer also may elect instead (or in addition) to stop buying from the guilty
producer, as well as to publicize the deficiency and its source. Such actions by multiple customers
can do serious damage to a producer’s income. Section 7, Quality and Income, is devoted to the ways
in which quality can influence income.
The Effect on Costs. The cost of poor quality consists of all costs that would disappear if there
were no deficiencies—no errors, no rework, no field failures, and so on. This cost of poor quality is
TABLE 2.1 Contrast, Big Q and Little Q
Quality is viewed as:
How to think about
Quality goals are
Cost of poor
Evaluation of
quality is based
mainly on:
Improvement is
directed at:
Training in
managing for
quality is:
Coordination is by:
Source: Planning for Quality, 2d ed. (1990). Juran Institute, Inc., Wilton, CT, pp. 1–12.
Content of little Q
Manufactured goods
Processes directly
related to manufacture
of goods
A technological problem
Clients who buy the
Based on culture of
functional departments
Among factory goals
Costs associated with
deficient manufactured
Conformance to factory
procedures, standards
Departmental performance
Concentrated in the
quality department
The quality manager
Content of big Q
All products, goods,
and services, whether
for sale or not
All process
support; business,
All industries,
service, government,
etc., whether for
profit or not
A business problem
All who are affected,
external and internal
Based on the
universal trilogy
In company business
All costs that would
disappear if
everything were
Responsiveness to
customer needs
Company performance
A quality council of
upper managers
shockingly high. In the early 1980s, I estimated that within the U.S. manufacturing industries, about a
third of the work done consisted of redoing what had already been done. Since then, estimates from a
sample of service industries suggest that a similar situation prevails in service industries generally.
Deficiencies that occur prior to sale obviously add to producers’ costs. Deficiencies that occur after
sale add to customers’ costs as well as to producers’ costs. In addition, they reduce producers’ repeat
sales. Section 8, Quality and Costs, is devoted to the ways in which quality can influence costs.
To attain quality, it is well to begin by establishing the “vision” for the organization, along with policies
and goals. (These matters are treated elsewhere in this handbook, especially in Section 13,
Strategic Deployment.) Conversion of goals into results (making quality happen) is then done
through managerial processes—sequences of activities that produce the intended results. Managing
for quality makes extensive use of three such managerial processes:
 Quality planning
 Quality control
 Quality improvement
These processes are now known as the “Juran trilogy.” They parallel the processes long used to manage
for finance. These financial processes consist of
Financial planning: This process prepares the annual financial budget. It defines the deeds to
be done in the year ahead. It translates those deeds into money. It determines the financial consequences
of doing all those deeds. The final result establishes the financial goals for the organization
and its various divisions and units.
Financial control: This process consists of evaluating actual financial performance, comparing
this with the financial goals, and taking action on the difference—the accountant’s “variance.”
There are numerous subprocesses for financial control: cost control, expense control, inventory
control, and so on.
Financial improvement: This process aims to improve financial results. It takes many forms:
cost-reduction projects, new facilities to improve productivity, new product development to
increase sales, acquisitions, joint ventures, and so on.
These processes are universal—they provide the basis for financial management, no matter what the
type of enterprise is.
The financial analogy helps managers realize that they can manage for quality by using the same
processes of planning, control, and improvement. Since the concept of the trilogy is identical to that
used in managing for finance, managers are not required to change their conceptual approach.
Much of their previous training and experience in managing for finance is applicable to managing
for quality.
While the conceptual approach does not change, the procedural steps differ. Figure 2.2 shows that
each of these three managerial processes has its own unique sequence of activities.
Each of the three processes is also a universal—it follows an unvarying sequence of steps. Each
sequence is applicable in its respective area, no matter what is the industry, function, culture, or
Figure 2.2 shows these unvarying sequences in abbreviated form. Extensive detail is provided in
other sections of this handbook as follows:
Section 3, The Quality Planning Process
Section 4, The Quality Control Process
Section 5, The Quality Improvement Process
The Juran Trilogy Diagram. The three processes of the Juran trilogy are interrelated. Figure
2.3 shows this interrelationship.
The Juran trilogy diagram is a graph with time on the horizontal axis and cost of poor quality on
the vertical axis. The initial activity is quality planning. The planners determine who the customers
are and what their needs are. The planners then develop product and process designs to respond to
those needs. Finally, the planners turn the plans over to the operating forces: “You run the process,
produce the product features, and meet the customers’ needs.”
Chronic and Sporadic. As operations proceed, it soon emerges that the process is unable to produce
100 percent good work. Figure 2.3 shows that over 20 percent of the work must be redone due
to quality deficiencies. This waste is chronic—it goes on and on. Why do we have this chronic
waste? Because the operating process was planned that way.
Under conventional responsibility patterns, the operating forces are unable to get rid of this
planned chronic waste. What they can do is to carry out quality control—to prevent things from getting
worse. Figure 2.3 also shows a sudden sporadic spike that has raised the defect level to over 40
percent. This spike resulted from some unplanned event such as a power failure, process breakdown,
or human error. As a part of their job of quality control, the operating forces converge on the scene
and take action to restore the status quo. This is often called “corrective action,” “troubleshooting,”
“putting out the fire,” and so on. The end result is to restore the error level back to the planned chronic
level of about 20 percent.
The chart also shows that in due course the chronic waste was driven down to a level far below
the original level. This gain came from the third process in the trilogy—quality improvement. In
effect, it was seen that the chronic waste was an opportunity for improvement, and steps were taken
to make that improvement.
The Trilogy Diagram and Product Deficiencies. The trilogy diagram (Figure 2.3) relates to product
deficiencies. The vertical scale therefore exhibits units of measure such as cost of poor quality,
error rate, percent defective, service call rate, and so on. On this same scale, perfection is at zero,
Quality planning
Establish quality
Identify who the
customers are
Determine the needs
of the customers
Develop product
features that
respond to customers’
Develop processes
able to produce the
product features
Establish process
controls; transfer
the plans to the
operating forces
Quality control
Evaluate actual
Compare actual
performance with
quality goals
Act on the
Quality improvement
Prove the need
Establish the
Identify the
improvement projects
Establish project
Provide the teams
with resources,
training, and
motivation to:
Diagnose the causes
Stimulate remedies
Establish controls to
hold the gains
FIGURE 2.2 The three universal processes of managing for quality. [Adapted from
Juran, J. M. (1989). The Quality Trilogy: A Universal Approach to Managing for Quality.
Juran Institute, Inc., Wilton, CT.]
and what goes up is bad. The results of reducing deficiencies are to reduce the cost of poor quality,
meet more delivery promises, reduce customer dissatisfaction, and so on.
The Trilogy Diagram and Product Features. When the trilogy diagram is applied to product features,
the vertical scale changes. Now the scale may exhibit units of measure such as millions of
instructions per second, mean time between failures, percent on-time deliveries, and so on. For such
diagrams, what goes up is good, and a logical, generic vertical scale is “product salability.” (For elaboration
on the Juran trilogy, see Juran 1986.)
Allocation of Time within the Trilogy. An interesting question for managers is, “How do
people allocate their time relative to the processes of the trilogy?” Figure 2.4 is a model designed to
show this interrelationship in a Japanese company (Itoh 1978).
In Figure 2.4 the horizontal scale represents the percentage allocation of any person’s time and
runs from zero to 100 percent. The vertical scale represents levels in the hierarchy. The diagram
shows that the upper managers spend the great majority of their time on planning and improvement.
They spend a substantial amount of time on strategic planning. The time they spend on control is
small and is focused on major control subjects.
At progressively lower levels of the hierarchy, the time spent on strategic planning declines,
whereas the time spent on control and maintenance grows rapidly. At the lowest levels, the time is
dominated by control and maintenance, but some time is still spent on planning and improvement.
A young recruit who joins an organization soon learns that it has in place numerous processes (systems)
to manage its affairs, including managing for quality. The recruit might assume that humans
have always used those processes to manage for quality and will continue to so in the future. Such
FIGURE 2.3 The Juran trilogy diagram. [Adapted from Juran, J. M. (1989). The Quality
Trilogy: A Universal Approach to Managing for Quality. Juran Institute, Inc., Wilton, CT.]
assumptions would be grossly in error. The processes used to manage for quality have undergone
extensive change over the millennia, and there is no end in sight.
Primitive Societies
The Family. Quality is a timeless concept. The origins of ways to manage for quality are hidden in
the mists of the ancient past. Yet we can be sure that humans have always faced problems of quality.
Primitive food-gatherers had to learn which fruits were edible and which were poisonous.
Primitive hunters had to learn which trees supplied the best wood for making bows or arrows. The
resulting know-how was then passed down from generation to generation.
The nuclear human organizational unit was the family. Isolated families were forced to create
self-sufficiency—to meet their own needs for food, clothing, and shelter. There was division of
work among family members. Production was for self-use, so the design, production, and use of a
product were all carried out by the same persons. Whereas the technology was primitive, the coordination
was superb. The same human beings received all inputs and took all remedial action. The
limiting factor for achieving quality was the primitive state of the technology.
The Village: Division of Labor. Villages were created to serve other essential human requirements
such as mutual defense and social needs. The village stimulated additional division of labor and
development of specialized skills. There emerged farmers, hunters, fishermen, and artisans of all
sorts—weavers, potters, shoemakers. By going through the same work cycle over and over again, the
artisans became intimately familiar with the materials used, the tools, the steps in the process, and
the finished product. The cycle included selling the product to users and receiving their feedback on
product performance. The experience derived from this intimate familiarity then enabled human
ingenuity to take the first steps toward the evolution of technology.
The Village Marketplace: Caveat Emptor. As villages grew, the village marketplace appeared,
where artisans and buyers met on scheduled market days. In this setting, producer and user met face
to face with the goods between them. The goods typically were natural products or were made from
natural materials. The producers and purchasers had long familiarity with the products, and the quality
of the products could to a high degree be judged by the unaided human senses.
FIGURE 2.4 The Itoh model. [Adapted from Management for Quality, 4th ed. (1987).
Juran Institute, Inc., Wilton, CT, p. 18.]
Under such a state of affairs, the village magistrates tended to avoid being drawn into quality disputes
between seller and buyer. This forced buyers to be vigilant so as to protect themselves against
poor quality. In effect, the seller was responsible for supplying the goods, but the buyer became
responsible for supplying the quality “assurance.” This arrangement was known as caveat emptor—
”let the buyer beware.” Thus buyers learned to beware by use of product inspection and test. They
looked closely at the cloth, smelled the fish, thumped the melon, tasted a grape. Their failure to
beware was at their own peril. In the village marketplace, caveat emptor was quite a sensible doctrine.
It is widely applied to this day in villages all over the world.
A further force in the village marketplace was the fact of common residence. Producer and buyer
both lived in the same village. Each was subject to scrutiny and character evaluation by the villagers.
Each also was subject to village discipline. For the artisan, the stakes were high. His status and livelihood
(and those of his family) were closely tied to his reputation as a competent and honest artisan.
In this way, the concept of craftsmanship became a quiet yet powerful stimulus to maintain a high
level of quality.
Effects of the Growth of Commerce. In due course villages expanded into towns and
cities, and improved transport opened the way to trade among regions.
A famous example of organized multiregional trade was the Hanseatic League which was centered
among the cities of northern Europe from about the 1200s to the 1600s. Its influence extended into
Scandinavia and Russia as well as to the Mediterranean and Black seas (von der Porten 1994).
Under trade among regions, producer and user could no longer meet face to face in the marketplace.
Products were now made by chains of suppliers and processors. Marketing was now done by
chains of marketers. The buyers’ direct point of contact was now with some merchant rather than
with the producer. All this reduced the quality protections inherent in the village marketplace to a
point requiring invention of new forms of quality assurance. One such invention was the quality
Quality Warranties. Early quality warranties were no doubt in oral form. Such warranties were
inherently difficult to enforce. Memories differed as to what was said and meant. The duration of the
warranty might extend beyond the life of the parties. Thus the written warranty was invented.
An early example was on a clay tablet found amid the ruins of Nippur in ancient Babylon. It involved
a gold ring set with an emerald. The seller guaranteed that for twenty years the emerald would not fall out
of the gold ring. If it did fall out of the gold ring before the end of twenty years, the seller agreed to pay
to the buyer an indemnity of ten mana of silver. The date is the equivalent of 429 B.C. (Bursk et al. 1962,
vol. I, p. 71).
Quality warranties are now widely used in all forms of trade and commerce. They stimulate producers
to give priority to quality and stimulate sellers to seek out reliable sources of supply. So great
is their importance that recent legislation has imposed standards to ensure that the wording of warranties
does not mislead the buyers.
Quality Specifications. Sellers need to be able to communicate to buyers the nature of what they
have to sell. Buyers need to be able to communicate to sellers the nature of what they want to buy.
In the village marketplace, oral communication could take place directly between producer and
buyer. With the growth of commerce, communication expanded to include chains of producers and
chains of merchants who often were widely separated. New forms of communications were needed,
and a major invention was the written quality specification. Now quality information could be communicated
directly between designer and producer or between seller and buyer no matter how great
the distance between them and how complex the nature of the product.
Like warranties, written specifications are of ancient origin. Examples have been found in Egyptian
papyrus scrolls over 3500 years old (Durant 1954). Early specifications focused on defining products
and the processes for producing them. In due course the concept was extended to defining the materials
from which the products were made. Then, as conflicts arose because sellers and buyers used different
methods of test, it became necessary to establish inspection and test specifications as well.
Measurement. The emergence of inspection and test specifications led to the evolution of measuring
instruments. Instruments for measuring length, volume, and time evolved thousands of years
ago. Instruments have continued to proliferate, with ever-increasing precision. In recent centuries,
the precision of measurement of time has increased by over ten orders of magnitude (Juran 1995,
Chapter 10).
Artisans and Guilds. The artisan’s possession of the skills of a trade was a source of income
and status as well as self-respect and respect from the community. However, as villages grew into
towns and cities, the numbers of artisans grew as well. The resulting competition became destructive
and threatened the benefits derived from craftsmanship.
To perpetuate their benefits, the artisans within a trade organized trade unions—guilds. Each
guild then petitioned the city authorities to confer on the guild members a monopoly on practicing
their trade.
Guilds flourished for centuries during the Middle Ages until the Industrial Revolution reduced
their influence. They used their monopolistic powers chiefly to provide a livelihood and security for
their members. The guilds also provided extensive social services to their members. (For elaboration,
see Bursk et al. 1962, vol. III, pp. 1656–1678.)
The Guild Hierarchy. Each guild maintained a hierarchy of (usually) three categories of workers:
the apprentice, the journeyman, and the master. Considerable formality surrounded the entry into
each category.
At the bottom was the apprentice or novice, whose entry was through an indenture—a formal
contract that bound the apprentice to serve a master for a specified period of years. In turn, the master
became responsible for teaching the trade to the apprentice.
To qualify for promotion, the apprentice was obliged to serve out the full term of the indenture.
In addition, he was required to pass an examination by a committee of masters. Beyond the oral part
of the examination, the apprentice was required to produce a perfect piece of work—a
masterpiece—that was then inspected by the examination committee. Success in the examination led
to a ceremonial admission to the status of journeyman.
The journeyman’s right to practice the trade was limited. He could become an employee of a master,
usually by the day. He also could journey to other towns, seeking employment in his trade. Only
after admission to the rank of master could he set up shop on his own.
Admission to the rank of master required first that there be an opening. Guilds imposed limits on
the numbers of masters in their areas. On the death or retirement of an active master, the guild would
decide whether to fill that opening. If so, a journeyman would be selected and admitted, again
through a formal ceremony.
Guilds and Quality Planning. Guilds were active in managing for quality, including quality planning.
They established specifications for input materials, manufacturing processes, and finished
products, as well as for methods of inspection and test.
Guilds and Quality Control. Guild involvement in quality control was extensive. They maintained
inspections and audits to ensure that artisans followed the quality specifications. They established
means of “traceability” to identify the producer. In addition, some applied their “mark” to finished
products as added assurance to consumers that quality met guild standards.
Control by the guilds also extended to sales. The sale of poor-quality goods was forbidden, and
offenders suffered a range of punishments—all the way from fines to expulsion from membership.
The guilds also established prices and terms of sale and enforced them.
Guilds and Quality Improvement. An overriding guild policy was solidarity—to maintain equality
of opportunity among members. To this end, internal competition among members was limited
to “honest” competition. Quality improvement through product or process innovation was not considered
to be “honest” competition. This limitation on quality improvement did indeed help to maintain
equality among members, but it also made the guild increasingly vulnerable to competition from
other cities that did evolve superior products and processes.
Guilds and External Forces. The guilds were able to control internal competition, but external
competition was something else. Some external competition came in the form of jurisdictional disputes
with other guilds, which consumed endless hours of negotiation. More ominous was competition
from other cities, which could be in quality as well as in price and value.
The policy of solidarity stifled quality improvement and thereby became a handicap to remaining
competitive. Thus the guilds urged the authorities to restrict imports of foreign goods. They also
imposed strict rules to prevent their trade secrets from falling into the hands of foreign competitors.
(The Venetian glass industry threatened capital punishment to those who betrayed such secrets.)
Inspection and Inspectors. The concepts of inspection and inspectors are of ancient origin.
Wall paintings and reliefs in Egyptian tombs show the inspections used during stone construction
projects. The measuring instruments included the square, level, and plumb bob for alignment control.
Surface flatness of stones was checked by “boning rods” and by threads stretched across the
faces of the stone blocks:
As shops grew in size, the function of inspection gave rise to the full-time job of inspector. In due
course, inspectors multiplied in numbers to become the basis for inspection departments, which in turn
gave birth to modern quality departments. (Singer et al. 1954, vol. I, p. 481).
Government Involvement in Managing for Quality. Governments have long involved
themselves in managing for quality. Their purposes have included protecting the safety and health of
citizens, defending and improving the economics of the state, and protecting consumers against
fraud. Each of these purposes includes some aspect of managing for quality.
Safety and Health of the Citizens. Early forms of protection of safety and health were after-thefact
measures. The Code of Hammurabi (c. 2000 B.C.) prescribed the death penalty for any builder
of a house that later collapsed and killed the owner. In medieval times, the same fate awaited the
baker who inadvertently had mixed rat poison with the flour.
Economics of the State. With the growth of trade between cities, the quality reputation of a city
could be an asset or a liability. Many cities took steps to protect their reputation by imposing quality
controls on exported goods. They appointed inspectors to inspect finished products and affix a seal
to certify as to quality. This concept was widely applied to high-volume goods such as textiles.
Continued growth of commerce then created competition among nations, including competition
in quality. Guilds tended to stifle quality improvement, but governments favored improving the quality
of domestic goods in order to reduce imports and increase exports. For example, in the late sixteenth
century, James VI of Scotland imported craftsmen from the Low Countries to set up a textile
factory and to teach their trade secrets to Scottish workers (Bursk et al. 1962, vol. IV, pp.
Consumer Protection. Many states recognized that as to some domestic trade practices, the rule of
caveat emptor did not apply. One such practice related to measurement. The states designed official
standard tools for measuring length, weight, volume, and so on. Use of these tools was then mandated,
and inspectors were appointed to ensure compliance. (See, for example, Juran 1995, chap. 1.)
The twentieth century witnessed a considerable expansion in consumer protection legislation. (For
elaboration, see Juran 1995, chap. 17.)
The Mark or Seal. A mark or seal has been applied to products over the centuries to serve multiple
purposes. Marks have been used to
Identify the producer, whether artisan, factory, town, merchant, packager, or still others: Such
identification may serve to fix responsibility, protect the innocent against unwarranted blame,
enable buyers to choose from among multiple makers, advertise the name of the maker, and
so on.
Provide traceability: In mass production, use of lot numbers helps to maintain uniformity of
product in subsequent processing, designate expiration dates, make selective product recalls, and
so on.
Provide product information, such as type and quantities of ingredients used, date when made,
expiration dates, model number, ratings (such as voltage, current), and so on.
Provide quality assurance: This was the major purpose served by the marks of the guilds and
towns. It was their way of telling buyers, “This product has been independently inspected and has
good quality.”
An aura of romance surrounds the use of seals. The seals of some medieval cities are masterpieces
of artistic design. Some seals have become world-renowned. An example is the British “hallmark”
that is applied to products made of precious metals.
The Industrial Revolution. The Industrial Revolution began in Europe during the mideighteenth
century. Its origin was the simultaneous development of power-driven machinery and
sources of mechanical power. It gave birth to factories that soon outperformed the artisans and small
shops and made them largely obsolete.
The Factory System: Destruction of Crafts. The goals of the factories were to raise productivity
and reduce costs. Under the craft system, productivity had been low due to primitive technology,
whereas costs had been high due to the high wages of skilled artisans. To reach their goals, the factories
reengineered the manufacturing processes. Under the craft system, an artisan performed every
one of the numerous tasks needed to produce the final product—pins, shoes, barrels, and so on.
Under the factory system, the tasks within a craft were divided up among several or many factory
workers. Special tools were designed to simplify each task down to a short time cycle. A worker then
could, in a few hours, carry out enough cycles of his or her task to reach high productivity.
Adam Smith, in his book, The Wealth of Nations, was one of the first to publish an explanation
of the striking difference between manufacture under the craft system versus the factory system. He
noted that pin making had been a distinct craft, consisting of 18 separate tasks. When these tasks
were divided among 10 factory workers, production rose to a per-worker equivalent of 4800 pins a
day, which was orders of magnitude higher than would be achieved if each worker were to produce
pins by performing all 18 tasks (Smith 1776). For other types of processes, such as spinning or weaving,
power-driven machinery could outproduce hand artisans while employing semiskilled or
unskilled workers to reduce labor costs.
The broad economic result of the factory system was mass production at low costs. This made
the resulting products more affordable and contributed to economic growth in industrialized countries,
as well as to the associated rise of a large “middle class.”
Quality Control under the Factory System. The factory system required associated changes in the
system of quality control. When craft tasks were divided among many workers, those workers were
no longer their own customers, over and over again. The responsibility of workers was no longer to
provide satisfaction to the buyer (also customer, user). Few factory workers had contact with buyers.
Instead, the responsibility became one of “make it like the sample” (or specification).
Mass production also brought new technological problems. Products involving assemblies of bits
and pieces demanded interchangeability of those bits and pieces. Then, with the growth of technology
and of interstate commerce, there emerged the need for standardization as well. All this required
greater precision throughout—machinery, tools, measurement. (Under the craft system, the artisan
fitted and adjusted the pieces as needed).
In theory, such quality problems could be avoided during the original planning of the manufacturing
processes. Here the limitation rested with the planners—the “master mechanics” and shop
supervisors. They had extensive, practical experience, but their ways were empirical, being rooted in
craft practices handed down through the generations. They had little understanding of the nature of
process variation and the resulting product variation. They were unschooled in how to collect and
analyze data to ensure that their processes had “process capability” to enable the production workers
to meet the specifications. Use of such new concepts had to await the coming of the twentieth
Given the limitations of quality planning, what emerged was an expansion of inspection by
departmental supervisors supplemented by full-time inspectors. Where inspectors were used, they
were made responsible to the respective departmental production supervisors. The concept of a special
department to coordinate quality activities broadly also had to await the coming of the twentieth
Quality Improvement. The Industrial Revolution provided a climate favorable for continuous quality
improvement through product and process development. For example, progressive improvements
in the design of steam engines increased their thermal efficiency from 0.5 percent in 1718 to 23.0
percent in 1906 (Singer et al. 1958, vol. IV). Inventors and entrepreneurs emerged to lead many
countries into the new world of technology and industrialization. In due course, some companies created
internal sources of inventors—research laboratories to carry out product and process development.
Some created market research departments to carry out the functions of entrepreneurship.
In contrast, the concept of continuous quality improvement to reduce chronic waste made little
headway. One likely reason is that most industrial managers give higher priority to increasing
income than to reducing chronic waste. The guilds’ policy of solidarity, which stifled quality
improvement, also may have been a factor. In any event, the concept of quality improvement to
reduce chronic waste did not find full application until the Japanese quality revolution of the twentieth
The Taylor System of Scientific Management. A further blow to the craft system came from F. W.
Taylor’s system of “scientific management.” This originated in the late nineteenth century when
Taylor, an American manager, wanted to increase production and productivity by improving manufacturing
planning. His solution was to separate planning from execution. He brought in engineers
to do the planning, leaving the shop supervisors and the work force with the narrow responsibility
of carrying out the plans.
Taylor’s system was stunningly successful in raising productivity. It was widely adopted in the
United States but not so widely adopted elsewhere. It had negative side effects in human relations,
which most American managers chose to ignore. It also had negative effects on quality. The American
managers responded by taking the inspectors out of the production departments and placing them in
newly created inspection departments. In due course, these departments took on added functions to
become the broad-based quality departments of today. (For elaboration, see Juran 1995, chap. 17.)
The Rise of Quality Assurance. The anatomy of “quality assurance” is very similar to that
of quality control. Each evaluates actual quality. Each compares actual quality with the quality goal.
Each stimulates corrective action as needed. What differs is the prime purpose to be served.
Under quality control, the prime purpose is to serve those who are directly responsible for conducting
operations—to help them regulate current operations. Under quality assurance, the prime
purpose is to serve those who are not directly responsible for conducting operations but who have a
need to know—to be informed as to the state of affairs and, hopefully, to be assured that all is well.
In this sense, quality assurance has a similarity to insurance. Each involves spending a small sum
to secure protection against a large loss. In the case of quality assurance, the protection consists of
an early warning that may avoid the large loss. In the case of insurance, the protection consists of
compensation after the loss.
Quality Assurance in the Village Marketplace. In the village marketplace, the buyers provided
much of the quality assurance through their vigilance—through inspection and test before buying the
product. Added quality assurance came from the craft system—producers were trained as apprentices
and were then required to pass an examination before they could practice their trade.
Quality Assurance through Audits. The growth of commerce introduced chains of suppliers and
merchants that separated consumers from the producers. This required new forms of quality assurance,
one being quality warranties. The guilds created a form of quality assurance by establishing
product and process standards and then auditing to ensure compliance by the artisans. In addition,
some political authorities established independent product inspections to protect their quality reputations
as exporters.
Audit of Suppliers’ Quality Control Systems. The Industrial Revolution stimulated the rise of large
industrial companies. These bought equipment, materials, and products on a large scale. Their early
forms of quality assurance were mainly through inspection and test. Then, during the twentieth century,
there emerged a new concept under which customers defined and mandated quality control
systems. These systems were to be instituted and followed by suppliers as a condition for becoming
and remaining suppliers. This concept was then enforced by audits, both before and during the life
of the supply contracts.
At first, this concept created severe problems for suppliers. One was the lack of standardization.
Each buying company had its own idea of what was a proper quality control system, so each supplier
was faced with designing its system to satisfy multiple customers. Another problem was that of
multiple audits. Each supplier was subject to being audited by each customer. There was no provision
for pooling the results of audits into some common data bank, and customers generally were
unwilling to accept the findings of audits conducted by personnel other than their own. The resulting
multiple audits were especially burdensome to small suppliers.
In recent decades, steps have been taken toward standardization by professional societies, by
national standardization bodies, and most recently, by the International Standards Organization
(ISO). ISO’s 9000 series of standards for quality control systems is now widely accepted among
European companies. There is no legal requirement for compliance, but as a marketing matter, companies
are reluctant to be in a position in which their competitors are certified as complying to ISO
9000 standards but they themselves are not.
There remains the problem of multiple audits. In theory, it is feasible for one audit to provide information
that would be acceptable to all buyers. This is already the case in quality audits conducted by
Underwriters’ Laboratories and in financial audits conducted by Dun & Bradstreet. Single audits may
in the future become feasible under the emerging process for certification to the ISO 9000 series.
Extension to Military Procurement. Governments have always been large buyers, especially for
defense purposes. Their early systems of quality assurance consisted of inspection and test. During
the twentieth century, there was a notable shift to mandating quality control systems and then
using audits to ensure conformance to the mandated systems. The North Atlantic Treaty
Organization (NATO) evolved an international standard—the Allied Quality Assurance
Publications (AQAP)—that includes provisions to minimize multiple audits. (For elaboration, see
Juran 1977.)
Resistance to Mandated Quality Control Systems. At the outset, suppliers resisted the mandated
quality control systems imposed by their customers. None of this could stop the movement toward
quality assurance. The economic power of the buyers was decisive. Then, as suppliers gained experience
with the new approach, they realized that many of its provisions were simply good business
practice. Thus the concept of mandated quality control systems seems destined to become a permanent
feature of managing for quality.
Shift of Responsibility. It should be noted that the concept of mandating quality control systems
involves a major change of responsibility for quality assurance. In the village marketplace, the pro-
ducer supplies the product, but the buyer has much of the responsibility for supplying the quality
assurance. Under mandated quality control systems, the producer becomes responsible for supplying
both the product and the quality assurance. The producer supplies the quality assurance by
 Adopting the mandated system for controlling quality
 Submitting the data that prove that the system is being followed
The buyers’ audits then consist of seeing to it that the mandated system is in place and that the system
is indeed being followed.
The Twentieth Century and Quality. The twentieth century witnessed the emergence of
some massive new forces that required responsive action. These forces included an explosive growth
in science and technology, threats to human safety and health and to the environment, the rise of the
consumerism movement, and intensified international competition in quality.
An Explosive Growth in Science and Technology. This growth made possible an outpouring of
numerous benefits to human societies: longer life spans, superior communication and transport,
reduced household drudgery, new forms of education and entertainment, and so on. Huge new industries
emerged to translate the new technology into these benefits. Nations that accepted industrialization
found it possible to improve their economies and the well-being of their citizenry.
The new technologies required complex designs and precise execution. The empirical methods of
earlier centuries were unable to provide appropriate product and process designs, so process yields
were low and field failures were high. Companies tried to deal with low yields by adding inspections
to separate the good from the bad. They tried to deal with field failures through warranties and customer
service. These solutions were costly, and they did not reduce customer dissatisfaction. The
need was to prevent defects and field failures from happening in the first place.
Threats to Human Safety and Health and to the Environment. With benefits from technology
came uninvited guests. To accept the benefits required changes in lifestyle, which, in turn, made
quality of life dependent on continuity of service. However, many products were failure-prone,
resulting in many service interruptions. Most of these were minor, but some were serious and even
frightening—threats to human safety and health, as well as to the environment.
Thus the critical need became quality. Continuity of the benefits of technology depended on
the quality of the goods and services that provided those benefits. The frequency and severity of the
interruptions also depended on quality—on the continuing performance and good behavior of the
products of technology. This dependence came to be known as “life behind the quality dikes.” (For
elaboration, see Juran 1970.)
Expansion of Government Regulation of Quality. Government regulation of quality is of ancient
origin. At the outset, it focused mainly on human safety and was conducted “after the fact”—laws
provided for punishing those whose poor quality caused death or injury. Over the centuries, there
emerged a trend to regulation “before the fact”—to become preventive in nature.
This trend was intensified during the twentieth century. In the field of human health, laws were
enacted to ensure the quality of food, pharmaceuticals, and medical devices. Licensing of practitioners
was expanded. Other laws were enacted relating to product safety, highway safety, occupational
safety, consumer protection, and so on.
Growth of government regulation was a response to twentieth-century forces as well as a force in
its own right. The rise of technology placed complex and dangerous products in the hands of amateurs—
the public. Government regulation then demanded product designs that avoided these dangers.
To the companies, this intervention then became a force to be reckoned with. (For elaboration,
see Juran 1995, chap. 17.)
The Rise of the Consumerism Movement. Consumers welcomed the features offered by the new
products but not the associated new quality problems. The new products were unfamiliar—most
consumers lacked expertise in technology. Their senses were unable to judge which of the competing
products to buy, and the claims of competing companies often were contradictory.
When products failed in service, consumers were frustrated by vague warranties and poor service.
“The system” seemed unable to provide recourse when things failed. Individual consumers were
unable to fight the system, but collectively they were numerous and hence potentially powerful, both
economically and politically. During the twentieth century, a “consumerism” movement emerged to
make this potential a reality and to help consumers deal more effectively with these problems. This
same movement also was successful in stimulating new government legislation for consumer protection.
(For elaboration, see Juran 1995, chap. 17.)
Intensified International Competition in Quality. Cities and countries have competed for centuries.
The oldest form of such competition was probably in military weaponry. This competition
then intensified during the twentieth century under the pressures of two world wars. It led to the
development of new and terrible weapons of mass destruction.
A further stimulus to competition came from the rise of multinational companies. Large companies
had found that foreign trade barriers were obstacles to export of their products. To get around
these barriers, many set up foreign subsidiaries that then became their bases for competing in foreign
markets, including competition in quality.
The most spectacular twentieth-century demonstration of the power of competition in quality
came from the Japanese. Following World War II, Japanese companies discovered that the West was
unwilling to buy their products—Japan had acquired a reputation for making and exporting shoddy
goods. The inability to sell became an alarm signal and a stimulus for launching the Japanese quality
revolution during the 1950s. Within a few decades, that revolution propelled Japan into a position
of world leadership in quality. This quality leadership in turn enabled Japan to become an economic
superpower. It was a phenomenon without precedent in industrial history.
The cumulative effect of these massive forces has been to “move quality to center stage.” Such a
massive move logically should have stimulated a corresponding response—a revolution in managing
for quality. However, it was difficult for companies to recognize the need for such a revolution—
they lacked the necessary alarm signals. Technological measures of quality did exist on the
shop floors, but managerial measures of quality did not exist in the boardrooms. Thus, except for
Japan, the needed quality revolution did not start until very late in the twentieth century. To make
this revolution effective throughout the world, economies will require many decades—the entire
twenty-first century. Thus, while the twentieth century has been the “century of productivity,” the
twenty-first century will be known as the “century of quality.”
The failure of the West to respond promptly to the need for a revolution in quality led to a widespread
crisis. The 1980s then witnessed quality initiatives being taken by large numbers of companies.
Most of these initiatives fell far short of their goals. However, a few were stunningly successful and produced
the lessons learned and role models that will serve as guides for the West in the decades ahead.
Lessons Learned. Companies that were successful in their quality initiatives made use of
numerous strategies. Analysis shows that despite differences among the companies, there was much
commonality—a lengthy list of strategies was common to most of the successful companies. These
common strategies included
Customer focus: Providing customer satisfaction became the chief operating goal.
Quality has top priority: This was written into corporate policies.
Strategic quality planning: The business plan was opened up to include planning for quality.
Benchmarking: This approach was adopted in order to set goals based on superior results
already achieved by others.
Continuous improvement: The business plan was opened up to include goals for quality
improvement. It was recognized that quality is a moving target.
Training in managing for quality: Training was extended beyond the quality department to all
functions and levels, including upper managers.
Big Q was adopted to replace little Q.
Partnering: Through cross-functional teams, partnering was adopted to give priority to company
results rather than to functional goals. Partnering was extended to include suppliers and
Employee empowerment: This was introduced by training and empowering the work force to
participate in planning and improvement, including the concept of self-directed teams.
Motivation: This was supplied through extending the use of recognition and rewards for
responding to the changes demanded by the quality revolution.
Measurements were developed to enable upper managers to follow progress toward providing
customer satisfaction, meeting competition, improving quality, and so on.
Upper managers took charge of managing for quality by recognizing that certain responsibilities
were not delegable—they were to be carried out by the upper managers, personally.
These responsibilities included
 Serve on the quality council
 Establish the quality goals
 Provide the needed resources
 Provide quality-oriented training
 Stimulate quality improvement
 Review progress
 Give recognition
 Revise the reward system
Inventions Yet to Come. Many of the strategies adopted by the successful companies are
without precedent in industrial history. As such, they must be regarded as experimental. They did
achieve results for the role model companies, but they have yet to demonstrate that they can achieve
comparable results in a broader spectrum of industries and cultures. It is to be expected that the
efforts to make such adaptations will generate new inventions, new experiments, and new lessons
learned. There is no end in sight.
This section of the handbook has drawn extensively from the following two books:
Juran, J. M. (ed.) (1995). A History of Managing for Quality. Sponsored by Juran Foundation, Inc. Quality Press,
Milwaukee Press, WI.
Upper Management and Quality: Making Quality Happen, 6th ed. (1993) Juran Institute, Inc., Wilton, CT.
The author is grateful to Juran Foundation, Inc., and Juran Institute, Inc., for permission to quote
from these works.
AT&T Quality Library (1990). Achieving Customer Satisfaction. Bell Laboratories Quality Information Center,
Indianapolis, IN.
Bursk, Edward C., Clark, Donald T., and Hidy, Ralph W. (1962). The World of Business, vol. I, p. 71, vol. III, pp.
1656–1678, vol. IV, pp. 2283–2285. Simon and Shuster, New York.
Davies, Norman De G. (1943). The Tomb of Rekh-mi-re at Thebes, vol. 2, Plate LXII. Metropolitan Museum of
Art, New York.
Durant, Will (1954). The Story of Civilization, Part I: Our Oriental Heritage, pp. 182–183. Simon and Schuster,
New York.
Itoh, Yasuro (1978). Upbringing of Component Suppliers Surrounding Toyota. International Conference on
Quality Control, Tokyo.
Juran, J. M. (1970). “Consumerism and Product Quality.” Quality Progress, July, pp. 18–27.
Juran, J. M. (1977). Quality and Its Assurance—An Overview. Second NATO Symposium on Quality and Its
Assurance, London.
Juran, J. M. (1986). “The Quality Trilogy: A Universal Approach to Managing for Quality.” Quality Progress,
August, pp. 19–24.
Juran, J. M. (1995) ed. A History of Managing for Quality. Sponsored by Juran Foundation, Inc. Quality Press,
Milwaukee, WI.
Singer, Charles, Holmyard, E. J., and Hall, A. R. (eds.) (1954). A History of Technology, vol. I, Fig. 313, p. 481.
Oxford University Press, New York.
Singer, Charles, Holmyard, E. J., and Hall, A. R. (eds.) (1958). A History of Technology, vol. IV, p. 164. Oxford
University Press, New York.
Smith, Adam (1776). The Wealth of Nations. Random House, New York, published in 1937.
Upper Management and Quality: Making Quality Happen, 6th ed. Juran Institute, Inc., Wilton, CT.
von der Porten, Edward (1994). “The Hanseatic League, Europe’s First Common Market.” National Geographic,
October, pp. 56–79.
John F. Early and O. John Coletti
Identification of Projects 3.4
Prepare Mission Statement 3.4
Basis for Establishing Quality Goals 3.5
New Product Policies 3.8
Establish Team 3.8
Types of External Customers 3.10
Internal Customers 3.11
Identifying Customers 3.11
Stated Needs and Real Needs 3.13
Perceived Needs 3.14
Cultural Needs 3.14
Needs Traceable to Unintended Use 3.14
Human Safety 3.14
“User Friendly” 3.15
Promptness of Service 3.15
Customer Needs Related to Deficiencies
Warranties 3.15
Effect of Complaint Handling on Sales
Keeping Customers Informed 3.16
Plan to Collect Customers’ Needs 3.16
Discovering Mustang Customer Needs
Analyze and Prioritize Customer Needs
Establish Units of Measurement and
Sensors 3.22
Translating and Measuring Mustang
Customer Needs 3.24
Group Together Related Customer Needs
Determine Methods for Identifying
Product Features 3.26
Standards, Regulations, and Policies
Criteria for Design 3.27
Criteria for Setting Product Feature Goals
Measuring Product Features Goals 3.29
Develop Detailed Product Features and
Goals 3.30
Optimize Product Features and Goals
Set and Publish Final Product Design
Review Product Goals 3.36
Identify Operating Conditions 3.36
Collect Known Information on Alternate
Processes 3.37
Select General Process Design 3.39
Identify Process Features and Goals 3.42
Identify Detailed Process Features and
Goals 3.42
Design for Critical Factors and Human
Error 3.43
Optimize Process Features and Goals
Establish Process Capability 3.45
Set and Publish Final Process Features
and Goals 3.45
Identify Controls Needed 3.46
Design Feedback Loop 3.47
Optimize Self-Control and Self-Inspection
Audit Plan for the Transfer 3.49
Implement Plan and Validate Transfer
*In the Fourth Edition, material covered by this section was supplied by Joseph M. Juran and Frank M. Gryna in sections on
Companywide Planning for Quality, Product Development, and Manufacturing Planning.
“Quality planning,” as used here, is a structured process for developing products (both goods and services)
that ensures that customer needs are met by the final result. The tools and methods of quality
planning are incorporated along with the technological tools for the particular product being developed
and delivered. Designing a new automobile requires automotive engineering and related disciplines,
developing an effective care path for juvenile diabetes will draw on the expert methods of specialized
physicians, and planning a new approach for guest services at a resort will require the techniques of an
experienced hotelier. All three need the process, methods, tools, and techniques of quality planning to
ensure that the final designs for the automobile, diabetic care, and resort services not only fulfill the
best technical requirements of the relevant disciplines but also meet the needs of the customers who
will purchase and benefit from the products.
The quality planning process and its associated methods, tools, and techniques have been developed
because in the history of modern society, organizations have rather universally demonstrated a consistent
failure to produce the goods and services that unerringly delight their customers. As a customer,
everyone has been dismayed time and time again when flights are delayed, radioactive contamination
spreads, medical treatment is not consistent with best practices, a child’s toy fails to function,
a new piece of software is not as fast or user-friendly as anticipated, government responds with glacial
speed (if at all), or a home washing machine with the latest high-tech gadget delivers at higher cost
clothes that are no cleaner than before. These frequent, large quality gaps are really the compound
result of a number of smaller gaps illustrated in Figure 3.1.
The first component of the quality gap is the understanding gap, that is, lack of understanding of
what the customer needs. Sometimes this gap opens up because the producer simply fails to consider
who the customers are and what they need. More often the gap is there because the supplying
organization has erroneous confidence in its ability to understand exactly what the customer really
needs. The final perception gap in Figure 3.1 also arises from a failure to understand the customer
and the customer needs. Customers do not experience a new suit of clothes or the continuity in service
from a local utility simply based on the technical merits of the product. Customers react to
how they perceive the good or service provides them with a benefit.
Understanding of Needs
Design of Product
Capability to Deliver Design
Actual Delivery
Customer Expectations
Customer Perception of Delivery
Understanding gap
Design gap
Process gap
Operations gap
Perception gap
FIGURE 3.1 The quality gap and its constituent gaps. [Inspired by A. Parasuraman, Valarie A. Zeithami, and
Leonard L. Berry (1985). “A Conceptual Model for Service Quality and Its Implications for Further Research.”
Journal of Marketing, Fall, pp. 41–50.]
The second constituent of the quality gap is a design gap. Even if there were perfect knowledge
about customer needs and perceptions, many organizations would fail to create designs for their
goods and services that are fully consistent with that understanding. Some of this failure arises from
the fact that the people who understand customers and the disciplines they use for understanding customer
needs are often systematically isolated from those who actually create the designs. In addition,
designers—whether they design sophisticated equipment or delicate human services—often lack the
simple tools that would enable them to combine their technical expertise with an understanding of
the customer needs to create a truly superior product.
The third gap is the process gap. Many splendid designs fail because the process by which the
physical product is created or the service is delivered is not capable of conforming to the design consistently
time after time. This lack of process capability is one of the most persistent and bedeviling
failures in the total quality gap.
The fourth gap is the operations gap. The means by which the process is operated and controlled
may create additional deficiencies in the delivery of the final good or service.
Quality planning provides the process, methods,
tools, and techniques for closing each of these
component gaps and thereby ensuring that the
final quality gap is at a minimum. Figure 3.2
summarizes at a high level the basic steps of
quality planning. The remainder of this section
will provide the details and examples for each of
these steps.
The first step, establish the project, provides the clear goals, direction, and infrastructure required
if the constituent quality gaps are to be closed. The next step provides for systematic identification
of all the customers. It is impossible to close the understanding gap if there is the least bit of uncertainty,
fuzziness, or ignorance about who all the customers are.
The discovery of customer needs in the third step provides the full and complete understanding
required for a successful product design to meet those needs. It also evaluates customer perceptions
explicitly so that the final perception gap can be avoided.
The develop product step uses both quality planning tools and the technology of the particular
industry to create a design that is effective in meeting the customer needs, thereby closing the design
gap. The process gap is closed in the next step, develop process. Quality planning techniques ensure
that the process is capable of delivering the product as it was designed, consistently, time after time.
Finally, the operations gap is closed by developing process controls that keep the process operating
at its full capability. Successful elimination of the operations gap also depends on an effective
transfer of the plans to the operating forces. A strong transfer plan, executed well, will provide operations
with all the processes, techniques, materials, equipment, skills, and so on to delight customers
on a continuing basis.
The remainder of this section will provide details, practical guidance, and examples for each of
these steps. Many detailed examples will be included of how Ford Motor Company applied the principles
of quality planning to develop its 1994 Ford Mustang.
A quality planning project is the organized work needed to prepare an organization to deliver a new
or revised product, following the steps associated with quality planning. Generally speaking, the
following activities are associated with establishing a quality planning project:
• Establish the project
• Identify the customers
• Discover the customer needs
• Develop the product
• Develop the process
• Develop the controls and
transfer to operations
FIGURE 3.2 Quality planning steps. (Juran Institute,
Inc., Copyright 1994. Used by permission.)
 Identify which projects are required to fulfill the organization’s strategy.
 Prepare a mission statement for each project.
 Establish a team to carry out the project.
 Plan the project.
Identification of Projects. Deciding which projects to undertake is usually the outgrowth of
the strategic and business planning of an organization. (See Section 13, Strategic Deployment for a
discussion of how specific projects are deployed from an organization’s vision, strategies, and goals.)
Typically, quality planning projects create new or updated products that are needed to reach specific
strategic goals, to meet new or changing customer needs, to fulfill legal or customer mandates, or
to take advantage of a new or emerging technology.
Upper management must take the leadership in identifying and supporting the critical quality
planning projects. Acting as a quality council or similar body, management needs to fulfill the following
key roles.
Setting Quality Goals. Top management identifies opportunities and needs to improve quality and
sets strategic goals for the organization.
Nominating and Selecting Projects. The quality council selects those major quality planning projects
critical to meeting strategic quality goals.
Selecting Teams. Once a project has been identified, the quality council appoints a team to see the
project through the remaining steps of the quality planning process.
Supporting Project Team. New techniques and processes are generally required to meet quality
goals. It is up to the quality council to see that each quality planning team is well prepared and
equipped to carry out its mission. The quality council’s support may include
 Providing education and training in quality planning tools and techniques
 Providing a trained facilitator to help the team work effectively and learn the quality planning
 Reviewing team progress
 Approving revision of the project mission
 Identifying/helping with any problems
 Coordinating related quality planning projects
 Helping with logistics, such as a meeting site
 Providing expertise in data analysis and survey design
 Furnishing resources for unusually demanding data collection
 Communicating project results
Monitoring Progress. The quality council is generally responsible for keeping the quality planning
process on track, evaluating progress, and making midcourse corrections to improve the effectiveness
of the entire process. Once the quality council has reviewed the sources for potential projects,
it will select one or more for immediate attention. Next, it must prepare a mission statement for the
Prepare Mission Statement. Once the quality council has identified the need for a project,
it should prepare a mission statement that incorporates the specific goal(s) of the project. The mission
statement is the written instruction for the team that describes the intent or purpose of the project. The
team mission describes
 The scope of the planning project, that is, the product and markets to be addressed
 The goals of the project, that is, the results to be achieved
Writing mission statements requires a firm understanding of the driving force behind the project.
The mission helps to answer the following questions:
 Why does the organization want to do the project?
 What will it accomplish once it is implemented?
A mission statement also fosters a consensus among those who either will be affected by the
project or will contribute the time and resources necessary to plan and implement the project goal.
 The team mission is to deliver to market a new low-energy, fluorocarbon-free refrigerator.
 The team will create accurate control and minimum cost for the inventory of all stores.
While these mission statements describe what will be done, they are still incomplete. They lack the
clarity and specificity that is required of a complete quality planning mission statement that incorporates
the goal(s) of a project. Well-written and effective mission statements define the scope of the
project by including one or more of the following.
Inherent performance: How the final product will perform on one or more dimensions, e.g., 24-
hour response time.
Comparative performance: How the final product will perform vis-?-vis the competition,
e.g., the fastest response time in the metropolitan area.
Customer reaction: How customers will rate the product compared with others available, e.g.,
one company is rated as having a better on-time delivery service compared with its closest rival.
Market: Who are or will be the customers or target audience for this product, and what share
of the market or market niche will it capture, e.g., to become the “preferred” source by all business
travelers within the continental United States.
Performance deficiencies: How will the product perform with respect to product failure, e.g.,
failure rate of less than 200 for every million hours of use.
Avoidance of unnecessary constraints: Avoid overspecifying the product for the team, e.g., if
the product is intended for airline carryon, specifying the precise dimensions in the mission may
be too restrictive. There may be several ways to meet the carryon market.
Basis for Establishing Quality Goals. In addition to the scope of the project, a mission
statement also must include the goal(s) of the project. An important consideration in establishing
quality goals is the choice of the basis for which the goal(s) are set.
Technology as a Basis. In many organizations, it has been the tradition to establish the quality
goals on a technological basis. Most of the goals are published in specifications and procedures that
define the quality targets for the supervisory and nonsupervisory levels.
The Market as a Basis. Quality goals that affect product salability should be based primarily on
meeting or exceeding market quality. Because the market and the competition undoubtedly will be
changing while the quality planning project is under way, goals should be set so as to meet or beat
the competition estimated to be prevailing when the project is completed. Some internal suppliers
are internal monopolies. Common examples include payroll preparation, facilities maintenance,
cafeteria service, and internal transportation. However, most internal monopolies have potential
competitors. There are outside suppliers who offer to sell the same service. Thus the performance of
the internal supplier can be compared with the proposals offered by an outside supplier.
Benchmarking as a Basis. “Benchmarking” is a recent label for the concept of setting goals
based on knowing what has been achieved by others. (See Section 12.) A common goal is the
requirement that the reliability of a new product be at least equal to that of the product it replaces
and at least equal to that of the most reliable competing product. Implicit in the use of benchmarking
is the concept that the resulting goals are attainable because they have already been
attained by others.
History as a Basis. A fourth and widely used basis for setting quality goals has been historical performance;
i.e., goals are based on past performance. Sometimes this is tightened up to stimulate
improvement. For some products and processes, the historical basis is an aid to needed stability. In
other cases, notably those involving chronically high costs of poor quality, the historical basis helps
to perpetuate a chronically wasteful performance. During the goal-setting process, the management
team should be on the alert for such misuse of the historical basis.
Quality Goals Are a Moving Target. It is widely recognized that quality goals must keep shifting
to respond to the changes that keep coming over the horizon: new technology, new competition,
threats, and opportunities. While organizations that have adopted quality management methods practice
this concept, they may not do as well on providing the means to evaluate the impact of those
changes and revise the goals accordingly.
Project Goals. Specific goals of the project, i.e., what the project team is to accomplish, are part
of an effective mission statement. In getting the job done, the team must mentally start at the finish.
The more focused it is on what the end result will look like, the easier it will be to achieve a successful
Measurement of the Goal. In addition to stating what will be done and by when, a project goal
must show how the team will measure whether or not it has achieved its stated goals. It is important
to spend some time defining how success is measured. Listed below are the four things that can be
1. Quality
2. Quantity
3. Cost
4. Time
An effective quality planning project goal must have five characteristics for it to provide a team
with enough information to guide the planning process. The goal must be
 Agreed to by those affected
 Realistic—It can be a stretch, but it must be plausible.
 Time specific—when it will be done
An example of a poorly written goal might look something like this: “To design a new car that is
best in class.” Contrast this with the following example: “To design, and put into production within
3 years, a new, midsized car that is best in class and priced, for the public, at under $20,000 (at time
of introduction). The design also should allow the company to sell the car and still have an average
return of between 4 and 6 percent.”
The second example is much more detailed, measurable, and time-specific compared with the
first. The target or end result is clearly stated and provides enough direction for the team to plan the
product features and processes to achieve the goal.
The Ford Mustang—Mission and Goals. Before moving ahead with any product development,
Ford agreed to a clear mission for the Mustang. The short version was “The Car The Star.” Whenever
a large group of people from various functional organizations is brought together to work on a project,
there is a natural tendency to bring a lot of their “home office priorities and objectives” to the team.
Unattended, these home office priorities can diffuse the team’s focus and create a significant amount
of conflict within the team.
To address this issue, the team chose the statement “The Car is the Star” to align the efforts of all
team members and to focus them on a single objective. This statement was a simple and effective
way to galvanize the team around the fact that there was one common purpose and a single superordinate
Specifically, this meant that all team members could adjudicate their daily decisions and actions
consistent with the overall team objective of making the car a reality and success. Program goals
were established for 18 separate parameters. These “18 panel charts” enabled the project to focus on
very specific success factors. (See Figure 3.3.)
Topic Description
Panel 1: Quality Things gone wrong/1000, things gone right/1000, repairs/1000
and customer satisfaction @ 3 months in service, plus things
gone wrong/1000 @ 4 years in service.
Panel 2: Timing Summary of the major milestones from the total program workplan.
Panel 3: Vehicle hardpoints Summary of the architectural hardpoints such as wheelbase,
tread, length, height, width, interior room, leg room, etc.
Panel 4: Vehicle dynamics Subjective targets for performance feel, ride, handling,
noise/vibration/harshness, brake performance, seat performance,
Panel 5: Weight Curb weight and emission test weight for all models.
Panel 6: Fuel economy Metro-highway fuel consumption is declared for all models.
Avoidance of gas guzzler is also declared.
Panel 7: Performance 0–60 mph elapsed time is declared for all models.
Panel 8: Complexity Number of discrete customer decisions and buildable combinations
is declared.
Panel 9: Serviceability Projection for the number of total service hours through 50,000
miles is declared.
Panel 10: Damageability Projected repair cost for a typical collision is declared.
Panel 11: Safety emissions Compliance with all applicable Federal Motor Vehicle Safety
Standards and Federal Clean Air Standards is declared.
Panel 12: Variable cost Variable cost versus prior model is declared.
Panel 13: Program investment Total investment for tooling, facilities, launch, and engineering is
Panel 14: Pricing Wholesale delivery price is declared for all models.
Panel 15: Volumes Five year projection for trend volume is declared.
Panel 16: Profitability Program profitability is declared in terms of fully accounted profits.
Panel 17: Features All product standard features and option by model are declared.
Panel 18: Export Export markets and volumes are declared.
FIGURE 3.3 Ford Mustang 18 panel chart goals.
New Product Policies. Companies need to have very clear policy guidance with respect to quality
and product development. Most of these should relate to all new products, but specific policies may
relate to individual products, product lines, or groups. Four of the most critical policies are as follows.
Deficiencies in New and Carryover Designs. Many organizations have established the clear policy
that no new product or component of a product will have a higher rate of deficiencies than the old
product or component that it is replacing. In addition, they often require that any carryover design
must have a certain level of performance; otherwise, it must be replaced with a more reliable design.
The minimum carryover reliability may be set by one or more of the following criteria: (1) competitor
or benchmark reliability, (2) customer requirements, or (3) a stretch goal beyond benchmark
or customer requirements.
Intended versus Unintended Use. Should stepladders be designed so that the user can stand on the
top step without damage, even though the step is clearly labeled “Do Not Step Here?” Should a hospital
design its emergency room to handle volumes of routine, nonemergency patients who show up
at its doors? These are policy questions that need to be settled before the project begins. The answers
can have a significant impact on the final product, and the answers need to be developed with reference
to the organization’s strategy and the environment within which its products are used.
Requirement of Formal Quality Planning Process. A structured, formal process is required to
ensure that the product planners identify their customers and design products and processes that will
meet those customer needs with minimum deficiencies. Structured formality is sometimes eschewed
as a barrier to creativity. Nothing could be more misguided. Formal quality planning identifies the
points at which creativity is demanded and then encourages, supports, and enables that creativity.
Formal planning also ensures that the creativity is focused on the customers and that creative designs
ultimately are delivered to the customer free of the destructive influences of deficiencies.
Custody of Designs and Change Control. Specific provision must be made to ensure that
approved designs are documented and accessible. Any changes to designs must be validated, receive
appropriate approvals, be documented, and be unerringly incorporated into the product or process.
Specific individuals must have the assigned authority, responsibility, and resources to maintain the
final designs and administer change control.
Ford Policies with Respect to the Mustang. Ford had three specific policies with respect to carryover
and new designs. New designs were required to be more reliable than the old. They also were
required to provide demonstrated cost-benefit contributions to the final product. Finally, major features
were expected to exceed the performance of the chief competitor—Camaro/Firebird.
Because Mustang needed to maintain its reputation as a reliable performance car, a more stringent
testing policy was established. In addition to the safety, economy, reliability, durability, and
other tests that all Ford cars must pass, the Mustang was required to pass tests such as 500 drag starts
without a failure and repeated hard braking without fading.
Establish Team. The cross-functional approach to quality planning is effective for several
Team involvement promotes sharing of ideas, experiences, and a sense of commitment to being
a part of and helping “our” organization achieve its goal.
The diversity of team members brings a more complete working knowledge of the product and
processes to be planned. Planning a product requires a thorough understanding of how things get
done in many parts of the organization.
Representation from various departments or functions promotes the acceptance and implementation
of the new plan throughout the organization. Products or processes designed with the active
participation of the affected areas tend to be technically superior and accepted more readily by
those who must implement them.
Guidelines for Team Selection. When selecting a team, the quality council identifies those parts
of the organization which have a stake in the outcome. There are several places to look:
 Those who will be most affected by the result of the project
 Departments or functions responsible for various steps in the process
 Those with special knowledge, information, or skill in the design of the project
 Areas that can be helpful in implementing the plan
The Mustang Team. Ford established a dedicated team of individuals from all key parts of the
company, including
Stakeholder organizations Role
Vehicle Engineering Team leader
Product and Business Planning Prepare the product assumptions and business
Body and Chassis Engineering Define the product upgrades to satisfy the
customer for comfort, convenience, and
safety and to meet the forecasted federal
motor vehicle safety standards.
Power Train Engineering Define the power train upgrades to satisfy the
customer wants for power and drivability
and meet the clean air and gas guzzler
Vehicle Design Provide excitement with a highly styled
Manufacturing and Assembly Ensure that there is a feasible, high-quality
Sales and Marketing Provide the “voice of the customer” and ensure
that the product hits the target market.
Purchasing Bring the suppliers to the party.
Finance Help the team develop a financially attractive
business package.
Besides their professional expertise for the job, many of the team members also were Mustang
enthusiasts who brought an understanding and passion for the customer needs that would be hard to
Because an automobile is a highly complex consumer product, the work was divided into five
chunk teams: body exterior and structure, body interior and electrical, power train, chassis, and vehicle.
Each chunk team had members from all the major stakeholders and was able to manage its portion
of the program with autonomy within the overall design and 18 panel chart parameters.
The overall coordination and direction of the work among the chunk teams was managed with a
weekly “major matters” meeting among the chunk team leaders and the program manager. These
meetings focused on the major program metrics incorporated in the 18 panel charts.
This step may seem unnecessary; of course, the planners and designers know who their customers
are: the driver of the automobile, the depositor in the bank account, the patient who takes the
medication. But these are not the only customers—not even necessarily the most important customers.
Customers comprise an entire cast of characters that needs to be understood fully.
Generally, there are two primary groups of customers: the external customers—those outside the
producing organization; and the internal customers—those inside the producing organization.
Types of External Customers. The term “customer” is often used loosely; it can refer to an
entire organization, a unit of a larger organization, or a person. There are many types of customers—
some obvious, others hidden. Below is a listing of the major categories to help guide complete customer
The purchaser Someone who buys the product for himself or
herself or for someone else, e.g., anyone
who purchases food for his or her family.
The end user/ultimate customer Someone who finally benefits from the product,
e.g., the patient who goes to health care
facility for diagnostic testing.
Merchants People who purchase products for resale,
wholesalers, distributors, travel agents and
brokers, and anyone who handles the product,
such as a supermarket employee who
places the product on the shelf.
Processors Organizations and people who use the product
or output as an input for producing their
own product, e.g., a refinery that receives
crude oil and processes it into different
products for a variety of customers.
Suppliers Those who provide input to the process, e.g.,
the manufacturer of the spark plugs for an
automobile or the law firm that provides
advice on the company’s environmental law
matters. Suppliers are also customers. They
have information needs with respect to
product specification, feedback on deficiencies,
predictability of orders, and so on.
Original equipment manufacturers (OEMs) Purchasers of a product to incorporate into
their own, e.g., a computer manufacturer
using another producer’s disk drives for its
Potential customers Those not currently using the product but
capable of becoming customers; e.g., a business
traveler renting a car may purchase a
similar automobile when the time comes to
buy one for personal use.
Hidden customers An assortment of different customers who are
easily overlooked because they may not
come to mind readily. They can exert great
influence over the product design: regulators,
critics, opinion leaders, testing services,
payers, the media, the public at large,
those directly or potentially threatened by
the product, corporate policymakers, labor
unions, professional associations.
Internal Customers. Everyone inside an organization plays three roles: supplier, processor,
and customer. Each individual receives something from someone, does something with it, and passes
it to a third individual. Effectiveness in meeting the needs of these internal customers can have a
major impact on serving the external customers. Identifying the internal customers will require some
analysis because many of these relationships tend to be informal, resulting in a hazy perception of
who the customers are and how they will be affected. For example, if a company decides to introduce
just-in-time manufacturing to one of its plants, this will have significant effect on purchasing,
shipping, sales, operations, and so on.
Most organizations try to set up a mechanism that will allow seemingly competing functions
to negotiate and resolve differences based on the higher goal of satisfying customer needs. This
might include conducting weekly meetings of department heads or publishing procedure manuals.
However, these mechanisms often do not work because the needs of internal customers are not
fully understood, and communication among the functions breaks down. This is why a major goal
in the quality planning process is to identify who the internal customers are, discover their needs,
and plan how those needs will be satisfied. This is also another reason to have a multifunctional
team involved in the planning; these are people who are likely to recognize the vested interests of
internal customers.
Identifying Customers. In addition to the general guidance just laid out, it is most often helpful
to draw a relatively high-level flow diagram of the processes related to the product being planned.
Careful analysis of this flow diagram often will provide new insight, identifying customers that
might have been missed and refining understanding of how the customers interact with the process.
Figure 3.4 is an example of such a diagram. A review of this diagram reveals that the role of “customer”
is really two different roles—placing the order and using the product. These may or may not
be played by the same individuals, but they are two distinct roles, and each needs to be understood
in terms of its needs.
Customers for Ford Mustang. The most critical customer for Ford Mustang was the “ultimate customer,”
namely, the person who would purchase the vehicle. The prime demographic target group
was the Mustang GT buyer. (The GT model is the image setter and the model that must appeal to the
enthusiast if the car line is to be successful.)
Other vital few external customers included
 Environmental Protection Agency (EPA)
 The National Highway and Traffic Safety Administration (NHTSA)
 The Federal Trade Commission (FTC)
Media (refers to both the internal and external sources)
 Internal media
Ford Communication Network (FCN)
Employee publications
 External media
Enthusiast magazines such as Motor Trend Magazine, Car and Driver, Road & Track,
Automobile Magazine, AutoWeek, etc.
Trade magazines such as Ward’s AutoWorld, Automotive News, Automotive Industries,
Automotive Engineering.
National television programs such as “Motorweek,” “MotorTrend.”
Local and regional television programs
Local and regional radio programs
Local and regional newspaper coverage
Internal Customers
Corporate management
Product Engineering Office
Assembly plant
Ford Customer Service Division
Public affairs
Internal media
 Ford Communication Network (FCN)
 Employee publications
FIGURE 3.4 The flow diagram and customers. [From Juran, J. M. (1988), Quality Control
Handbook, 4th ed. McGraw-Hill, New York, p. 6.6.]
The third step of quality planning is to discover the needs of both internal and external customers for
the product. Some of the key activities required for effective discovery of customer needs include
 Plan to collect customers’ needs.
 Collect a list of customers’ needs in their language.
 Analyze and prioritize customers’ needs.
 Translate their needs into “our” language.
 Establish units of measurement and sensors.
Our own experience tells us that the needs of human beings are both varied and complex. This
can be particularly challenging to a planning team because the actions of customers are not always
consistent with what they say they want. The challenge for quality planning is to identify the most
important needs from the full array of those needs expressed or assumed by the customer. Only then
can the product delight the customers.
When designing a product, there are actually two related but distinct aspects of what is being
developed: the technology elements of what the product’s features will actually do or how it will
function and the human elements of the benefits customers will receive from using the product. The
two must be considered together.
Discovering customer needs is a complex task. Experience shows that customers usually do not
state, in simple terms, exactly what they want; often they do not even mention some of their most
basic needs. Accuracy of bank statements, competence of a physician, reliability of a computer, and
grammatical correctness of a publication may be assumed and never stated without probing.
One of the ways customers express their needs is in terms of problems they experience and their
expectation that a product will solve their problems. For example, a customer may state, “I cannot
always answer my telephone personally, but I do not want callers to be either inconvenienced or disgusted
with nonresponsive answering systems,” or, “ My mother’s personal dignity and love of people
are very important to me. I want to find an extended care facility that treats her like a person, not
a patient.” Even when the need is not expressed in such terms, the art and science of discovering
needs are to understand exactly the benefit that the customer expects.
When a product’s features meet a customer’s need, it gives the customer a feeling of satisfaction.
If it fails to deliver the promised feature defect-free, the customer feels dissatisfaction. Even if a
product functions the way it has been designed, a competing product, by virtue of superior service
or performance, may provide customers with greater satisfaction.
Stated Needs and Real Needs. Customers commonly state their needs as seen from their
viewpoint and in their language. Customers may state their needs in terms of the goods or services
they wish to buy. However, their real needs are the benefits they believe they will receive.
To illustrate:
Customer wishes to buy… Benefit customer needs might include…
Fresh pasta Nourishment and taste
Newest personal computer Write reports quickly and easily
Find information on the Web
Help children learn math
Health insurance Security against financial disaster
Access to high quality health care
Choice in health care providers
Airline ticket Transportation, Comfort, Safety, and Convenience
Failure to grasp the difference between stated needs and real needs can undermine a quality planning
project. Understanding the real needs does not mean that the planners can dismiss the customers’
statements and substitute their own superior technical understanding as being the customers’
real needs. Understanding the real needs means asking and answering such questions as
Why is the customer buying this product?
What service does he or she expect from it?
How will the customer benefit from it?
How does the customer use it?
What has created customer complaints in the past?
Why have customers selected competitor products over ours?
Perceived Needs. Customers understandably state their needs based on their perceptions.
These may differ entirely from the supplier’s perceptions of what constitutes product quality.
Planners can mislead themselves by considering whether the customers’ perceptions are wrong or
right rather than focusing on how these perceptions influence their buying habits. While such differences
between customers and suppliers are potential troublemakers, they also can be an opportunity.
Superior understanding of customer perceptions can lead to competitive advantage.
Cultural Needs. The needs of customers, especially internal customers, go beyond products
and processes. They include primary needs for job security, self-respect, respect of others, continuity
of habit patterns, and still other elements of what we broadly call the “cultural values”; these are seldom
stated openly. Any proposed change becomes a threat to these important values and hence will
be resisted until the nature of the threat is understood.
Needs Traceable to Unintended Use. Many quality failures arise because a customer
uses the product in a manner different from that intended by the supplier. This practice takes
many forms. Patients visit emergency rooms for nonemergency care. Untrained workers are
assigned to processes requiring trained workers. Equipment does nor receive specified preventive
Factors such as safety may add to the cost, yet they may well result in a reduced overall cost by
helping to avoid the higher cost arising from misuse of the product. What is essential is to learn the
What will be the actual use (and misuse)?
What are the associated costs?
What are the consequences of adhering only to intended use?
Human Safety. Technology places dangerous products into the hands of amateurs who do not
always possess the requisite skills to handle them without accidents. It also creates dangerous
byproducts that threaten human health, safety, and the environment. The extent of all this is so great
that much of the effort of product and process planning must be directed at reducing these risks to
an acceptable level. Numerous laws, criminal and civil, mandate such efforts.
“User Friendly.” The amateur status of many users has given rise to the term “user friendly”
to describe that product feature that enables amateurs to make ready use of technological products.
For example, the language of published information should be
Readily understood (Notorious offenders have included legal documents, owners’ operating manuals,
administrative forms, etc. Widely used forms such as governmental tax returns should be
field tested on a sample of the very people who will later be faced with filling out such forms.)
Broadly compatible (For example, new releases of software should be “upward compatible with
earlier releases.”)
Promptness of Service. Services should be prompt. In our culture, a major element of competition
is promptness of service. Interlocking schedules (as in mail delivery or airline travel) are
another source of a growing demand for promptness. Still another example is the growing use of justin-
time manufacturing, which requires dependable deliveries of materials to minimize inventories.
All such examples demonstrate the need to include the element of promptness in planning to meet
customer needs.
Customer Needs Related to Deficiencies. In the event of product failure, a new set of
customer needs emerges—how to get service restored, and how to get compensated for the associated
losses and inconvenience. Clearly, the ideal solution to all this is to plan quality so that
there will be no failures. At this point, we will look at what customers need when failures do
Warranties. The laws governing sales imply that there are certain warranties given by the supplier.
However, in our complex society, it has become necessary to provide specific, written contracts
to define just what is covered by the warranty and for how long a time. In addition, it should be clear
who has what responsibilities.
Effect of Complaint Handling on Sales. While complaints deal primarily with product
dissatisfaction, there is a side effect on salability. Research in this area has pointed out the
following: Of the customers who were dissatisfied with products, nearly 70 percent did not
complain. The proportions of these who did complain varied depending on the type of product
involved. The reasons for not complaining were principally (1) the effort to complain was not
worth it, (2) the belief that complaining would do no good, and (3) lack of knowledge about
how to complain. More than 40 percent of the complaining customers were unhappy with the
responsive action taken by the suppliers. Again, percentages varied depending on the type of
Future salability is strongly influenced by the action taken on complaints. This strong influence
also extends to brand loyalty. Even customers of popular brands of “large ticket” items, such as
durable goods, financial services, and automobile services, will reduce their intent to buy when they
perceive that their complaints are not addressed.
This same research concluded that an organized approach to complaint handling provides a high
return on investment. The elements of such an organized approach may include
 A response center staffed to provide 24-hour access by consumers and/or a toll-free telephone
 Special training for the employees who answer the telephones
 Active solicitation of complaints to minimize loss of customers in the future
Keeping Customers Informed. Customers are quite sensitive to being victimized by secret
actions of a supplier, as the phrase “Let the buyer beware!” implies. When such secrets are later discovered
and publicized, the damage to the supplier’s quality image can be considerable. In a great
many cases, the products are fit for use despite some nonconformances. In other cases, the matter may
be debatable. In still other cases, the act of shipment is at the least unethical and at the worst illegal.
Customers also have a need to be kept informed in many cases involving product failures.
There are many situations in which an interruption in service will force customers to wait for an
indefinite period until service is restored. Obvious examples are power outages and delays in public
transportation. In all such cases, the customers become restive. They are unable to solve the
problem—they must leave that to the supplier. Yet they want to be kept informed as to the nature
of the problem and especially as to the likely time of solution. Many suppliers are derelict in keeping
customers informed and thereby suffer a decline in their quality image. In contrast, some airlines
go to great pains to keep their customers informed of the reasons for a delay and of the
progress being made in providing a remedy.
Plan to Collect Customers’ Needs. Customer needs keep changing. There is no such thing
as a final list of customer needs. While it can be frustrating, planning teams must realize that even
while they are in the middle of the planning process, forces such as technology, competition, social
change, and so on, can create new customer needs or may change the priority given to existing needs.
It becomes extremely important to check with customers frequently and monitor the marketplace.
Some of the most common ways to collect customer needs include
 Customer surveys, focus groups, and market research programs and studies
 Routine communications, such as sales and service calls and reports, management reviews, house
 Tracking customer complaints, incident reports, letters, and telephone contacts
 Simulated-use experiments and planning processes that involve the customer
 Employees with special knowledge of the customer: sales, service, clerical, secretarial, and supervisory
who come into contact with customers
 Customer meetings
 User conferences for the end user
 Information on competitors’ products
 Personal visits to customer locations; observe and discuss
How product is used
Unintended uses
Service failures by others
What current or new features will relieve onerous tasks
Changes in habits and culture
Changes in sales
Existence of or changes in premium prices
Sale of spare parts or after-sale service
Purchase of options
 Government or independent laboratory data
 Changes in federal, state, and local regulations that will identify current need or new opportunity
 Competitive analysis and field intelligence comparing products with those of competitors
 Personal experience dealing with the customer and the product (However, it is important to be cautious
about giving personal experience too much weight without direct verification by customers.
The analysts must remember that looking at customer needs and requirements from a personal
viewpoint can be a trap.)
Often customers do not express their needs in terms of the benefits they wish to receive from purchasing
and using the product.
Discovering Mustang Customer Needs. In designing the Mustang, Ford relied heavily
on the following sources for customer needs:
 Quantitative market research
 Qualitative market research
 Inspection studies
 Observational research
 Dealer input
 Customer surveys
 Direct interaction with customers
 Media feedback
 Product evaluation reports—internal and external
 Competitive evaluations
Collect List of Customers’ Needs in Their Language. For a list of customers’ needs
to have significant meaning in planning a new product, they must be stated in terms of benefits
sought. Another way of saying this is to capture needs in the customer’s voice. By focusing on the
benefits sought by the customer rather than on the means of delivering the benefit, designers will
gain a better understanding of what the customer needs and how the customer will be using the product.
Stating needs in terms of the benefits sought also can reveal opportunities for improved quality
that often cannot be seen when concentrating on the product features alone.
Analyze and Prioritize Customer Needs. The information actually collected from customers
is often too broad, too vague, and too voluminous to be used directly in designing a product.
Both specificity and priority are needed to ensure that the design really meets the needs and that time
is spent on designing for those needs which are truly the most important. The following activities
help provide this precision and focus:
 Organizing, consolidating, and prioritizing the list of needs for both internal and external customers
 Determining the importance of each need for both internal and external customers
 Breaking down each need into precise terms so that a specific design response can be identified
 Translating these needs into the supplying organization’s language
 Establishing specific measurements and measurement methods for each need
One of the best planning tools to analyze and organize customers’ needs is the “quality planning
Quality Planning Spreadsheets. Quality planning generates a large amount of information that is
both useful and necessary, but without a systematic way to approach the organization and analysis
of this information, the planning team may be overwhelmed by the volume and miss the message it
Although planners have developed various approaches for organizing all this information, the
most convenient and basic planning tool is the quality planning spreadsheet. The spreadsheet is a
highly versatile tool that can be adapted to a number of situations. The quality planning process
makes use of several kinds of spreadsheets, such as
 Customer needs spreadsheet
 Needs analysis spreadsheet
 Product design spreadsheet
 Process design spreadsheet
 Process control spreadsheet
Besides recording information, these tools are particularly useful in analyzing relationships
among the data that have been collected and in facilitating the stepwise conversion of customer needs
into product features and then product features into process characteristics and plans. This conversion
is illustrated in Figure 3.5. Analysis of customers and their needs provides the basis for designing
the product. The summary of that design feeds the process design, which feeds the control
For most planning projects, simple matrix spreadsheets will suffice. For other projects, more
complex quality functional deployment spreadsheets are helpful in computing design trade-offs. All
these spreadsheets are designed to allow the team to record and compare the relationships among
many variables at the same time. We will illustrate some of these spreadsheets at the appropriate
point in the planning process. Figure 3.6 illustrates the generic layout of any one of these spreadsheets.
In general, the row headings are the “what’s” of the analysis—the customers to be satisfied,
the needs to be met, and so on. The columns are the “how’s”—the needs that, when met, will satisfy
the customer, the product features that will meet the needs, and so on. The bottom row of the
spreadsheet generally contains specific measurable goals for the “how” at the top. The body of the
spreadsheet expresses with symbols or numerics the impact of the “how” on the “what”—e.g., none,
moderate, strong, very strong. Other columns can be added to give specific measures of the importance
of the respective rows, benchmarks, and so on.
Customer Needs Spreadsheet. Figure 3.7 provides a simple example of a customer needs
spreadsheet. The left column lists, in priority order, all the external and internal customers. The
column headings are the various needs that have been discovered. By either checking or entering
a designation for importance, it is possible to create a simple but comprehensive picture of the
importance of meeting each need. All product development must operate within a budget.
Prioritizing the customers and their needs ensures that the budget is focused on what is most
Precise Customer Needs. Once the needs that must be met have been prioritized, they must be
described in sufficiently precise terms to design a product based on them. A customer needs spreadsheet
helps assemble this analysis. At this point, customer needs are probably a mixture of relatively
broad expectations such as “ease of use” and more specific requests such as “access on Saturday.”
Figure 3.8 illustrates how broad needs (called “primary”) are broken into succeeding levels of specificity
(“secondary,” “tertiary,” etc.) Note that primary and secondary do not mean more and less
important, they mean, respectively, less specific and more specific. Each need must be broken down
to the level at which it can (1) be measured and (2) serve as an unambiguous guide for product
design. In some cases two levels of detail may suffice, in others four or five may be required. Figure
3.8 illustrates how this might be done for the primary need “convenience” associated with a group
medical practice.
Mustang Customer Needs. In addition to basic transportation, the following needs surfaced as
most important for Mustang customers:
FIGURE 3.5 Spreadsheets in quality planning. (Juran Institute, Inc. Copyright 1994. Used by permission.)
FIGURE 3.6 Generic planning spreadsheet. (Juran Institute, Inc.
Copyright 1994. Used by permission.)
How Well
How Much
Customer Needs
Color separators
Catchy cover lines
Stable circulation
It sellls
Enough time
Material complete
No last minute changes
Very Strong
FIGURE 3.7 Customer needs spreadsheet for a magazine. (Juran Institute, Inc. Copyright 1994. Used
by permission.)
Primary need Secondary need Tertiary need
Convenience Hours of operation Open between 5:00 and 9:00 p.m.
Saturday hours
Transportation access Within three blocks of bus stop
Ample parking
Short wait Urgent appointment within 24 hours
Routine appointment within 14 days
Waiting time at appointment less than 15 minutes
Complementary available Pharmacy on site
services Lab on site
Low cost of ownership
Fuel economy
Handling—in all types of weather
Ride quality—regardless of road surface
Translate Their Needs into “Our” Language. The precise customer needs that have been identified
may be stated in any of several languages, including
The customer’s language
The supplier’s (“our”) language
A common language
An old aphorism claims that the British and Americans are separated by a common language. The
appearance of a common language or dialect can be an invitation to trouble because both parties
believe that they understand each other and expect to be understood. Failure to communicate because
of the unrecognized differences can build additional misunderstanding that only compounds the difficulty.
It is imperative, therefore, for planners to take extraordinary steps to ensure that they properly
understand customer needs by systematically translating them. The need to translate applies to
both internal and external customers. Various company functions employ local dialects that are often
not understood by other functions.
Vague terminology constitutes one special case for translation that can arise even (and often especially)
between customers and suppliers that believe they are speaking the same dialect. Identical
words have multiple meanings. Descriptive words do not describe with technological precision.
Aids to Translation. Numerous aids are available to clear up vagueness and create a bridge across
languages and dialects. The most usual are listed following:
A glossary is a list of terms and their definitions. It is a published agreement on the precise meanings
of key terms. The publication may be embellished by other forms of communication, such
as sketches, photographs, and videotapes.
FIGURE 3.8 Needs analysis spreadsheet for a medical office. (Juran Institute, Inc. Copyright 1994. Used
by permission.)
Samples can take many forms, such as physical goods (e.g., textile swatches, color chips,
audio cassettes) or services (e.g., video recordings to demonstrate “samples” of good service—
courtesy, thoughtfulness, etc.) They serve as specifications for product features. They make
use of human senses beyond those associated with word images.
A special organization to translate communications with external customers may be required
because of the high volume of translation. A common example is the order-editing department,
which receives orders from clients. Some elements of these orders are in client language. Order
editing translates these elements into supplier language, e.g., product code numbers, supplier
acronyms, and so on.
Standardization is used by many mature industries for the mutual benefit of customers and suppliers.
This standardization extends to language, products, processes, and so on. All organizations make
use of short designations for their products, such as code numbers, acronyms, words, phrases, and so
on. Such standardized nomenclature makes it easy to communicate with internal customers.
Measurement is the most effective remedy for vagueness and multiple dialects—“Say it in numbers.”
This is the first, but not the last, point in the planning process where measurement is critical.
Quality planning also requires measurement of product features, process features, process
capability, control subjects, and so on.
Establish Units of Measurement and Sensors. Sound quality planning requires precise
communication between customers and suppliers. Some of the essential information can be conveyed
adequately by words. However, an increasingly complex and specialized society demands
higher precision for communicating quality-related information. The higher precision is best attained
when we say it in numbers.
Quantification requires a system of measurement. Such a system consists of
A unit of measurement, which is a defined amount of some quality feature, permits evaluation of
that feature in numbers, e.g., hours of time to provide service, kilowatts of electric power, or concentration
of a medication.
A sensor, which is a method or instrument of measurement, carries out the evaluation and states
the findings in numbers in terms of the unit of measure, e.g., a clock for telling time, a thermometer
for measuring temperature, or an x-ray to measure bone density.
By measuring customer needs, one has established an objective criterion for whether or not the
needs are met. In addition, only with measurement can one answer questions such as, Is our quality
getting better or worse? Are we competitive with others? Which one of our operations provides the
best quality? How can we bring all operations up to the level of the best?
Units of Measure for Product Features. The first task in measurement is to identify the appropriate
unit of measurement for each customer need. For product features, we know of no simple, convenient,
generic formula that is the source of many units of measure. The number and variety of
product features are simply enormous. In practice, each product feature requires its own unique unit
of measure. A good starting point is to ask the customers what their units of measure are for evaluating
product quality. If the supplier’s units of measure are different, the stage is set for customer dissatisfaction,
and the team will need to come up with a unit of measure acceptable to both parties.
Even if the customers have not developed an explicit unit of measure, ask them how they would
know whether the need were met. Their response may carry with it an implicit unit of measure.
Application to Goods. Units of measure for quality features of goods make extensive use of “hard”
technological units. Some of these are well known to the public: time in minutes, temperature in
degrees, or electric current in amperes. Many others are known only to the specialists.
There are also “soft” areas of quality for goods. Food technologists need units of measure for flavor,
tenderness, and still other properties of food. Household appliances must be “handsome” in
appearance. Packaging must be “attractive.” To develop units of measure for such features involves
much effort and ingenuity.
Application to Services. Evaluation of service quality includes some technological units of measure.
A widespread example is promptness, which is measured in days, hours, and so on.
Environmental pollutants (e.g., noise, radiation, etc.) generated by service companies are likewise
measured using technological units of measure.
Service quality also involves features such as courtesy of service personnel, decor of surroundings,
and readability of reports. Since these features are judged by human beings, the units of measure
(and the associated sensors) must be shown to correlate with a jury of customer opinion.
The Ideal Unit of Measure. The criteria for an ideal unit of measure are summarized below. An
ideal unit of measure
 Is understandable
 Provides an agreed basis for decision making
 Is conducive to uniform interpretation
 Is economical to apply
 Is compatible with existing designs of sensors, if other criteria also can be met
Measuring Abstractions. Some quality features seem to stand apart from the world of physical
things. Quality of service often includes courtesy as a significant quality feature. Even in the case of
physical goods, we have quality features, such as beauty, taste, aroma, feel, or sound. The challenge
is to establish units of measure for such abstractions.
The approach to dealing with abstractions is to break them up into identifiable pieces. Once
again, the customer may be the best source to start identifying these components. For example, hotel
room appearance is certainly a quality feature, but it also seems like an abstraction. However, we can
divide the feature into observable parts and identify those specifics which collectively constitute
“appearance,” e.g., the absence of spots or bare patches on the carpet, clean lavatory, linens free from
discoloration and folded to specified sizes, windows free of streaks, bedspreads free of wrinkles and
hanging to within specific distances from the floor, and so on. Once units of measure have been
established for each piece or component, they should be summarized into an index, e.g., number of
soiled or damaged carpets to total number of hotel rooms, number of rooms with missing linens to
total number of rooms, or number of customer complaints.
Establish the Sensor. To say it in numbers, we need not only a unit of measure, but we also need to
evaluate quality in terms of that unit of measure. A key element in making the evaluation is the sensor.
A “sensor” is a specialized detecting device or measurement tool. It is designed to recognize the
presence and intensity of certain phenomena and to convert this sense knowledge into information.
In turn, the resulting information becomes an input to decision making because it enables us to evaluate
actual performance.
Technological instruments are obviously sensors. So are the senses of human beings. Trends in
some data series are used as sensors. Shewhart control charts are sensors.
Precision and Accuracy of Sensors. The “precision” of a sensor is a measure of the ability of the
sensor to reproduce its results over and over on repeated tests. For most technological sensors, this
reproducibility is high and is also easy to quantify.
At the other end of the spectrum are the cases in which we use human beings as sensors:
inspectors, auditors, supervisors, and appraisers. Human sensors are notoriously less precise than
technological sensors. Such being the case, planners are well advised to understand the limitations
inherent in human sensing before making decisions based on the resulting data.
The “accuracy” of a sensor is the degree to which the sensor tells the truth—the extent to which
its evaluation of some phenomenon agrees with the “true” value as judged by an established standard.
The difference between the observed evaluation and the true value is the “error,” which can be
positive or negative.
For technological sensors, it is usually easy to adjust for accuracy by recalibrating. A simple
example is a clock or watch. The owner can listen to the time signals provided over the radio. In contrast,
the precision of a sensor is not easy to adjust. The upper limit of precision is usually inherent
in the basic design of the sensor. To improve precision beyond its upper limit requires a redesign.
The sensor may be operating at a level of precision below that of its capability owing to misuse, inadequate
maintenance, and so on. For this reason, when choosing the appropriate sensor for each need,
planners will want to consider building in appropriate maintenance schedules along with checklists
on actions to be taken during the check.
Translating and Measuring Mustang Customer Needs. The customer need for performance
illustrates how high-level needs breakdown into a myriad of detailed needs. Performance
included all the following detailed, precise needs:
Performance feel off the line
Wide-open throttle (WOT) 0 to 60 mi/h elapsed time
WOT 14-mile elapsed time
WOT 40 to 60 mi/h passing time
WOT 30 to 70 mi/h passing time
Part-throttle response
Seat-of-the-pants feel that can only be measured by a jury of customers
Competitor performance was used as a minimum benchmark, but Ford knew that its competitors
also were working on new models and had to stretch the needs analysis to include “what-if” scenarios
that were tested with panels of consumers and automotive experts.
Product Design Spreadsheet. All the information on the translation and measurement of a customer
need must be recorded and organized. Experience recommends placing these data so that
they will be close at hand during product design. The example in Figure 3.9 shows a few needs all
prepared for use in product design. The needs, their translation, and their measurement are all
placed to the left of the spreadsheet. The remainder of the spreadsheet will be discussed in the next
Once the customers and their needs are fully understood, we are ready to design the product that will
meet those needs best. Product development is not a new function for a company. Most companies
have some process for designing and bringing new products to market. In this step of the quality
planning process, we will focus on the role of quality in product development and how that role combines
with the technical aspects of development and design appropriate for a particular industry.
Within product development, product design is a creative process based largely on technological or
functional expertise.
The designers of products traditionally have been engineers, systems analysts, operating managers,
and many other professionals. In the quality arena, designers can include any whose experience,
position, and expertise can contribute to the design process. The outputs of product design are
detailed designs, drawings, models, procedures, specifications, and so on.
The overall quality objectives for this step are two:
1. Determine which product features and goals will provide the optimal benefit for the customer
2. Identify what is needed so that the designs can be delivered without deficiencies.
Very Strong
No double
All appointments
Pt. comes
All info.
easy to find
No “holds” used
Pt. followed
MD’s instructions
Do not have to
Units of
Product Features
Product Feature Goals
Review by
Cross resource checking
Auto search for open times
Check resource constraints
FAX information to
scheduling source
Mail instructions to patient
100% of time for all
information entered
One key stroke
Cannot change
appt. w/o
author from source
Reminder always
generated for receiver
For all appointments
Review by
Review by person
doing procedure
Review by
by scheduler
FIGURE 3.9 Product design spreadsheet for outpatient appointment function. (Juran Institute, Inc. Copyright 1994.
Used by permission.)
In the case of designing services, the scope of this activity is sometimes puzzling. For example,
in delivering health care, where does the product of diagnosing and treating end and the processes
of laboratory testing, chart reviews, and so on begin? One useful way to think about the distinction
is that the product is the “face to the customer.” It is what the customer sees and experiences. The
patient sees and experiences the physician interaction, waiting time, clarity of information, and so
on. The effectiveness and efficiency of moving blood samples to and around the laboratory have an
effect on these product features but are really features of the process that delivers the ultimate product
to the customer.
Those who are designing physical products also can benefit from thinking about the scope of
product design. Remembering that the customer’s needs are the benefits that the customer wants
from the product, the design of a piece of consumer electronics includes not only the contents of the
box itself but also the instructions for installation and use and the “help line” for assistance.
There are six major activities in this step:
• Group together related customer needs.
• Determine methods for identifying product features.
• Select high-level product features and goals.
• Develop detailed product features and goals.
• Optimize product features and goals.
• Set and publish final product design.
Group Together Related Customer Needs. Most quality planning projects will be
confronted with a large number of customer needs. Based on the data developed in the preceding
steps, the team can prioritize and group together those needs which relate to similar functionality.
This activity does not require much time, but it can save a lot of time later. Prioritization ensures
that the scarce resources of product development are spent most effectively on those items which
are most important to the customer. Grouping related needs together allows the planning team to
“divide and conquer,” with subteams working on different parts of the design. Such subsystem or
component approaches to design, of course, have been common for years. What may be different
here is that the initial focus is on the components of the customers’ needs, not the components
of the product. The component design for the product will come during the later activities in
this step.
Determine Methods for Identifying Product Features. There are many complementary
approaches for identifying the best product design for meeting customers’ needs. Most design
projects do not use all of them. Before starting to design, however, a team should develop a systematic
plan for the methods it will use in its own design. Here are some of the options.
Benchmarking. This approach identifies the best in class and the methods behind it that make it
best. See Section 12 for details.
Basic Research. One aspect of research might be a new innovation for the product that does not
currently exist in the market or with competitors. Another aspect of basic research looks at exploring
the feasibility of the product and product features. While both these aspects are important, be
careful that fascination with the technological abilities of the product do not overwhelm the primary
concern of its benefits to the customer.
Market Experiments. Introducing and testing ideas for product features in the market allow one to
analyze and evaluate concepts. The focus group is one technique that can be used to measure customer
reactions and determine whether the product features actually will meet customer needs. Some
organizations also try out their ideas, on an informal basis, with customers at trade shows and association
meetings. Still others conduct limited test marketing with a prototype product.
Creativity. Developing product features allows one to dream about a whole range of possibilities
without being hampered by any restrictions or preconceived notions. Quality planning is a proven,
structured, data-based approach to meeting customers’ needs. But this does not mean it is rigid and
uncreative. At this point in the process, the participants in planning must be encouraged and given the
tools they need to be creative so as to develop alternatives for design. After they have selected a number
of promising alternatives, then they will use hard analysis and data to design the final product.
Planning teams can take advantage of how individuals view the world: from their own perspective.
Every employee potentially sees other ways of doing things. The team can encourage people to
suggest new ideas and take risks. Team members should avoid getting “stuck” or take too much time
to debate one particular idea or issue. They can put it aside and come back to it later with a fresh
viewpoint. They can apply new methods of thinking about customers’ needs or problems, such as the
 Changing key words or phrases. For example, call a “need” or “problem” an “opportunity.” Instead
of saying, “deliver on time,” say, “deliver exactly when needed.”
 Random association. For example, take a common word such as “apple” or “circus” and describe your
business, product, or problem as the word. For example, “Our product is like a circus because…”
 Central idea. Shift your thinking away from one central idea to a different one. For example, shift
the focus from the product to the customer by saying, “What harm might a child suffer, and how
can we avoid it?” rather than, “How can we make the toy safer?”
 Putting yourself in the other person’s shoes. Examine the question from the viewpoint of the other
person, your competitor, your customer—and build their case before you build your own.
 Dreaming. Imagine that you had a magic wand that you could wave to remove all obstacles to
achieving your objectives. What would it look like? What would you do first? How would it
change your approach?
 The spaghetti principle. When you have difficulty considering a new concept or how to respond to
a particular need, allow your team to be comfortable enough to throw out a new idea, as if you
were throwing spaghetti against the wall, and see what sticks. Often even “wild” ideas can lead to
workable solutions.
The initial design decisions are kept as simple as possible at this point. For example, the idea
of placing the control panel for the radio on the steering wheel would be considered a high-level
product feature. Its exact location, which controls, and how they function can be analyzed later in
more detail. It may become the subject of more detailed product features as the planning project
Standards, Regulations, and Policies. This is also the time to be certain that all relevant
standards, regulations, and policies have been identified and addressed. While some of these requirements
are guidelines for how a particular product or product feature can perform, others mandate
how they must perform. These may come from inside the organization, and others may come from
specific federal, state, or local governments, regulatory agencies, or industry associations. All product
features and product feature goals must be analyzed against these requirements before making
the final selection of product features to be included in the design.
It is important to note that if there is a conflict when evaluating product features against any standards,
policies, or regulations, it is not always a reason to give up. Sometimes one can work to gain
acceptance for a change when it will do a better job of meeting customer needs. This is especially
true when it comes to internal policies. However, an advocate for change must be prepared to back
the arguments up with the appropriate data.
Criteria for Design. As part of the preparation for high level design, the design team must
agree on the explicit criteria to be used in evaluating alternative designs and design features. All
designs must fulfill the following general criteria:
 Meet the customers’ needs
 Meet the suppliers’ and producers’ needs
 Meet (or beat) the competition
 Optimize the combined costs of the customers and suppliers
In addition to the preceding four general criteria, the team members should agree explicitly on
the criteria that it will use to make its selection. (If the choices are relatively complex, the team
should consider using the formal discipline of a selection matrix.) One source for these criteria will
be the team’s mission statement and goals. Some other types of criteria the team may develop include
 The impact of the feature on the needs
 The relative importance of the needs being served
 The relative importance of the customers whose needs are affected
 The feasibility and risks of the proposed feature
 The impact on product cost
 The relationship to competitive features uncovered in benchmarking
 The requirements of standards, policies, regulations, mandates, and so on
As part of the decision on how to proceed with design, teams also must consider a number of
other important issues regarding what type of product feature will be the best response to customers’
needs. When selecting product features, they need to consider whether to
 Develop an entirely new functionality
 Replace selected old features with new ones
 Improve or modify existing features
 Eliminate the unnecessary
Regulations and Standards for Ford’s Mustang. The Federal Motor Vehicle Safety Standards
(FMVSS) are, of course, a prime concern for designing any automobile. Ford had established its own
safety standards that were more extensive than the federal mandates and included a significant margin
of additional safety on all quantitative standards.
Select High-Level Product Features and Goals. This phase of quality planning will stimulate the
team to consider a whole array of potential product features and how each would respond to the
needs of the customer. This activity should be performed without being constrained by prior assumptions
or notions as to what worked or did not work in the past. A response that previously failed to
address a customer need or solve a customer problem might be ready to be considered again because
of changes in technology or the market.
The team begins by executing its plan for identifying the possible product features. It should then
apply its explicit selection criteria to identify the most promising product features.
The product design spreadsheet in Figure 3.9 is a good guide for this effort. Use the right side of
the spreadsheet to determine and document the following:
 Which product features contribute to meeting which customer needs
 That each priority customer need is addressed by at least one product feature
 That the total impact of the product features associated with a customer need is likely to be sufficient
for meeting that need
 That every product feature contributes to meeting at least one significant customer need
 That every product feature is necessary for meeting at least one significant customer need (i.e.,
removing that feature would leave a significant need unmet)
Now the team must set goals for each feature. In quality terms, a goal is an aimed-at quality target
(such as aimed-at values and specification limits). As discussed earlier, this differs from quality
standards in that the standard is a mandated model to be followed that typically comes from an external
source. While these standards serve as “requirements” that usually dictate uniformity or how the
product is to function, product feature goals are often voluntary or negotiated. Therefore, the quality
planning process must provide the means for meeting both quality standards and quality goals.
Criteria for Setting Product Feature Goals. As with all goals, product feature goals
must meet certain criteria. While the criteria for establishing product feature goals differ slightly
from the criteria for project goals verified in step 1, there are many similarities. Product feature goals
should encompass all the important cases and be
Measuring Product Features Goals. Establishing the measurement for a product feature
goal requires the following tasks:
 Determine the unit of measure: meters, seconds, days, percentages, and so on.
 Determine how to measure the goal (i.e., determine what is the sensor).
 Set the value for the goal.
The work done in measuring customer needs should be applied now. The two sets of measurements
may be related in one of the following ways:
 Measurement for the need and for the product feature goal may use the same units and sensors.
For example, if the customer need relates to timeliness measured in hours, one or more product
features normally also will be measured in hours, with their combined effects meeting the customer
 Measurement for the product feature may be derived in a technical manner from the need measurement.
For example, a customer need for transporting specified sizes and weights of loads may
be translated into specific engineering measurements of the transport system.
 Measurement for the product feature may be derived from a customer behavioral relationship with
the product feature measure. For example, automobile manufacturers have developed the specific
parameters for the dimensions and structure of an automobile seat that translate into the customer
rating it “comfortable.”
Since we can now measure both the customer need and the related product feature goals, it is possible
for the quality planning team to ensure that the product design will go a long way toward meeting
the customers’ needs, even before building any prototypes or conducting any test marketing.
For large or complex projects, the work of developing product features is often divided among a
number of different individuals and work groups. After all these groups have completed their work,
the overall quality planning team will need to integrate the results. Integration includes
 Combining product features when the same features have been identified for more than one cluster
 Identifying and resolving conflicting or competing features and goals for different clusters
 Validating that the combined design meets the criteria established by the team
Develop Detailed Product Features and Goals. For large and highly complex products,
it will usually be necessary to divide the product into a number of components and even subcomponents
for detailed design. Each component will typically have its own design team that will complete
the detailed design described below. In order to ensure that the overall design remains integrated,
consistent, and effective in meeting customer needs, these large, decentralized project require
 A steering or core team that provides overall direction and integration
 Explicit charters with quantified goals for each component
 Regular integrated design reviews for all components
 Explicit integration of designs before completion of the product design phase
Once the initial detailed product features and goals have been developed, then the technical designers
will prepare a preliminary design, with detailed specifications. This is a necessary step before a team
can optimize models of product features using a number of quality planning tools and ultimately set
and publish the final product features and goals.
It is not uncommon for quality planning teams to select product features at so high a level that
they are not specific enough to respond to precise customer needs. Just as in the identification of customers’
primary needs, high-level product features need to be broken down further into terms that
are clearly defined and which can be measured.
Optimize Product Features and Goals. Once the preliminary design is complete, it must
be optimized. That is, the design must be adjusted so that it meets the needs of both customer and
supplier while minimizing their combined costs and meeting or beating the competition.
Finding the optimum can be a complicated matter unless it is approached in an organized fashion
and follows quality disciplines. For example, there are many designs in which numerous variables
converge to produce a final result. Some of these designs are of a business nature, such as
design of an information system involving optimal use of facilities, personnel, energy, capital, and
so on. Other such designs are technological in nature, involving optimizing the performance of hardware.
Either way, finding the optimum is made easier through the use of certain quality disciplines.
Finding the optimum involves balancing the needs, whether they are multicompany needs or
within-company needs. Ideally, the search for the optimum should be done through the participation
of suppliers and customers alike. There are several techniques that help achieve this optimum.
Design Review. Under this concept, those who will be affected by the product are given the opportunity
to review the design during various formative stages. This allows them to use their experience
and expertise to make such contributions as
 Early warning of upcoming problems
 Data to aid in finding the optimum
 Challenge to theories and assumptions
Design reviews can take place at different stages of development of the new product. They can
be used to review conclusions about customer needs and hence the product specifications (characteristics
of product output). Design reviews also can take place at the time of selecting the optimal
product design. Typical characteristics of design reviews include the following:
 Participation is mandatory.
 Reviews are conducted by specialists, external to the planning team.
 Ultimate decisions for changes remain with the planning team.
 Reviews are formal, scheduled, and prepared for with agendas.
 Reviews will be based on clear criteria and predetermined parameters.
 Reviews can be held at various stages of the project.
Ground rules for good design reviews include
 Adequate advance planning of review agenda and documents
 Clearly defined meeting structure and roles
 Recognition of interdepartmental conflicts in advance
 Emphasis on constructive, not critical, inputs
 Avoidance of competitive design during review
 Realistic timing and schedules for the reviews
 Sufficient skills and resources provided for the review
 Discussion focus on untried/unproved design ideas
 Participation directed by management
Joint Planning. Planning teams should include all those who have a vested interest in the outcome
of the design of the product along with individuals skilled in product design. Under this concept, the
team, rather than just the product designers, bears responsibility for the final design.
Structured Negotiation. Customers and suppliers are tugged by powerful local forces to an extent
that can easily lead to a result other than the optimum. To ensure that these negotiating sessions proceed
in as productive a fashion as possible, it is recommended that ground rules be established before
the meetings. Here are some examples:
 The team should be guided by a spirit of cooperation, not competition, toward the achievement of
a common goal.
 Differences of opinion can be healthy and can lead to a more efficient and effective solution.
 Everyone should have a chance to contribute, and every idea should be considered.
 Everyone’s opinions should be heard and respected without interruptions.
 Avoid getting personal; weigh pros and cons of each idea, looking at its advantages before its disadvantages.
 Challenge conjecture; look at the facts.
 Whenever the discussion bogs down, go back and define areas of agreement before discussing
areas of disagreement.
 If no consensus can be reached on a particular issue, it should be tabled and returned to later on in
the discussion.
Create New Options. Often teams approach a product design with a history of how things were
done in the past. Optimization allows a team to take a fresh look at the product and create new
options. Some of the most common and useful quality tools for optimizing the design include the
Competitive analysis provides feature-by-feature comparison with competitors’ products. (See
the following for an example.)
Salability analysis evaluates which product features stimulate customers to be willing to buy the
product and the price they are willing to pay. (See the following for an example.)
Value analysis calculates not only the incremental cost of specific features of the product but also
the cost of meeting specific customer needs and compares the costs of alternative designs.
(See the following for an example.)
Criticality analysis identifies the “vital few” features that are vulnerable in the design so that they
can receive priority for attention and resources.
Failure mode and effect analysis (FMEA) calculates the combined impact of the probability of a
particular failure, the effects of that failure, and the probability that the failure can be detected
and corrected, thereby establishing a priority ranking for designing in failure-prevention countermeasures.
(See Section 19 under Reliability Analysis.)
Fault-tree analysis aids in the design of preventive countermeasures by tracing all possible combinations
of causes that could lead to a particular failure. (See Section 19 under Reliability
Analysis; also see Section 48.)
Design for manufacture and assembly evaluates the complexity and potential for problems during
manufacture to make assembly as simple and error-free as possible. Design for maintainability
evaluates particular designs for the ease and cost of maintaining them during their useful life.
Competitive Analysis. Figure 3.10 is an example of how a competitive analysis might be displayed.
The data for a competitive analysis may require a combination of different approaches such as laboratory
analysis of the competitors’ products, field testing of those products, or in-depth interviews
and on-site inspections where willing customers are using a competitor’s product.
Note that by reviewing this analysis, the planning team can identify those areas in which the
design is vulnerable to the competition, as well as those in which the team has developed an advantage.
Based on this analysis, the team will then need to make optimization choices about whether to
upgrade the product or not. The team may need to apply a value analysis to make some of these
Salability Analysis. An example of salability analysis is shown in Figure 3.11. This analysis is similar
to a competitive analysis, except that the reference point is the response of customers to the proposed
design rather than a comparison with the features of the competitors’ designs. Note, however,
that elements of competitive and salability analyses can be combined, with the salability analysis
incorporating customer evaluation of both the proposed new design and existing competitive designs.
Complex products, such as automobiles, with multiple optional features and optional configurations
offer a unique opportunity to evaluate salability. Observed installation rates of options on both
the existing car line and competitors’ cars provide intelligence on both the level of market demand
for the feature and the additional price that some segments of the market will pay for the feature—
although the other segments of the market may place little or no value on it.
Value Analysis. Value analysis has been quite common in architectural design and the development
of custom-engineered products, but it also can be applied successfully to other environments as well,
as illustrated in Figure 3.12. By comparing the costs for meeting different customer needs, the design
team can make a number of significant optimization decisions. If the cost for meeting low-priority
needs is high, the team must explore alternative ways to meet those needs and even consider not
addressing them at all if the product is highly price sensitive. If very important needs have not consumed
much of the expense, the team will want to make certain that it has met those needs fully and
completely. While low expense for meeting a high-priority need is not necessarily inappropriate, it
does present the designers with the challenge of making certain that lower-priority needs are not
being met using resources that could be better directed toward the higher-priority needs. It is not
uncommon for products to be overloaded with “bells and whistles” at the expense of the fundamental
functionality and performance.
Mustang’s Performance Features. One of the critical challenges for engineering the performance
of Mustang was to develop the ideal power-to-weight ratio that would meet the performance needs
of the customers. However, as is the case with most ratios of this sort, the ratio can be improved
either by reducing the weight or by increasing the power. What is more, fuel economy is also affected
by the weight.
Design trade-offs among weight, power, and fuel economy involved not only detailed engineering
calculations but also careful assessments of customer reactions and the competing views of different
functions within the company. Reaching a final design that met the customer needs and fulfilled sound
engineering principles required strong project leadership in addition to the data.
Product Feature & Goal
Product A Product B
Check if Product Feature is Present
Ours Product A Product B
Feature Performance vs. Goal (*)
Product A Product B
Check if Product Feature is Present
Ours Product A Product B
Feature Performance vs. Goal (*)
Identify if
Risk or
Identify if
Risk or
Retreive messages
from all touch
tone phones easily
Yes Yes Yes 4 5 4 —
Yes No Yes 3 — 5 O
No No Yes — — 4 O
Yes Yes 4 — R
Yes Yes 3 4 R
Change message
from any remote
2 lines built in
Below Add Features in
Competitors Product Not
Included in Ours
No cassette used to record
Telephone and answering
machine in one unit
FIGURE 3.10 Competitive analysis. (Juran Institute, Inc. Copyright 1994. Used by permission.)
Name of Product
Car Repair Service—
Basis for Rating
Prior Use
How Do
See Differences
Between Our
Products and
Buy If Price Were
Not Important?
Would Customer
Buy If Price Were
Of All Products
Listed, Prioritize
Which Would
Customers Buy
and Its Basis?
Identify if
Risk or
Opportunity Price Yes
How Do
Rate Product?
U Y $175 Y 2-F E
O N $145 Y 3-P G
O $175 Y 1-F
Competitor A—
Competitor B—
Name of Product
Pick-up and delivery
of car to be repaired
Product Feature Goal:
Same Day Service
Basis for Rating
Prior Use
How Do Customers
See Differences
Between Our
Features Against
Competing Features?
Positively (+)
Negatively (–)
No Difference
Does the
Addition of the
Feature Make
the Product:
More Salable
Less Salable
No Difference
Identify if
Risk or
How Do
Rate Product?
O — O S
U +
— R E
Competitor A—
Not Offered
Competitor B—
Offered. Also provides
loaner car to customer
FIGURE 3.11 Salability analysis for automobile maintenance service. (Juran Institute, Inc. Copyright 1994. Used
by permission.)
Set and Publish Final Product Design. After the design has been optimized and tested, it
is time to select the product features and goals to be included in the final design. This is also the stage
where the results of product development are officially transmitted to other functions through various
forms of documentation. These include the specifications for the product features and product
feature goals, as well as the spreadsheets and other supporting documents. All this is supplemented
by instructions, both oral and written. To complete this activity, the team must first determine the
process for authorizing and publishing product features and product feature goals. Along with the
features and goals, the team should include any procedures, specifications, flow diagrams, and other
spreadsheets that relate to the final product design. The team should pass along results of experiments,
field testing, prototypes, and so on, that are appropriate. If an organization has an existing
process for authorizing product goals, it should be reexamined in light of recent experience. Ask
these questions, Does the authorization process guarantee input from key customers—both internal
and external? Does it provide for optimization of the design? If an organization has no existing goal
authorization process, now is a good time to initiate one.
Once the product is developed, it is necessary to determine the means by which the product will be
created and delivered on a continuing basis. These means are, collectively, the “process.” “Process
development” is the set of activities for defining the specific means to be used by operating personnel
for meeting product quality goals. Some related concepts include
(listed in
Cost of
Product Feature & Goals
Product: Store Front Prenatal Clinic
Walk in
handled by
Nurse, 5
days a week
2 days a
Worker, 5
days a
to use
60,000 30,000 10,000 10,000 20,000 40,000 170,000
in staff
70,000 10,000 15,000 95,000
25,000 25,000
Sensitivity 5,000 5,000
15,000 15,000
Cost for
60,000 100,000 40,000 45,000 20,000
65,000 330,000
under 1
5 days a
On-site Billing
Clerk takes
insurance from
all eligible
FIGURE 3.12 Value analysis for prenatal clinic. (Juran Institute, Inc. Copyright 1994. Used by permission.)
Subprocesses: Large processes may be decomposed into these smaller units for both the development
and operation of the process.
Activities: The steps in a process or subprocess.
Tasks: The detailed step-by-step description for execution of an activity.
In order for a process to be effective, it must be goal oriented, with specific measurable outcomes;
systematic, with the sequence of activities and tasks fully and clearly defined and all inputs and outputs
fully specified; and capable, i.e., able to meet product quality goals under operating conditions
and legitimate, with clear authority and accountability for its operation.
The eleven major activities involved in developing a process are
 Review product goals.
 Identify operating conditions.
 Collect known information on alternate processes.
 Select general process design.
 Identify process features and goals.
 Identify detailed process features and goals.
 Design for critical factors and human error.
 Optimize process features and goals.
 Establish process capability.
 Set and publish final process features and goals.
 Set and publish final process design.
Review Product Goals. Ideally, this review will be relatively simple. Product quality goals
should have been validated with the prior participation of those who would be affected. In many
companies, however, product and process design often are executed by different teams. There is no
real joint participation on either group’s part to contribute to the results that both the teams are
expected to produce. This lack of participation usually reduces the number of alternative designs that
could have been readily adopted in earlier stages but become more difficult and more expensive to
incorporate later. In addition, those who set the product goals have a vested interest in their own decisions
and exhibit cultural resistance to proposals by the process design team to make changes to the
product design. If the product and process design efforts are being performed by different groups,
then review and confirmation of the product quality goals are absolutely critical.
Review of product quality goals ensures that they are understood by those most affected by the
process design. The review helps achieve the optimum. Process designers are able to present product
designers with some realities relative to the costs of meeting the quality goals. The review
process should provide a legitimate, unobstructed path for challenging costly goals.
Identify Operating Conditions. Seeking to understand operating conditions requires investigation
of a number of dimensions.
User’s Understanding of the Process. By “users,” we mean those who either contribute to the
processes in order to meet product goals or those who employ the process to meet their own needs.
Users consist, in part, of internal customers (organization units or persons) responsible for running
the processes to meet the quality goals. Operators or other workers are users. Process planners
need to know how these people will understand the work to be done. The process must be designed
either to accommodate this level of understanding or to improve the level of understanding.
How the Process Will be Used. Designers always know the intended use of the process they develop.
However, they may not necessarily know how the process is actually used (and misused) by the end
user. Designers can draw on their own experiences but usually must supplement these with direct
observation and interviews with those affected.
The Environments of Use. Planners are well aware that their designs must take account of environments
that can influence process performance. Planners of physical processes usually do take
account of such environmental factors as temperature, vibration, noise level, and so on. Planners who
depend heavily on human responses, particularly those in the service areas, should address the
impact of the environment on human performance in their process designs. For example, a team
designing the process for handling customer inquiries should consider how environmental stress can
influence the performance of the customer service representatives. This stress can result from large
numbers of customer complaints, abusive customers, lack of current product information, and so on.
Collect Known Information on Alternative Processes. Once the goals and environment
are clear, the planning team needs reliable information on alternative processes available for
meeting those goals in the anticipated environment.
Process Anatomy. At the highest level, there are some basic process anatomies that have specific
characteristics that planners should be aware of. A “process anatomy” is a coherent structure that
binds or holds the process together. This structure supports the creation of the goods or the delivery
of the service. The selection of a particular anatomy also will have a profound influence on how the
product is created and the ability of the organization to respond to customers’ needs. Figure 3.13
illustrates these.
The Autonomous Department. The “autonomous process” is defined as a group of related activities
that are usually performed by one department or a single group of individuals. In this process
form, the department or group of individuals receives inputs from suppliers, such as raw materials,
parts, information, or other data, and converts them into finished goods and services, all within a single
self-contained department.
An example of an autonomous process is the self-employed professional, e.g., a physician, consultant,
or artisan. In financial services, it might be the loan-approval department. In manufacturing,
a well-known example is a tool room. It starts with tool steel and engineering drawings and creates
punches, dies, fixtures, and gauges to be used on the manufacturing floor. Even though we refer to
this kind of process anatomy as “autonomous,” outputs or deliverables from other processes are still
required from outside sources that serve as inputs into this process. The self-employed physician, for
example, may purchase equipment and materials from supply houses, pharmaceutical companies,
and so on.
The Assembly Tree. The “assembly tree” is a familiar process that incorporates the outputs of several
subprocesses. Many of these are performed concurrently and are required for final assembly or
to achieve an end result at or near the end of the process. This kind of process anatomy is widely
used by the great mechanical and electronic industries that build automotive vehicles, household
appliances, electronic apparatus, and so on. It is also used to define many processes in a hospital,
such as in the case of performing surgery in the operating room. The branches or leaves of the tree
represent numerous suppliers or in-house departments making parts and components. The elements
are assembled by still other departments.
In the office, certain processes of data collection and summary also exhibit features of the assembly
tree. Preparation of major accounting reports (e.g., balance sheet, profit statement) requires
assembly of many bits of data into progressively broader summaries that finally converge into the
consolidated reports. The assembly-tree design has been used at both the multifunctional and departmental
levels. In large operations, it is virtually mandatory to use staff specialists who contribute different
outputs at various multifunctional levels. An example of this is the budget process. While it is
not mandatory to use staff specialists for large departmental processes, this is often the case. This
can be illustrated by the design department, where various design engineers contribute drawings of
a project that contribute to the overall design.
Flow of basic materials
Outflow of finished
and tested product
Vendor departments
In-house departments
Final assembly
Vendor departments
In-house departments
To test
and usage
To test and usage
Assembly Tree
Autonomous Department
Process Anatomies
FIGURE 3.13 Process anatomies. (Juran Institute, Inc. Copyright 1994. Used by
The Procession. Another familiar form, the “procession process,” uses a sequential approach as
the basis for the process. This differs from the assembly tree, in which many of the activities are performed
concurrently. The procession approach tends to take a more linear approach, whereby the
lower-level processes are performed sequentially. It mandates that certain activities must be completed
before others can begin because the outputs of each of the subprocesses serve as the inputs for
each succeeding subprocess.
The intent of selecting a process anatomy is to determine the overall structure or architecture of the
process that produces the product features and meets product feature goals. It does not necessarily follow
that choosing one process anatomy over another locks the team into using that same architecture
exclusively throughout the entire system. Quite the contrary, the team may select the assembly-tree
process as the structure for the overall system but use a combination of autonomous and procession
anatomies as the basis for subprocesses at the functional, departmental, or unit level.
Process Quality Management. Increasingly, many planners are applying a fourth, less traditional
form of management known as “process quality management” to their major processes. This new,
alternative management form has come about in response to an increased realization that many of
today’s business goals and objectives are becoming even more heavily dependent on large, complex,
cross-functional business processes. Process quality management emphasizes that there are several
critical processes that are crucial to an organization if it is to maintain and grow its business. (See
Section 6 for a full discussion.)
Measuring the Process. In selecting a specific process design, the team will need to acquire information
on the effectiveness and efficiency of alternative designs, including
 Deficiency rates
 Cycle time
 Unit cost
 Output rate
To acquire the needed data, the planners must typically use a number of different approaches,
 Analyzing the existing process
 Analyzing similar or related processes
 Testing alternative processes
 Analyzing new technology
 Acquiring information from customers
 Simulating and estimating
Select General Process Design. Just as product design began with a high-level description
expanded to the details, process design should begin by describing the overall process flow with a
high-level process-flow diagram. From this diagram it will be possible to identify the subprocesses
and major activities that can then be designed at a more detailed level. In developing the high-level
flow, as well as the greater detail later, the team should ensure that it meets the following criteria:
 Will deliver the quality goals for the product
 Incorporates the countermeasures for criticality analysis, FMEA, and fault-tree analysis
 Meets the project goals
 Accounts for actual, not only intended, use
 Is efficient in consumption of resources
 Demands no investments that are greater than planned
While some process designs will largely repeat existing designs and some others will represent
“green field” or “blank sheet” redesigns, most effective process redesigns are a combination of the
tried and true existing processes with some significant quantum changes in some parts of the process.
The preceding criteria should be the guides for whether a particular part of the process should be
incorporated as it is, improved, or replaced with a fundamentally different approach.
This is the point in process design to think as creatively as possible, using some of the same techniques
discussed under product development. Consider the impact of radically different anatomies.
Would the customer be served better with dedicated, multispecialty units or with highly specialized
expert functionality accessed as needed? What approach is mostly likely to reduce deficiencies? How
can cycle time by cut dramatically? Is there a new technology that would allow us to do it differently?
Can we develop such a technology?
Once the high-level flow is completed, each activity and decision within the flow diagram needs
to be fully documented with a specification of the following for each:
 Goals for outputs
 Cycle time
 General description of the conversion of inputs to outputs
Clear specification of these factors makes it possible to divide up the work of detailed design later
and still be confident that the final design will be consistent and coordinated.
Once the initial new process flow is completed, it should be reviewed for opportunities to
improve it, such as
 Eliminate sources of error that lead to rework loops.
 Eliminate or reduce redundant subprocesses, activities, or tasks.
 Decrease the number of handoffs.
 Reduce cycle time.
 Replace tasks, activities, or processes that have outputs with defects.
 Correct sequencing issues in the process to reduce the amount of activity or rework.
Carryover of Process Designs. For each subprocess or major activity, one of the following questions
must be answered in the affirmative:
If it is a carryover design, is the process capable of meeting the product quality goals?
If it is a new design, can we demonstrate that it is at least as effective at meeting product quality
goals while also maintaining or improving cost and cycle time?
Testing Selected Processes. One of the key factors for a successful design is incorporating the
lessons learned from testing the product, the product features, and the overall process and subprocesses
to ensure that they meet quality goals. Testing should be conducted throughout the entire
quality planning process to allow for changes, modifications, and improvements to the plan before it
is transferred to operations. Testing is performed at various points to analyze and evaluate alternate
designs of the overall process and subprocesses.
There are a number of options for testing the efficiency and effectiveness of a process prior to
full-scale implementation. They include the following:
Pilot test: A pilot test tests the overall process on a small scale or with a small segment of the
total population. The segment to receive testing will vary depending on the process itself. Testing
may be limited to a particular location, department, or function.
Modular test: Sometimes it is not possible to test the entire process at one time, but it may be
possible to test crucial elements of the process separately. A modular test is a test of individual
segments of the process. Generally, the outputs of certain subprocesses influence the ability of
other processes to perform efficiently and effectively. These critical processes require their own
tests to isolate problems that may occur and allow improvements to be made.
Simulation: This design technique observes and manipulates a mathematical or physical model that
represents a real-world process for which, for technical or economic reasons, direct experimentation
is not possible. Different circumstances can be applied to test how the process will perform under
varying conditions, inputs, and worst-case scenarios.
Dry run: A dry run is a walk-through of the new process, with the planning team playing a dominant
operating role in the process. This is a test of the process under operating conditions. The purpose
is to test the process. Any resulting product is not sent to customers. Usually the team has
worked so closely with designing the process that it can lose sight of how the various pieces actually
fit together. The dry run gives the team one last opportunity to step back and see, from a conceptual
standpoint, whether the process can work as designed before other tests are performed or before
the process is transferred to operations.
Acceptance test: This is a highly structured form of testing common in complex systems, such as
computer systems. A test plan is designed by a special team not directly involved in the design of the
process being tested. The test plan sets up the proper environmental conditions, inputs, relevant interventions,
and operating conditions. The test is intended to stress, in relevant ways, the important functional
and other features in which the process could fail. In some cases, it is vital that the new process
design be tested under operating conditions by the people who will actually operate it—assuming
they are different from the planning team. The team may not have understood problem operating conditions;
there may be unforeseen problems or resistance that cannot be overcome. Without such a test,
these factors could contribute to a very costly mistake. Therefore, in such cases, acceptance testing
under real conditions is essential.
Comparisons or benchmarks. Other units inside and outside the organization may already be
using a process similar to the one designed. The process can be validated by comparing it with
existing similar processes.
Test Limitations. All tests have some limitations. The following are common limitations that
should be understood and addressed.
Differences in operating conditions: Dry runs and modular testing obviously differ from operating
conditions. Even pilot tests and benchmarks will differ in some details from the actual, full
implementation. Some common differences between conditions for testing and conditions for
full-scale use include
 People operating the process
 Customers of the process
 Extreme values and unusual conditions
 Interactions with other processes and other parts of the organization.
Differences in size: Especially with critical failures, such as breakdown of equipment, loss of key
personnel, or any other potential failure, as in the case of complications in a surgical procedure, a test
might not be large enough to allow these rare failures to occur with any high degree of certainty.
Cultural resistance: Cultural reactions to tests differ from reactions to permanent changes. Such
reactions might be either more or less favorable than full-scale implementation. Tests may go well
because they lack the cultural impact of full implementation. They may go poorly because
participants will not give the test the same careful attention they would give the “real” work.
Other effects. Sometimes designing a new process or redesigning an existing process may create
or exacerbate problems in other processes. For example, improved turnaround time in approving
home loans may create a backlog for the closing department. Such interactions among
processes might not occur in an isolated test.
Identify Process Features and Goals. A “process feature” is any property, attribute, and
so on that is needed to create the goods or deliver the service and achieve the product feature goals
that will satisfy a customer need. A “process goal” is the numeric target for one of the features.
Whereas product features answer the question, “What characteristics of the product do we need
to meet customers needs?” process features answer the question, “What mechanisms do we need to
create or deliver those characteristics (and meet quality goals) over and over again without deficiencies?”
Collectively, process features define a process. The flow diagram is the source of many, but
not all, of these features and goals.
As the process design progresses from the macro level down into details, a long list of specific
process features emerges. Each of these is aimed directly at producing one or more product features.
For example:
 Creating an invoice requires a process feature that can perform arithmetic calculations so that
accurate information can be added.
 Manufacturing a gear wheel requires a process feature that can bore precise holes into the center
of the gear blank.
 Selling a credit card through telemarketing requires a process feature that accurately collects customer
Most process features fall into one of the following categories:
 Procedures—a series of steps followed in a regular, definite order
 Methods—an orderly arrangement of a series of tasks, activities, or procedures
 Equipment and supplies—”physical” devices and other hard goods that will be needed to perform
the process
 Materials—tangible elements, data, facts, figures, or information (these, along with equipment
and supplies, also may make up inputs required as well as what is to be done to them)
 People—numbers of individuals, skills they will require, goals, and tasks they will perform
 Training—skills and knowledge required to complete the process
 Other resources—additional resources that may be needed
 Support processes—can include secretarial support, occasionally other support, such as outsources
of printing services, copying services, temporary help, and so on.
Just as in the case of product design, process design is easier to manage and optimize if the
process features and goals are organized into a spreadsheet indicating how the process delivers the
product features and goals. Figure 3.14 illustrates such a spreadsheet.
The spreadsheet serves not only as a convenient summary of the key attributes of the process, it
also facilitates answering two key questions that are necessary for effective and efficient process
design. First, will every product feature and goal be attained by the process? Second, is each process
feature absolutely necessary for at least one product feature; i.e., are there any unnecessary or redundant
process features? Also, verify that one of the other process features cannot be used to create the
same effect on the product.
Often high-level process designs will identify features and goals that are required from companywide
macro processes. Examples might include cycle times from the purchasing process, specific
data from financial systems, and new skills training. Because the new process will depend on these
macro processes for support, now is the time to verify that they are capable of meeting the goals. If
they are not, the macro processes will need to be improved as part of the process design, or they will
need to be replaced with an alternative delivery method.
Identify Detailed Process Features and Goals. In most cases, it will be most efficient
and effective for individual subteams to carry out the detailed designs of subprocesses and major
activities. These detailed designs will have the process features and goals as their objectives and criteria.
Each subprocess team will develop the design to the level at which standard operating procedures
can be developed, software coded, equipment produced or purchased, and materials acquired.
Design for Critical Factors and Human Error. One key element of process design is
determining the effect that critical factors will have on the design. “Critical factors” are those aspects
which present serious danger to human life, health, and the environment or risk the loss of very large
sums of money. Some examples of such factors involve massive scales of operations: airport traffic
control systems, huge construction projects, systems of patient care in hospital, and even the process
for managing the stock market. Planning for such factors should obviously include ample margins
of safety as to structural integrity, fail-safe provisions, redundancy systems, multiple alarms, and so
on. Criticality analysis and failure-mode and effect analysis (see Section 19) are helpful tools in identifying
those factors which require special attention at this point.
Workers vary in their capabilities to perform specific tasks and activities. Some workers perform
well, whereas others do not perform nearly as well. What is consistent about all workers is that they
are a part of the human family, and human beings are fallible. Collectively, the extent of human
errors is large enough to require that the process design provides for means to reduce and control
human error. Begin by analyzing the data on human errors, and then apply the Pareto principle. The
vital few error types individually become candidates for special process design. The human errors
that can be addressed by process design fall into these major classes:
 Technique errors arising from individuals lacking specific, needed skills
 Errors aggravated by lack of feedback
 Errors arising from the fact that humans cannot remain indefinitely in a state of complete, ready
Very Strong
Time to perform job
appointment time
All materials
environmentally safe
Less than one hour 100
percent of time
forecast on P.C.
to determine
to/from and
work needed
10 gallons
per minute
One person
per 10,000
sq. ft. of yd.
approved by
State Dept. of
Forecast time
always within
10 percent
of actual
99 percent of jobs
within 15 minutes
of appointment
All naturally occuring/no
Product Feature Product Feature Goal
Process Features
Process Feature Goals
FIGURE 3.14 Process design spreadsheet for a lawn care service. (Juran Institute, Inc. Copyright 1994. Used by
Technique Errors. Some workers consistently outperform others on specific quality tasks. The
likely reason is possession of a special “knack.” In such cases, designers should study the methods
used by the respective workers to discover the methodologic differences. These differences usually
include the knack—a small difference in method that produces a big difference in performance. Once
the knack is discovered, the process designers can arrange to include the knack in the technology.
Alternatively, the knack can be brought into the workers’ training program so that all workers are
brought up to the level of the best.
Lack of Instant Feedback. A useful principle in designing human tasks is to provide instant feedback
to the worker so that the performance of the work conveys a message about the work to the worker. For
example, a worker at a control panel pushes a switch and receives three feedbacks: the feel of the shape
of the switch handle, the sound of an audible click signaling that the switch went all the way, and the
sight of a visual illumination of a specific color and shape. Providing such feedback is part of self-control
and allows the worker to modify his or her performance to keep the process within its quality goals.
Human Inattention Errors. A technique for designing human work is to require human attention
as a prerequisite for completing the work; i.e., the task cannot be performed unless the person doing
it devotes attention to it and to nothing else. A widespread case in point is inspection of documents,
products, or whatever. Human checking can be done in two very different ways.
By passive deeds: Listening, looking, reading. Such deeds are notoriously subject to lapses in
human attention. Also, such deeds leave no trail behind them. We have no way of knowing
whether the human being in question is really paying attention or is in a state of inattention. For
example, a person providing visual inspection of a product moving along an assembly line or
someone proofreading a report may become fatigued. They can easily experience a momentary
lapse in their attention, causing them to miss spotting a defect or to fail to notice that a column
of numbers does not add up correctly.
By active deeds: Operating a keyboard, writing, spelling. Such deeds cannot be performed at
all without paying attention to the task at hand and to the exclusion of all else. These active deeds
do leave a trail behind them. They are therefore far less error-prone than passive checking. An
example would be someone having to attach the leads of a voltage meter to a circuit board to
check its resistance or a blood bank technician retesting each sample to verify blood type.
Inadvertent human errors and other types of errors can also be reduced by “errorproofing”—
building processes so that the error either cannot happen or is unlikely to happen.
Principles of Errorproofing. Research has indicated that there are a number of different classifications
of errorproofing methods, and these are spelled out below.
Elimination: This consists of changing the technology to eliminate operations that are errorprone.
For example, in some materials handling operations, the worker should insert a protective
pad between the lifting wire and the product so that the wire will not damage the product.
Elimination could consist of using nylon bands to do the lifting.
Replacement: This method retains the error-prone operation but replaces the human worker
with a nonhuman operator. For example, a human worker may install the wrong component into
an assembly. A properly designed robot avoids such errors. Nonhuman processes, so long as they
are properly maintained, do not have lapses in attention, do not become weary, do not lose their
memory, and so on.
Facilitation: Under this method, the error-prone operation is retained, and so is the human
worker. However, the human worker is provided with a means to reduce any tendency toward
errors. Color coding of parts is an example.
Detection: This method does nothing to prevent the human error from happening. Instead, it
aims to find the error at the earliest opportunity so as to minimize the damage done. A widespread
example is automated testing between steps in a process.
Mitigation: Here again, the method does nothing to prevent the human error from happening.
However, means are provided to avoid serious damage done. A common example is providing a
fuse to avoid damage to electrical equipment.
Optimize Process Features and Goals. After the planners have designed for critical factors
and made modifications to the plan for ways of reducing human error, the next activity is to
optimize first the subprocesses and then the overall process design. In step 4, develop product, the
concept of optimization was introduced. The same activities performed for optimizing product features
and product feature goals also apply to process planning. Optimization applies to both the
design of the overall process and the design of individual subprocesses.
Establish Process Capability. Before a process begins operation, it must be demonstrated
to be capable of meeting its quality goals. The concepts and methods for establishing process capability
are discussed in detail in Section 22, under Process Capability. Any planning project must measure
the capability of its process with respect to the key quality goals. Failure to achieve process
capability should be followed by systematic diagnosis of the root causes of the failure and improvement
of the process to eliminate those root causes before the process becomes operational.
Reduction in Cycle Time. Process capability relates to the effectiveness of the process in meeting
customer needs. One special class of needs may relate to subprocess cycle time—the total time
elapsed from the beginning of a process to the end. Reducing cycle time has almost become an
obsession for many organizations. Pressures from customers, increasing costs, and competitive
forces are driving companies to discover faster ways of performing their processes. Often these targeted
processes include launching new products, providing service to customers, recruiting new
employees, responding to customer complaints, and so on. For existing processes, designers follow
the well-known quality-improvement process to reduce cycle time. Diagnosis identifies causes for
excessive time consumption. Specific remedies are then developed to alleviate these causes. (See
Section 5, The Quality Improvement Process.)
Set and Publish Final Process Features and Goals. After the planning team has established
the flow of the process, identified initial process features and goals, designed for critical
processes and human error, optimized process features and goals, and established process capabilities,
it is ready to define all the detailed process features and goals to be included in the final design.
This is also the stage where the results of process development are officially transmitted to other
functions through various forms of documentation. These include the specifications for the product
features and product feature goals as well as the spreadsheets and other supporting documents. All
this is supplemented by instructions, both oral and written.
Filling out the process design spreadsheet is an ongoing process throughout process development.
The spreadsheet should have been continually updated to reflect design revisions from such activities
as reviewing alternative options, designing for critical factors and human error, optimizing, testing
process capability, and so on. After making the last revision to the process design spreadsheet, it
should be checked once more to verify the following:
 That each product feature has one or more process features with strong or very strong relation.
This will ensure the effective delivery of the product feature without significant defects. Each
product feature goal will be met if each process goal is met.
 That each process feature is important to the delivery of one or more product features. Process
features with no strong relationship to other product features are unnecessary and should be
The completed process design spreadsheet and detailed flow diagrams are the common information
needed by managers, supervisors, and workers throughout the process. In addition, the planning
team must ensure that the following are also specified for each task within the process:
 Who is responsible for doing it
 How the task is to be competed
 Its inputs
 Its outputs
 Problems that can arise during operations and how to deal with them
 Specification of equipment and materials to be used
 Information required by the task
 Information generated by the task
 Training, standard operating procedures, job aids that are needed
In this step, planners develop controls for the processes, arrange to transfer the entire product plan
to operational forces, and validate the implementation of the transfer. There are seven major activities
in this step.
 Identify controls needed.
 Design feedback loop.
 Optimize self-control and self-inspection.
 Establish audit.
 Demonstrate process capability and controllability.
 Plan for transfer to operations.
 Implement plan and validate transfer.
Once planning is complete, these plans are placed in the hands of the operating departments. It
then becomes the responsibility of the operational personnel to manufacture the goods or deliver the
service and to ensure that quality goals are met precisely and accurately. They do this through a
planned system of quality control. Control is largely directed toward continuously meeting goals and
preventing adverse changes from affecting the quality of the product. Another way of saying this is
that no matter what takes place during production (change or loss of personnel, equipment or electrical
failure, changes in suppliers, etc.), workers will be able to adjust or adapt the process to these
changes or variations to ensure that quality goals can be achieved.
Identify Controls Needed. Process control consists of three basic activities:
 Evaluate the actual performance of the process.
 Compare actual performance with the goals.
 Take action on the difference.
Detailed discussions of these activities in the context of the feedback loop are contained in
Section 4, The Quality Control Process.
Control begins with choosing quality goals. Each quality goal becomes the target at which the
team directs its efforts. All control is centered around specific things to be controlled. We will call
these things “control subjects.” Each control subject is the focal point of a feedback loop. Control
subjects are a mixture of
Product features: Some control is carried out by evaluating features of the product itself (e.g.,
the invoice, the gear wheel, the research report, etc.) Product controls are associated with the deci-
sion: Does this product conform to specifications or goals? Inspection is the major activity for
answering this question. This inspection is usually performed at points where the inspection results
make it possible to determine where breakdowns may have occurred in the production process.
Process features: Much control consists of evaluating those process features which most directly
affect the product features, e.g., the state of the toner cartridge in the printer, the temperature
of the furnace for smelting iron, or the validity of the formulas used in the researcher’s report.
Some features become candidates for control subjects as a means of avoiding or reducing failures.
These control subjects typically are chosen from previously identified critical factors or
from conducting FMEA, FTA, and criticality analysis. Process controls are associated with the
decision: Should the process run or stop?
Side-effect features: These features do not affect the product, but they may create troublesome
side effects, such as irritations to employees, offense to the neighborhood, threats to the environment,
and so on.
These three types of control subjects may be found at several different stages of the process:
 During operations, including
Running control
Product control
Supporting operations control
Facility and equipment control
Design Feedback Loop. Once the control subjects are selected, it is time to design the
remainder of the feedback loop by
 Setting the standards for control—i.e., the levels at which the process is out of control and the
tools, such as control charts, that will be used to make the determination
 Deciding what action is needed when those standards are not met, e.g., troubleshooting.
 Designating who will take those actions
A detailed process flow diagram should be used to identify and document the points at which
control measurements and actions will be taken. Then each control point should be documented on
a control spreadsheet similar to Figure 3.15.
Optimize Self-Control and Self-Inspection. As discussed in more detail in Section 22,
Operations, self-control takes place when workers know what they are supposed to do. Goals and
targets are clearly spelled out and visible.
 Workers know what they are doing. Their output is measured, and they receive immediate feedback
on their performance.
 Workers have the ability and the means to regulate the outcomes of the process. They need a capable
process along with the tools, training, and authority to regulate it.
In addition to providing the optimal conditions for process operation and control, establishing
self-control has a significant, positive impact on the working environment and the individuals in it.
Whenever possible, the design of the quality control system should stress self-control by the operating
forces. Such a design provides the shortest feedback loop but also requires the designers to
ensure that the process capability is adequate to meet the product quality goals.
Once self-control is established, self-inspection should be developed. Self-inspection permits
the worker to check that the product adheres to quality standards before it is passed on to the next
station in the production cycle. Production and front-line workers are made to feel more responsible
for the quality of their work. Feedback on performance is immediate, thereby facilitating
process adjustments. Traditional inspection also has the psychological disadvantage of using an
“outsider” to report the defects to the worker. The costs of a separate inspection department can be
However, some prerequisite criteria must first be established:
Quality is number one: Quality must undoubtedly be made the highest priority. If this is not
clear, the workers succumb to schedule and cost pressures and classify as acceptable products that
should be rejected.
Mutual confidence: Managers must trust the workers enough to be willing to delegate the
responsibility and the authority to carry out the work. Workers must also have enough confidence
in managers to be willing to accept this responsibility and authority.
Training: Workers should be trained to make the product conformance decisions and should
also be tested to ensure that they make good decisions.
Specifications must be unequivocally clear.
The quality audit and audit of control systems are treated elsewhere in detail—see, for example,
Section 22, under Audit of Operations Quality. While the audit of a control system is a function independent
of the planning team, the planning team does have the responsibility for ensuring that adequate
documentation is available to make an effective audit possible and that there are provisions of
resources and time for conducting the audit on an ongoing basis.
FIGURE 3.15 Control spreadsheet. [From Juran, J. M. (1988), Quality Control Handbook, 4th ed. McGraw-Hill, New York, 6.9.]
Demonstrate Process Capability and Controllability. While process capability must be addressed
during the design of the process, it is during implementation that initial findings of process capability
and controllability must be verified.
Plan for Transfer to Operations. In many organizations, receipt of the process by operations is
structured and formalized. An information package is prepared consisting of certain standardized
essentials: goals to be met, facilities to be used, procedures to be followed, instructions, cautions, and
so on. There are also supplements unique to the project. In addition, provision is made for briefing
and training the operating forces in such areas as maintenance, dealing with crisis, and so on. The
package is accompanied by a formal document of transfer of responsibility. In some organizations,
this transfer takes place in a near-ceremonial atmosphere.
The structured approach has value. It tends to evolve checklists and countdowns that help ensure
that the transfer is orderly and complete. If the organization already has a structure for transfer, project
information may be adapted to conform with established practice. If the company has a loose structure
or none at all, the following material will aid in planning the transfer of the project.
Regardless of whether the organization has a structure or not, the team should not let go of the
responsibility of the project until it has been validated that the transfer has taken place and everyone
affected has all the information, processes, and procedures needed to produce the final product.
Transfer of Know-How. During process design, the planners acquire a great deal of know-how
about the process. The operating personnel could benefit from this know-how if it were transferred.
There are various ways of making this transfer, and most effective transfers make use of several complementary
channels of communication, including
Process specifications
On-the-job training
Formal training courses
Prior participation
Audit Plan for the Transfer. As part of the plan for formal transfer, a separate audit plan
should also be developed as a vehicle for validating the transfer of the plan. This kind of audit is
different from the control audits described previously. The purpose of this audit is to evaluate how
successful the transfer was. For the audit to have real meaning, specific goals should be established
during the planning phase of the transfer. Generally, these goals relate to the quality goals established
during the development of the product, product features, and process features. The team may
decide to add other goals inherent to the transfer or to modify newly planned quality goals during
the first series of operations. For example, during the first trial runs for producing the product, total
cycle time may exceed expected goals by 15 percent. This modification takes into account that
workers may need time to adjust to the plan. As they become more skilled, gain experience with the
process, and get more comfortable with their new set of responsibilities, cycle time will move closer
to targeted quality goals.
The audit plan for the transfer should include the following:
 Goals to meet
 How meeting the goals will be measured
 The time phasing for goals, measurement, and analysis
 Who will audit
 What reports will be generated
 Who will have responsibility for corrective action for failure to meet specific goals
Implement Plan and Validate Transfer. The final activity of the quality planning process
is to implement the plan and validate that the transfer has occurred. A great deal of time and effort
has gone into creating the product plan, and validating that it all works is well worth the effort.
Designs for World Class Quality (1995). Juran Institute, Wilton, CT.
Juran, Joseph M. (1992). Quality by Design. Free Press, New York.
Parasuraman, A., Zeithami, Valarie A., and Berry, Leonard L. (1985). “A Conceptual Model for Service Quality
and Its Implications for Further Research.” Journal of Marketing, Fall, pp. 41–50.
Veraldi, L. C. (1985). “The Team Taurus Story.” MIT Conference paper, Chicago, Aug. 22. Center for Advanced
Engineering Study, MIT, Cambridge, MA.
J. M. Juran
A. Blanton Godfrey
Quality Control Defined 4.2
The Relation to Quality Assurance 4.3
The Feedback Loop 4.3
LOOP 4.5
Choose the Control Subject 4.5
Establish Measurement 4.6
Establish Standards of Performance:
Product Goals and Process Goals 4.6
Measure Actual Performance 4.7
The Sensor 4.7
Compare to Standards 4.7
Take Action on the Difference 4.7
The Process 4.8
The PDCA Cycle 4.8
Control by Nonhuman Means 4.9
Control by the Work Force 4.10
Control by the Managerial Hierarchy 4.10
The Customers and Their Needs 4.10
Who Plans? 4.11
Quality Control Concepts 4.11
The Flow Diagram 4.11
Control Stations 4.11
Setup (Startup) Control 4.12
Running Control 4.13
Product Control 4.13
Facilities Control 4.13
Concept of Dominance 4.14
Seriousness Classification 4.14
Process Capability 4.14
Who Does What? 4.16
Special and Common Causes of
Variation 4.16
The Shewhart Control Chart 4.16
Points Within Control Limits 4.17
Points Outside of Control Limits 4.17
Statistical Control Limits and Quality
Tolerances 4.18
Self-Control; Controllability 4.19
Effect on the Process Conformance
Decision 4.20
USE 4.20
The Product Conformance Decision 4.20
Self-Inspection 4.20
The Fitness for Use Decision 4.21
Disposition of Unfit Product 4.22
Corrective Action 4.23
Diagnosis of Sporadic Change 4.23
Corrective Action—Remedy 4.23
Statistical Process Control (SPC) 4.25
The Merits 4.25
The Risks 4.25
Information for Decision Making 4.25
Format of Quality Manuals 4.27
Quality Control Defined. This section describes the quality control process. “Quality control”
is a universal managerial process for conducting operations so as to provide stability—to
prevent adverse change and to “maintain the status quo.”
To maintain stability, the quality control process evaluates actual performance, compares actual
performance to goals, and takes action on the difference.
Quality control is one of the three basic managerial processes through which quality can be
managed. The others are quality planning and quality improvement, which are discussed in
Sections 3 and 5, respectively. The Juran trilogy diagram (Figure 4.1) shows the interrelation of
these processes.
Figure 4.1 is used in several other sections in this handbook to describe the relationships between
quality planning, quality improvement, and quality control and the fundamental managerial processes
in total quality management. What is important for this section is to concentrate on the two “zones of
control.” In Figure 4.1 we can easily see that although the process is in control in the middle of the
chart, we are running the process at an unacceptable level of waste. What is necessary here is not more
control but improvement—actions to change the level of performance.
After the improvements have been made, a new level of performance has been achieved. Now it
is important to establish new controls at this level to prevent the performance level from deteriorating
to the previous level or even worse. This is indicated by the second zone of control.
The term “control of quality” emerged early in the twentieth century (Radford 1917, 1922). The
concept was to broaden the approach to achieving quality, from the then-prevailing after-the-fact
inspection, to what we now call “defect prevention.” For a few decades, the word “control” had a
broad meaning which included the concept of quality planning. Then came events which narrowed
the meaning of “quality control.” The “statistical quality control” movement gave the impression that
quality control consisted of using statistical methods. The “reliability” movement claimed that quality
control applied only to quality at the time of test but not during service life.
In the United States, the term “quality control” now often has the narrow meaning defined previously.
The term “total quality management” (TQM) is now used as the all-embracing term. In
Lessons Learned
Chronic Waste
Cost of Poor Quality
Quality Improvement
New Zone of
Quality Control
Original Zone of
Planning Quality Control (During Operations)
Quality Control
FIGURE 4.1 The Juran trilogy diagram. (Juran Institute, Inc., Wilton, CT.)
Europe, the term “quality control” is also acquiring a narrower meaning. Recently, the European
umbrella quality organization changed its name from European Organization for Quality Control to
European Organization for Quality. In Japan, the term “quality control” retains a broad meaning.
Their “total quality control” is roughly equivalent to our term “total quality management.” In 1997
the Union of Japanese Scientists and Engineers (JUSE) adopted the term total quality management
(TQM) to replace total quality control (TQC) to more closely align themselves with the more common
terminology used in the rest of the world.
The quality control process is one of the steps in the overall quality planning sequence described
in Section 3, The Quality Planning Process, and briefly again in Section 14, Total Quality
Management. Figure 4.2 shows the input-output features of this step.
In Figure 4.2 the input is operating process features developed to produce the product features
required to meet customer needs. The output consists of a system of product and process controls
which can provide stability to the operating process.
The Relation to Quality Assurance. Quality control and quality assurance have much in
common. Each evaluates performance. Each compares performance to goals. Each acts on the difference.
However they also differ from each other. Quality control has as its primary purpose to
maintain control. Performance is evaluated during operations, and performance is compared to goals
during operations. The resulting information is received and used by the operating forces.
Quality assurance’s main purpose is to verify that control is being maintained. Performance is
evaluated after operations, and the resulting information is provided to both the operating forces and
others who have a need to know. Others may include plant, functional, or senior management; corporate
staffs; regulatory bodies; customers; and the general public.
The Feedback Loop. Quality control takes place by use of the feedback loop. A generic form
of the feedback loop is shown in Figure 4.3.
The progression of steps in Figure 4.3 is as follows:
1. A sensor is “plugged in” to evaluate the actual quality of the control subject—the product or
process feature in question. The performance of a process may be determined directly by evaluation
of the process feature, or indirectly by evaluation of the product feature—the product “tells”
on the process.
2. The sensor reports the performance to an umpire.
3. The umpire also receives information on what is the quality goal or standard.
4. The umpire compares actual performance to standard. If the difference is too great, the umpire
energizes an actuator.
5. The actuator stimulates the process (whether human or technological) to change the performance
so as to bring quality into line with the quality goal.
FIGURE 4.2 The input-output diagram for the quality control process.
6. The process responds by restoring conformance.
Note that in Figure 4.3 the elements of the feedback loop are functions. These functions are universal
for all applications, but responsibility for carrying out these functions can vary widely. Much
control is carried out through automated feedback loops. No human beings are involved. Common
examples are the thermostat used to control temperature and the cruise control used in automobiles
to control speed.
Another frequent form of control is self-control carried out by a human being. An example of
such self-control is the village artisan who performs every one of the steps of the feedback loop.
The artisan chooses the control subjects, sets the quality goals, senses what is the actual quality
performance, judges conformance, and becomes the actuator in the event of nonconformance. For
a case example involving numerous artisans producing Steinway pianos, see Lenehan (1982).
Self-directing work teams also perform self-control as is meant here. See Section 15 for a further
discussion of this concept.
This concept of self-control is illustrated in Figure 4.4. The essential elements here are the need for
the worker or work-force team to know what they are expected to do, to know how they are actually
doing, and to have the means to adjust their performance. This implies they have a capable process and
have the tools, skills, and knowledge necessary to make the adjustments and the authority to do so.
A further common form of feedback loop involves office clerks or factory workers whose work
is reviewed by umpires in the form of inspectors. This design of a feedback loop is largely the
result of the Taylor system of separating planning from execution. The Taylor system emerged a
FIGURE 4.3 The generic feedback loop. (Making Quality Happen, Juran
Institute, Inc., senior executive workshop, p. F-3, Wilton, CT.)
FIGURE 4.4 Self-control. (“Quality Control,” Leadership
for the Quality Century, Juran Institute, Inc., senior executive
workshop, p. 5, Wilton, CT.)
century ago and contributed greatly to increasing productivity. However, the effect on quality control
was negative.
The feedback loop is a universal. It is fundamental to any problem in quality control. It applies to all
types of operations, whether in service industries or manufacturing industries, whether for profit or
not. It applies to all levels in the hierarchy, from the chief executive officer to the work force, inclusive.
However, there is wide variation in the nature of the elements of the feedback loop.
In Figure 4.5 a simple flowchart is shown describing the quality control process with the simple
universal feedback loop imbedded.
Choose the Control Subject. Each feature of the product (goods and services) or
process becomes a control subject—a center around which the feedback loop is built. The critical
first step is to choose the control subject. Control subjects are derived from multiple sources
which include:
Stated customer needs for product features
Technological analysis to translate customer needs into product and process features
Process features which directly impact the product features
Industry and government standards
Needs to protect human safety and the environment
Needs to avoid side effects such as irritations to employees or offense to the neighboring community
At the worker level, control subjects consist mainly of product and process features set out in
specifications and procedures manuals. At managerial levels the control subjects are broader and
increasingly business-oriented. Emphasis shifts to customer needs and to competition in the marketplace.
This shift in emphasis then demands added, broader control subjects which, in turn, have an
influence on the remaining steps of the feedback loop.
FIGURE 4.5 The quality control process. (“Quality Control,”
Leadership for the Quality Century, Juran Institute, Inc., senior
executive workshop, p. 2, Wilton, CT.)
Establish Measurement. After choosing the control subject, the next step is to establish the
means of measuring the actual performance of the process or the quality level of the goods or services.
Measurement is one of the most difficult tasks in quality management and is discussed in almost every
section of this handbook, especially in the industry sections. In establishing the measurement we need
to clearly specify the means of measurement (the sensor), the frequency of measurement, the way the
data will be recorded, the format for reporting the data, the analysis to be made on the data to convert
the data to usable information, and who will make the measurement. See Section 9, Measurement,
Information, and Decision-Making, for a thorough discussion of this subject.
Establish Standards of Performance: Product Goals and Process Goals. For
each control subject it is necessary to establish a standard of performance—a quality goal (also
called targets, objectives, etc.). A standard of performance is an aimed-at achievement toward
which effort is expended. Table 4.1 gives some examples of control subjects and the associated
The prime goal for products is to meet customer needs. Industrial customers often specify their
needs with some degree of precision. Such specified needs then become quality goals for the producing
company. In contrast, consumers tend to state their needs in vague terms. Such statements
must then be translated into the language of the producer in order to become product goals.
Other goals for products which are also important are those for reliability and durability.
Whether the products meet these goals can have a critical impact on customer satisfaction and
loyalty and on overall costs. The failures of products under warranty can seriously impact the
profitability of a company through both direct costs and indirect costs (loss of repeat sales, word
of mouth, etc.).
The processes which produce products have two sets of quality goals:
1. To produce products which do meet customer needs. Ideally, each and every unit of product
should meet customer needs.
2. To operate in a stable and predictable manner. In the dialect of the quality specialist, each process
should be “under control.” We will later elaborate on this, under the heading Process
Conformance. These goals may be directly related to the costs of producing the goods or services.
Quality goals may also be established for departments or persons. Performance against such goals
then becomes an input to the company’s reward system. Ideally such goals should be:
Legitimate: They should have undoubted official status.
Measurable: So that they can be communicated with precision.
Attainable: As evidenced by the fact that they have already been attained by others.
Equitable: Attainability should be reasonably alike for individuals with comparable responsibilities.
TABLE 4.1 Examples of Control Subjects and Associated Quality Goals
Control subject Goal
Vehicle mileage Minimum of 25 mi/gal highway driving
Overnight delivery 99.5% delivered prior to 10:30 a.m. next morning
Reliability Fewer than three failures in 25 years of service
Temperature Minimum 505°F; maximum 515°F
Purchase-order error rate No more than 3 errors/1000 purchase orders
Competitive performance Equal or better than top three competitors on six factors
Customer satisfaction 90% or better rate, service outstanding or excellent
Customer retention 95% retention of key customers from year to year
Customer loyalty 100% of market share of over 80% of customers
Quality goals may be set from a combination of the following bases:
Goals for product features and process features are largely based on technological analysis.
Goals for departments and persons should be based on benchmarking rather than historical performance.
For elaboration, see Section 12, Benchmarking.
Quality goals at the highest levels are in the early stages of development. The emerging practice
is to establish goals on matters such as meeting customers’ changing needs, meeting competition,
maintaining a high rate of quality improvement, improving the effectiveness of business processes,
and revising the planning process so as to avoid creating new failure-prone products and processes.
Measure Actual Performance. The critical step in quality control is to measure the actual
performance of the product or the process. To make this measurement we need a sensor, a device to
make the actual measurement.
The Sensor. A “sensor” is a specialized detecting device. It is designed to recognize the presence
and intensity of certain phenomena, and to convert the resulting data into “information.” This
information then becomes the basis of decision making. At lower levels of organization the information
is often on a real-time basis and is used for current control. At higher levels the information
is summarized in various ways to provide broader measures, detect trends, and identify the vital few
The wide variety of control subjects requires a wide variety of sensors. A major category is the
numerous technological instruments used to measure product features and process features. Familiar
examples are thermometers, clocks, yardsticks, and weight scales. Another major category of sensors
is the data systems and associated reports which supply summarized information to the managerial
hierarchy. Yet another category involves the use of human beings as sensors. Questionnaires
and interviews are also forms of sensors.
Sensing for control is done on a huge scale. This has led to the use of computers to aid in the
sensing and in conversion of the resulting data into information. For an example in an office environment
(monitoring in telephone answering centers), see Bylinsky (1991). For an example in a factory
environment (plastic molding), see Umscheid (1991).
Most sensors provide their evaluations in terms of a unit of measure—a defined amount of some
quality feature—which permits evaluation of that feature in numbers. Familiar examples of units of
measure are degrees of temperature, hours, inches, and tons. For a discussion of units of measure,
see Section 9, Measurement, Information, and Decision-Making. A considerable amount of sensing
is done by human beings. Such sensing is subject to numerous sources of error.
Compare to Standards. The act of comparing to standards is often seen as the role of an
umpire. The umpire may be a human being or a technological device. Either way, the umpire may
be called on to carry out any or all of the following activities:
1. Compare the actual quality performance to the quality goal.
2. Interpret the observed difference; determine if there is conformance to the goal.
3. Decide on the action to be taken.
4. Stimulate corrective action.
These activities require elaboration and will shortly be examined more closely.
Take Action on the Difference. In any well-functioning quality control system we need a
means of taking action on the difference between desired standards of performance and actual performance.
We need an actuator. This device (human or technological or both) is the means for stimulating
action to restore conformance. At the worker level it may be a keyboard for giving orders to
an office computer or a calibrated knob for adjusting a machine tool. At the management level it may
be a memorandum to subordinates.
The Process. In all of the preceding discussion we have assumed a process. This may also be
human or technological or both. It is the means for producing the product features, each of which is
a control subject. All work is done by a process which consists of an input, labor, technology, procedures,
energy, materials, and output. For a more complete discussion of process, see Section 6,
Process Management.
The PDCA Cycle. There are many ways of dividing the feedback loop into elements and steps.
Some of them employ more than six elements; others employ fewer than six. A popular example of
the latter is the so-called PDCA cycle (also the Deming wheel) as shown in Figure 4.6. Deming
(1986) referred to this as the Shewhart cycle, which is the name many still use when describing this
version of the feedback loop.
In this example the feedback loop is divided into four steps labeled Plan, Do, Check, and Act.
These steps correspond roughly to the six steps discussed previously:
“Plan” includes choosing control subjects and setting goals.
“Do” includes running the process.
“Check” includes sensing and umpiring.
“Act” includes stimulating the actuator to take corrective action.
An early version of the PDCA cycle was included in W. Edwards Deming’s first lectures in Japan
(Deming 1950). Since then, additional versions have been devised and published. For elaboration,
see Koura (1991).
Some of these versions have attempted to label the PDCA cycle in ways which make it serve as
a universal series of steps for both quality control and quality improvement. The authors feel that this
confuses matters, since two very different processes are involved. (The process for quality improvement
is discussed in Section 5.)
Control subjects run to large numbers, but the number of “things” to be controlled is far larger. These
things include the published catalogs and price lists sent out, multiplied by the number of items in
Study the results.
What did we learn?
What can we predict?
What could be the most important
accomplishments of this team?
What changes might be desirable?
What data are available? Are new
observations needed? If yes, plan
a change or test. Decide how to
use the observations.
Carry out the change or test decided
upon, preferably on a small scale.
Observe the effects
of the change or test.
4. 1.
3. 2.
Step 5. Repeat Step 1, with knowledge accumulated.
Step 6. Repeat Step 2, and onward.
FIGURE 4.6 The PDCA cycle. (Deming, 1986.)
each; the sales made, multiplied by the number of items in each sale; the units of product produced,
multiplied by the associated numbers of quality features; and so on for the numbers of items associated
with employee relations, supplier relations, cost control, inventory control, product and process
development, etc.
A study in one small company employing about 350 people found that there were over a billion
things to be controlled (Juran 1964, pp. 181–182).
There is no possibility for upper managers to control huge numbers of control subjects. Instead,
they divide up the work of control, using a plan of delegation somewhat as depicted in Figure 4.7.
This division of work establishes three areas of responsibility for control: control by nonhuman
means, control by the work force, and control by the managerial hierarchy.
Control by Nonhuman Means. At the base of the pyramid are the automated feedback
loops and error-proofed processes which operate with no human intervention other than maintenance
of facilities (which, however, is critical), These nonhuman methods provide control over the great
majority of things. The control subjects are exclusively technological, and control takes place on a
real-time basis.
The remaining controls in the pyramid require human intervention. By a wide margin, the most
amazing achievement in quality control takes place during a biological process which is millions of
years old—the growth of the fertilized egg into an animal organism. In human beings the genetic
instructions which program this growth consist of a sequence of about three billion “letters.” This
sequence—the human genome—is contained in two strands of DNA (the double helix) which
“unzip” and replicate about a million billion times during the growth process from fertilized egg to
birth of the human being.
Given such huge numbers, the opportunities for error are enormous. (Some errors are harmless,
but others are damaging and even lethal.) Yet the actual error rate is of the order of about one in 10
billion. This incredibly low error rate is achieved through a feedback loop involving three processes
(Radman and Wagner 1988):
A high-fidelity selection process for attaching the right “letters,” using chemical lock-and-key
Control by
Upper Managers
Control by
Control by
the Work Force
Automated Controls
FIGURE 4.7 The pyramid of control. (Making Quality Happen, Juran Institute, Inc.,
senior executive workshop, p. F-5, Wilton, CT.)
A proofreading process for reading the most recent letter, and removing it if incorrect
A corrective action process to rectify the errors which are detected
Control by the Work Force. Delegating such decisions to the work force yields important
benefits in human relations and in conduct of operations. These benefits include shortening the feedback
loop; providing the work force with a greater sense of ownership of the operating processes,
often referred to as “empowerment”; and liberating supervisors and managers to devote more of their
time to planning and improvement.
It is feasible to delegate most quality control decisions to the work force. Many companies
already do. However, to delegate process control decisions requires meeting the criteria of “selfcontrol.”
To delegate product control decisions requires meeting the criteria for “self-inspection.”
(See later in this section under Self-Control and Self-Inspection, respectively.)
Control by the Managerial Hierarchy. The peak of the pyramid of control consists of the
“vital few” control subjects. These are delegated to the various levels in the managerial hierarchy,
including the upper managers.
Managers should avoid getting deeply into making decisions on quality control. Instead, they
Make the vital few decisions.
Provide criteria to distinguish the vital few decisions from the rest. For an example of providing
such criteria see Table 4.3 under the heading: The Fitness for Use Decision.
Delegate the rest under a decision making process which provides the essential tools and
The distinction between vital few matters and others originates with the control subjects. Table
4.2 shows how control subjects at two levels—work force and upper management—affect the elements
of the feedback loop.
Planning for control is the activity which provides the system—the concepts, methodology, and
tools—through which company personnel can keep the operating processes stable and thereby produce
the product features required to meet customer needs. The input-output features of this system
(also plan, process) were depicted in Figure 4.2.
The Customers and Their Needs. The principal customers of quality control systems are
the company personnel engaged in control—those who carry out the steps which form the feedback
loop. Such personnel require (1) an understanding of customers’ quality needs and (2) a definition
TABLE 4.2 Contrast of Quality Control at Two Levels—Work Force and Upper Management
At work force levels At managerial levels
Control goals Product and process Business oriented,
features in specifications product salability,
and procedures competitiveness
Sensors Technological Data systems
Decisions to be made Conformance or not? Meet customer needs or not?
Source: Making Quality Happen, Juran Institute, Inc., senior executive workshop, p. F-4, Wilton, Ct.
of their own role in meeting those needs. However, most of them lack direct contact with customers.
Planning for quality control helps to bridge that gap by supplying a translation of what are customers’
needs, along with defining responsibility for meeting those needs. In this way, planning for
quality control includes providing operating personnel with information on customer needs (whether
direct or translated) and definition of the related control responsibilities of the operating personnel.
Planning for quality control can run into extensive detail. See, for example, Duyck (1989) and Goble
Who Plans? Planning for quality control has in the past been assigned variously to
Staff planners who also plan the operating processes
Staff quality specialists
Multifunctional teams of planners and operating personnel
Departmental managers and supervisors
The work force
Planning for quality control of critical processes has traditionally been the responsibility of those
who plan the operating process. For noncritical processes the responsibility was usually assigned to
quality specialists from the Quality Department. Their draft plans were then submitted to the operating
heads for approval.
Recent trends have been to increase the use of the team concept. The team membership includes
the operating forces and may also include suppliers and customers of the operating process. The
recent trend has also been to increase participation by the work force. For elaboration, see Juran
(1992, pp. 290–291). The concept of self-directing work teams has been greatly expanded in recent
years and includes many of these ideas. See Section 15 for more details on this topic.
Quality Control Concepts. The methodologies of Quality Control are built around various
concepts such as the feedback loop, process capability, self-control, etc. Some of these concepts are
of ancient origin; others have evolved in this century. During the discussion of planning for quality
control, we will elaborate on some of the more widely used concepts.
The Flow Diagram. The usual first step in planning for quality control is to map out the flow
of the operating process. (Design of that process is discussed in Section 3, The Quality Planning
Process.) The tool for mapping is the “flow diagram.” Figure 4.8 is an example of a flow diagram.
(For more examples of this tool, see Appendix V.)
The flow diagram is widely used during planning of quality controls. It helps the planning team to
Understand the overall operating process. Each team member is quite knowledgable about his/her
segment of the process, but less so about other segments and about the interrelationships.
Identify the control subjects around which the feedback loops are to be built. [For an example,
see Siff (1984).] The nature of these control subjects was discussed previously under the heading,
The Control Subject.
Design control stations. (See the following section.)
Control Stations. A “control station” is an area in which quality control takes place. In the
lower levels of organization, a control station is usually confined to a limited physical area.
Alternatively, the control station can take such forms as a patrol beat or a “control tower.” At higher
levels, control stations may be widely dispersed geographically, as is the scope of a manager’s
A review of numerous control stations shows that they are usually designed to provide evaluations
and/or early warnings in the following ways:
At changes of jurisdiction, where responsibility is transferred from one organization to another
Before embarking on some significant irreversible activity such as signing a contract
After creation of a critical quality feature
At the site of dominant process variables
At areas (“windows”) which allow economical evaluation to be made
The flow diagram not only discloses the progression of events in the operating process, it also suggests
which stages should become the centers of control activity. Several of these stages apply to the
majority of operating processes.
Setup (Startup) Control. The end result of this form of control is the decision of whether or
not to “push the start button.” Typically this control involves
A countdown listing the preparatory steps needed to get the process ready to produce. Such
countdowns sometime come from suppliers. Airlines provide checklists to help travelers plan
FIGURE 4.8 The flow diagram.
their trips; electric power companies provide checklists to help householders prepare the house
for winter weather.
Evaluation of process and/or product features to determine whether, if started, the process will
meet the goals.
Criteria to be met by the evaluations.
Verification that the criteria have been met.
Assignment of responsibility. This assignment varies, depending largely on the criticality of
the quality goals. The greater the criticality, the greater is the tendency to assign the verification
to specialists, supervisors and “independent” verifiers rather than to nonsupervisory
Running Control. This form of control takes place periodically during the operation of the
process. The purpose is to make the “run or stop” decision—whether the process should continue to
produce product or whether it should stop.
Running control consists of closing the feedback loop, over and over again. The process and/or
product performance is evaluated and compared with goals. If the product and/or process conforms
to goals, and if the process has not undergone some significant adverse change, the decision is “continue
to run.” If there is nonconformance or if there has been a significant change, then corrective
action is in order.
The term “significant” has meanings beyond those in the dictionary. One of these meanings
relates to whether an indicated change is a real change or is a false alarm due to chance variation.
The design for process control should provide the tools needed to help the operating forces distinguish
between real changes and false alarms. Statistical process control (SPC) methodology is aimed
at providing such tools (see Section 45).
Product Control. This form of control takes place after some amount of product has been produced.
The purpose of the control is to decide whether or not the product conforms to the product
quality goals. Assignment of responsibility for this decision differs from company to company.
However, in all cases those who are to make the decision must be provided with the facilities and
training which will enable them to understand the product quality goals, evaluate the actual product
quality, and decide whether there is conformance.
Since all this involves making a factual decision, it can in theory be delegated to anyone, including
members of the work force. In practice, this delegation is not made to those whose assigned priorities
might bias their judgment. In such cases the delegation is usually to those whose
responsibilities are free from such biases, for example, “independent” inspectors. Statistical quality
control (SQC) is a methodology frequently employed to yield freedom from biases.
Facilities Control. Most operating processes employ physical facilities: equipment, instruments,
and tools. Increasingly the trend has been to use automated processes, computers, robots, etc.
This same trend makes product quality more and more dependent on maintenance of the facilities.
The elements of design for facilities control are well known:
Establish a schedule for conducting facilities maintenance.
Establish a checklist—a list of tasks to be performed during a maintenance action.
Train the maintenance forces to perform the tasks.
Assign clear responsibility for adherence to schedule.
The weakest link in facilities control has been adherence to schedule. To ensure strict adherence
to schedule requires an independent audit.
In cases involving introduction of new technology, a further weak link is training the maintenance
forces (White 1988).
During the 1980s the auto makers began to introduce computers and other electronics into their vehicles.
It soon emerged that many repair shop technicians lacked the technological education base needed
to diagnose and remedy the associated field failures. To make matters worse, the auto makers did not give
high priority to standardizing the computers. As a result a massive training backlog developed.
For an excellent treatise on facilities maintenance, see Nowlan and Heap (1978).
Concept of Dominance. Control subjects are so numerous that planners are well advised to
identify the vital few control subjects so that they will receive appropriate priority. One tool for identifying
the vital few is the concept of dominance.
Operating processes are influenced by many variables, but often one variable is more important
than all the rest combined. Such a variable is said to be the “dominant variable.” Knowledge of which
process variable is dominant helps planners during allocation of resources and priorities. The more
usual dominant variables include:
1. Set-up dominant: Some processes exhibit high stability and reproducibility of results, over
many cycles of operation. A common example is the printing process. The design for control
should provide the operating forces with the means for precise setup and validation before operations
2. Time-dominant: Here the process is known to change progressively with time, for example,
depletion of consumable supplies, heating up, and wear of tools. The design for control should
provide means for periodic evaluation of the effect of progressive change and for convenient
3. Component-dominant: Here the main variable is the quality of the input materials and components.
An example is the assembly of electronic or mechanical equipments. The design for control
should be directed at supplier relations, including joint planning with suppliers to upgrade the
quality of the inputs.
4. Worker-dominant: In these processes, quality depends mainly on the skill and knack possessed
by the workers. The skilled trades are well-known examples. The design for control should
emphasize aptitude testing of workers, training and certification, quality rating of workers, and
error-proofing to reduce worker errors.
5. Information-dominant: Here the processes are of a “job-shop” nature, so that there is frequent
change in what product is to be produced. As a result, the job information changes frequently. The
design for control should concentrate on providing an information system which can deliver accurate,
up-to-date information on just how this job differs from its predecessors.
Seriousness Classification. Another way of identifying the vital few control subjects is through
“seriousness classification.” Under this concept each product feature is classified into one of several
defined classes such as critical, major, and minor. These classifications then guide the planners in allocation
of resources, assignment of priorities, choice of facilities, frequency of inspection and test, etc.
For elaboration, see Section 22, Operations, under Classification of Defects.
Process Capability. One of the most important concepts in the quality planning process is
“process capability.” The prime application of this concept is during planning of the operating
processes. This application is treated in more depth in Section 22, Operations.
This same concept also has application in quality control. To explain this, a brief review is in order.
All operating processes have an inherent uniformity for producing product. This uniformity can often
be quantified, even during the planning stages. The process planners can use the resulting information
for making decisions on adequacy of processes, choice of alternative processes, need for revision of
processes, and so forth, with respect to the inherent uniformity and its relationship to process goals.
Applied to planning for quality control, the state of process capability becomes a major factor in
decisions on frequency of measuring process performance, scheduling maintenance of facilities, etc.
The greater the stability and uniformity of the process, the less the need for frequent measurement
and maintenance.
Those who plan for quality control should have a thorough understanding of the concept of
process capability and its application to both areas of planning—planning the operating processes as
well as planning the controls.
The work of the planners is usually summarized on a control spreadsheet. This spreadsheet is a major
planning tool. An example can be seen in Figure 4.9.
In this spreadsheet the horizontal rows are the various control subjects. The vertical columns consist
of elements of the feedback loop plus other features needed by the operating forces to exercise
control so as to meet the quality goals.
Some of the contents of the vertical columns are unique to specific control subjects. However,
certain vertical columns apply widely to many control subjects. These include unit of measure, type
of sensor, quality goal, frequency of measurement, sample size, criteria for decision-making, and
responsibility for decision making.
Who Does What? The feedback loop involves multiple tasks, each of which requires a clear
assignment of responsibility. At any control station there may be multiple people available to perform
those tasks. For example, at the work-force level, a control station may include setup specialists,
operators, maintenance personnel, inspectors, etc. In such cases it is necessary to agree on who
should make which decisions and who should take which actions. An aid to reaching such agreement
is a special spreadsheet similar to Figure 4.9.
In this spreadsheet the essential decisions and actions are listed in the left-hand column. The
remaining columns are headed up by the names of the job categories associated with the control
station. Then, through discussion among the cognizant personnel, agreement is reached on who is
to do what.
The spreadsheet (Figure 4.9) is a proven way to find answers to the long-standing, but vague,
question, “Who is responsible for quality?” This question has never been answered because it is
FIGURE 4.9 Spreadsheet for “Who does what?” (Making Quality Happen, Juran Institute, Inc., senior executive workshop, p. F-8,
Wilton, CT.)
Wave solder
Solder temperature
Degree F
( F)
505 F Continuous N/A 510 F
reduce heat
500 F
increase heat
5 ft/min
reduce speed
4 ft/min
increase speed
Feet per
Conveyor speed N/A 1/hour 4.5 ft/min ft/min
At 1.5%, drain
bath, replace
% Total
Alloy purity 15 grams 1/month 1.5% max
inherently unanswerable. However if the question is restated in terms of decisions and actions, the
way is open to agree on the answers. This clears up the vagueness.
Does the process conform to its quality goals? The umpire answers this question by interpreting the
observed difference between process performance and process goals. When current performance
does differ from the quality goals, the question arises: What is the cause of this difference?
Special and Common Causes of Variation. Observed differences usually originate in
one of two ways: (1) the observed change is caused by the behavior of a major variable in the process
(or by the entry of a new major variable) or (2) the observed change is caused by the interplay of
multiple minor variables in the process.
Shewhart called (1) and (2) “assignable” and “nonassignable” causes of variation, respectively
(Shewhart 1931). Deming later coined the terms “special” and “common” causes of variation
(Deming 1986). In what follows we will use Deming’s terminology.
“Special” causes are typically sporadic, and often have their origin in single variables. For such
cases it is comparatively easy to conduct a diagnosis and provide remedies. “Common” causes are
typically chronic and usually have their origin in the interplay among multiple minor variables, As a
result, it is difficult to diagnose them and to provide remedies. This contrast makes clear the importance
of distinguishing special causes from common causes when interpreting differences. The need
for making such distinctions is widespread. Special causes are the subject of quality control; common
causes are the subject of quality improvement.
The Shewhart Control Chart. It is most desirable to provide umpires with tools which can
help to distinguish between special causes and common causes. An elegant tool for this purpose is
the Shewhart control chart (or just control chart) shown in Figure 4.10.
In Figure 4.10 the horizontal scale is time, and the the vertical scale is quality performance. The
plotted points show quality performance as time progresses.
The chart also exhibits three horizontal lines. The middle line is the average of past performance
and is therefore the expected level of performance. The other two lines are statistical “limit lines.”
FIGURE 4.10 The Shewhart control chart. (“Quality Control,” Leadership for the Quality
Century, Juran Institute, Inc., senior executive workshop, p. 4, Wilton, CT.)
Limit (UCL)
of All
Limit (LCL)
Number of Samples Average Wall Thickness
of Samples
They are intended to separate special causes from common causes, based on some chosen level of
odds, such as 20 to 1.
Points Within Control Limits. Point A on the chart differs from the historical average.
However, since point A is within the limit lines, this difference could be due to common causes (at
odds of less than 20 to 1.) Hence we assume that there is no special cause.
In the absence of special causes, the prevailing assumptions include:
Only common causes are present.
The process is in a state of “statistical control.”
The process is doing the best it can.
The variations must be endured.
No action need be taken—taking action may make matters worse (a phenomenon known as
“hunting” or “tampering.”
The preceding assumptions are being challenged by a broad movement to improve process uniformity.
Some processes exhibit no points outside of control chart limits, yet the interplay of minor
variables produces some defects.
In one example, a process in statistical control was nevertheless improved by an order of magnitude.
The improvement was by a multifunctional improvement team which identified and addressed
some of the minor variables. This example is a challenge to the traditional assumption that variations
due to common causes must be endured (Pyzdek 1990).
In other cases the challenge is more subtle. There are again no points outside the control limits,
but in addition, no defects are being produced. Nevertheless the customers demand greater and
greater uniformity. Examples are found in business processes (precision of estimating), as well as in
manufacture (batch-to-batch uniformity of chemicals, uniformity of components going into random
assembly). Such customer demands are on the increase, and they force suppliers to undertake projects
to improve the uniformity of even the minor variables in the process. There are many types of
control charts. See Section 45, Statistical Process Control, for a more detailed discussion of this
important tool.
Points Outside of Control Limits. Point B also differs from the historical average, but is
outside of the limit lines. Now the odds are heavily against this being due to common causes—over
20 to 1. Hence we assume that point B is the result of special causes. Traditionally such “out-of-control”
points became nominations for corrective action.
Ideally all such nominations should stimulate prompt corrective action to restore the status quo.
In practice many out-of-control changes do not result in corrective action. The usual reason is that
the changes involving special causes are too numerous—the available personnel cannot deal with all
of them. Hence priorities are established based on economic significance or on other criteria of
importance. Corrective action is taken for the high-priority cases; the rest must wait their turn. Some
changes at low levels of priority may wait a long time for corrective action.
A further reason for failure to take corrective action is a lingering confusion between statistical
control limits and quality tolerances. It is easy to be carried away by the elegance and sensitivity of
the control chart. This happened on a large scale during the 1940s and 1950s. Here are two examples
from the personal experience of the one of the authors:
A large automotive components factory placed a control chart at every machine.
A viscose yarn factory created a “war room” of over 400 control charts.
In virtually all such cases the charts were maintained by the quality departments but ignored by
the operating personnel. Experience with such excesses has led managers and planners to be wary of
employing control charts just because they are sensitive detectors of change. Instead, the charts
should be justified based on value added. Such justifications include:
Customer needs are directly involved.
There is risk to human safety or the environment.
Substantial economics are at stake.
The added precision is needed for control.
Statistical Control Limits and Quality Tolerances. For most of human history quality
goals consisted of product features or process features, usually defined in words. The growth of technology
then stimulated the growth of measurement plus a trend to define quality goals in numbers.
In addition, there emerged the concept of limits or “tolerances” around the goals. For example:
At least 95 percent of the shipments shall meet the scheduled delivery date.
The length of the bar shall be within 1 mm of the specified number.
Such quality goals had official status. They were set by product or process designers, and published
as official specifications. The designers were the official quality legislators—they enacted the
laws. Operating personnel were responsible for obeying the quality laws—meeting the specified
goals and tolerances.
Statistical control limits in the form of control charts were virtually unknown until the 1940s. At
that time, these charts lacked official status. They were prepared and published by quality specialists
from the Quality Department. To the operating forces, control charts were a mysterious, alien concept.
In addition, the charts threatened to create added work in the form of unnecessary corrective
action. The operating personnel reasoned as follows: It has always been our responsibility to take
corrective action whenever the product becomes nonconforming. These charts are so sensitive that
they detect process changes which do not result in nonconforming product. We are then asked to take
corrective action even when the products meet the quality goals and tolerances.
So there emerged a confusion of responsibility. The quality specialists were convinced that the
control charts provided useful early-warning signals which should not be ignored. Yet the quality
departments failed to recognize that the operating forces were now faced with a confusion of responsibility.
The latter felt that so long as the products met the quality goals there was no need for corrective
action. The upper managers of those days were of no help—they did not involve themselves
in such matters. Since the control charts lacked official status, the operating forces solved their problem
by ignoring the charts. This contributed to the collapse, in the 1950s, of the movement known
as “statistical quality control.”
The 1980s created a new wave of interest in applying the tools of statistics to the control of quality.
Many operating personnel underwent training in “statistical process control.” This training
helped to reduce the confusion, but some confusion remains. To get rid of the confusion, managers
Clarify the responsibility for corrective action on points outside the control limits. Is this action
mandated or is it discretionary?
Establish guidelines on action to be taken when points are outside the statistical control limits but
the product still meets the quality tolerances.
The need for guidelines for decision making is evident from Figure 4.11. The guidelines for quadrants
A and C are obvious. If both process and product conform to their respective goals, the process
may continue to run. If neither process nor product conform to their respective goals, the process
should be stopped, and remedial action should be taken. The guidelines for quadrants B and D are
often vague, and this vagueness has been the source of a good deal of confusion. If the choice of
action is delegated to the work force, the managers should establish clear guidelines.
Numerous efforts have been made to design control chart limits in ways which help operating
personnel to detect whether product quality is threatening to exceed the product quality limits. For a
recent example, see Carr (1989). Another approach, based on product quality related to product quality
limits, is “PRE-Control.” See Juran (1988, pp. 24.31–24.38).
Self-Control; Controllability. Workers are in a state of self-control when they have been
provided with all the essentials for doing good work. These essentials include:
Means of knowing what are the quality goals.
Means of knowing what is their actual performance.
Means for changing their performance in the event that performance does not conform to goals.
To meet this criterion requires an operating process which (1) is inherently capable of meeting
the goals and (2) is provided with features which make it possible for the operating forces to
adjust the process as needed to bring it into conformance with the goals.
These criteria for self-control are applicable to processes in all functions and all levels, from general
manager to nonsupervisory worker.
It is all too easy for managers to conclude that the above criteria have been met. In practice, there
are many details to be worked out before the criteria can be met. The nature of these details is evident
from checklists which have been prepared for specific processes in order to ensure meeting the
criteria for self-control. Examples of these checklists include those designed for product designers,
production workers, and administrative and support personnel. Examples of such checklists can be
found by referring to the subject index of this handbook.
If all the criteria for self-control have been met at the worker level, any resulting product nonconformances
are said to be worker-controllable. If any of the criteria for self-control have not been
met, then management’s planning has been incomplete—the planning has not fully provided the
means for carrying out the activities within the feedback loop. The nonconforming products resulting
from such deficient planning are then said to be management-controllable. In such cases it is
risky for managers to hold the workers responsible for quality.
Responsibility for results should, of course, be keyed to controllability. However, in the past
many managers were not aware of the extent of controllability as it prevailed at the worker level.
Studies conducted by Juran during the 1930s and 1940s showed that at the worker level the proportion
of management-controllable to worker-controllable nonconformances was of the order of 80 to
20. These findings were confirmed by other studies during the 1950s and 1960s. That ratio of 80 to
20 helps to explain the failure of so many efforts to solve the companies’ quality problems solely by
motivating the work force.
FIGURE 4.11 Example of areas of decision making. (Making Quality Happen, Juran Institute, Inc., senior executive
workshop, p.F-21, Wilton, CT.)
Effect on the Process Conformance Decision. Ideally the decision of whether the
process conforms to process quality goals should be made by the work force. There is no shorter
feedback loop. For many processes this is the actual arrangement. In other cases the process conformance
decision is assigned to nonoperating personnel—independent checkers or inspectors. The reasons
The worker is not in a state of self-control.
The process is critical to human safety or to the environment.
Quality does not have top priority.
There is lack of mutual trust between the managers and the work force.
There are two levels of product features, and they serve different purposes. One of these levels serves
such purposes as:
Meeting customer needs
Protecting human safety
Protecting the environment
Product features are said to possess “fitness for use” if they are able to serve the above purposes.
The second level of product features serves purposes such as:
Providing working criteria to those who lack knowledge of fitness for use
Creating an atmosphere of law and order
Protecting innocents from unwarranted blame
Such product features are typically contained in internal specifications, procedures, standards,
etc. Product features which are able to serve the second list of purposes are said to possess conformance
to specifications, etc. We will use the shorter label “conformance.”
The presence of two levels of product features results in two levels of decision making: Is the
product in conformance? Is the product fit for use? Figure 4.12 shows the interrelation of these decisions
to the flow diagram.
The Product Conformance Decision. Under prevailing policies, products which conform
to specification are sent on to the next destination or customer. The assumption is that products
which conform to specification are also fit for use. This assumption is valid in the great majority of
The combination of large numbers of product features when multiplied by large volumes of product
creates huge numbers of product conformance decisions to be made. Ideally these decisions
should be delegated to the lowest levels of organization—to the automated devices and the operating
work force. Delegation of this decision to the work force creates what is called “self-inspection.”
Self-Inspection. We define “self-inspection” as a state in which decisions on the product are
delegated to the work force. The delegated decisions consist mainly of: Does product quality conform
to the quality goals? What disposition is to be made of the product?
Note that self-inspection is very different from self-control, which involves decisions on the
The merits of self-inspection are considerable:
The feedback loop is short; the feedback often goes directly to the actuator—the energizer for
corrective action.
Self-inspection enlarges the job of the work force—it confers a greater sense of job ownership.
Self-inspection removes the police atmosphere created by use of inspectors, checkers, etc.
However, to make use of self-inspection requires meeting several essential criteria:
Quality is number one: Quality must have undoubted top priority.
Mutual confidence: The managers must have enough trust in the work force to be willing to
make the delegation, and the work force must have enough confidence in the managers to be willing
to accept the responsibility.
Self-control: The conditions for self-control should be in place so that the work force has all the
means necessary to do good work.
Training: The workers should be trained to make the product conformance decisions.
Certification: The recent trend is to include a certification procedure. Workers who are candidates
for self-inspection undergo examinations to ensure that they are qualified to make good
decisions. The successful candidates are certified and may be subject to audit of decisions thereafter.
For examples, see Nowak (1991, Military Airlift Command) and Pearl (1988, Corning
Glass Works).
In many companies these criteria are not fully met, especially the criterion of priority. If some
parameter other than quality has top priority, there is a real risk that evaluation of product conformance
will be biased. This problem happens frequently when personal performance goals are in conflict
with overall quality goals. For example, a chemical company found that it was rewarding sales
personnel on revenue targets without regard to product availability or even profitability. The sales people
were making all their goals, but the company was struggling.
The Fitness for Use Decision. The great majority of products do conform to specifications.
For the nonconforming products there arises a new question: Is the nonconforming product nevertheless
fit for use?
FIGURE 4.12 Flow diagram of decisions on conformance and fitness for use.
Not OK
A complete basis for making this decision requires answers to questions such as:
Who will be the user(s)?
How will this product be used?
Are there risks to structural integrity, human safety, or the environment?
What is the urgency for delivery?
How do the alternatives affect the producer’s and the user’s economics?
To answer such questions can involve considerable effort. Companies have tried to minimize the
effort through procedural guidelines. The methods in use include:
Treat all nonconforming product as unfit for use: This approach is widely used for products
which can pose risks to human safety or the environment—products such as pharmaceuticals or
nuclear energy.
Create a mechanism for decision making: An example is the Material Review Board so
widely used in the defense industry. This device is practical for matters of importance, but is
rather elaborate for the more numerous cases in which little is at stake.
Create a system of multiple delegation: Under such a system, the “vital few” decisions are
reserved for a formal decision-making body such as a Material Review Board. The rest are delegated
to other people.
Table 4.3 is an example of a table of delegation used by a specific company. (Personal communication
to one of the authors.)
For additional discussion on the fitness-for-use decision, see Juran (1988, pp. 18.32–18.36).
Disposition of Unfit Product. Unfit product is disposed of in various ways: scrap, sort,
rework, return to supplier, sell at a discount, etc. The internal costs can be estimated to arrive at an
economic optimum. However, the effects go beyond money: schedules are disrupted, people are
blamed, etc. To minimize the resulting human abrasion, some companies have established rules of
conduct such as:
TABLE 4.3 Multiple Delegations of Decision Making on Fitness for Use*
Amount of product or money at stake is
Effect of nonconformance is on Small Large
Internal economics only Department head directly involved, Plant managers involved,
quality engineer quality manager
Economic relations Supplier, purchasing agent, Supplier, manager
with supplier quality engineer
Economic relations with client Client, salesperson, quality engineer Client: for Marketing,
Manufacturing, Technical,
Field performance of the product Product designer, salesperson, Client: managers for
quality engineer Technical, Manufacturing,
Marketing Quality
Risk of damage to society or Product design manager, General manager and team of
of nonconformance to compliance officer, lawyer, upper managers
government regulations quality managers
*For those industries whose quality mission is really one of conformance to specification (for example, atomic energy,
space), the real decision maker on fitness for use is the client or the government regulator.
Choose that alternative which minimizes the total loss to all parties involved. Now there is less
to argue about, and it becomes easier to agree on how to share the loss.
Avoid looking for blame. Instead, treat the loss as an opportunity for quality improvement.
Use “charge backs” sparingly. Charging the vital few losses to the departments responsible has
merit from an accounting viewpoint. However, when applied to the numerous minor losses, this
is often uneconomic as well as detrimental to efforts to improve quality.
Failure to use products which meet customer needs is a waste. Sending out products which do not
meet customer needs is worse. Personnel who are assigned to make product conformance decisions
should be provided with clear definitions of responsibility as well as guidelines for decision making.
Managers should, as part of their audit, ensure that the processes for making product conformance
decisions are appropriate to company needs.
Corrective Action. The final step in closing the feedback loop is to actuate a change which
restores conformance with quality goals. This step is popularly known as “troubleshooting” or “firefighting.”
Note that the term “corrective action” has been applied loosely to two very different situations,
as shown in Figure 4.1. The feedback loop is well designed to eliminate sporadic nonconformance
like that “spike” in Figure 4.1; the feedback loop is not well designed to deal with the area of
chronic waste shown in the figure. Instead, the need is to employ the quality improvement process
of Section 5.
We will use the term “corrective action” in the sense of troubleshooting—eliminating sporadic
Corrective action requires the journeys of diagnosis and remedy. These journeys are simpler than
for quality improvement. Sporadic problems are the result of adverse change, so the diagnostic journey
aims to discover what has changed. The remedial journey aims to remove the adverse change
and restore conformance.
Diagnosis of Sporadic Change. During the diagnostic journey the focus is on “What has
changed.” Sometimes the causes are not obvious, so the main obstacle to corrective action is diagnosis.
The diagnosis makes use of methods and tools such as:
Autopsies to determine with precision the symptoms exhibited by the product and process.
Comparison of products made before and after the trouble began to see what has changed; also
comparison of good and bad products made since the trouble began.
Comparison of process data before and after the problem began to see what process conditions
have changed.
Reconstruction of the chronology, which consists of logging on a time scale (of hours, days,
etc.): (1) the events which took place in the process before and after the sporadic change, that
is, rotation of shifts, new employees on the job, maintenance actions, etc., and (2) the timerelated
product information, that is, date codes, cycle time for processing, waiting time, move
dates, etc.
Analysis of the resulting data usually sheds a good deal of light on the validity of the various theories
of causes. Certain theories are denied. Other theories survive to be tested further.
Operating personnel who lack the training needed to conduct such diagnoses may be forced to
shut down the process and request assistance from specialists, the maintenance department, etc. They
may also run the process “as is” in order to meet schedules and thereby risk failure to meet the quality
Corrective Action—Remedy. Once the cause(s) of the sporadic change is known, the worst
is over. Most remedies consist of going back to what was done before. This is a return to the famil-
iar, not a journey into the unknown (as is the case with chronic problems). The local personnel are
usually able to take the necessary action to restore the status quo.
Process designs should provide means to adjust the process as required to attain conformance
with quality goals. Such adjustments are needed at start-up and during running of the process. This
aspect of design for process control ideally should meet the following criteria:
There should be a known relationship between the process variables and the product results.
Means should be provided for ready adjustment of the process settings for the key process variables.
A predictable relationship should exist between the amount of change in the process settings and
the amount of effect on the product features.
If such criteria are not met, the operating personnel will, in due course, be forced to cut and try
in order to carry out remedial action. The resulting frustrations become a disincentive to putting high
priority on quality. Burgam (1985) found:
In one foundry an automated process design for controlling the amount of metal poured failed to provide
adequate regulation. As a result, human regulation took over. The workers then played safe by overpouring,
since underpoured castings had to be scrapped. The result was much waste until a new
technology solved the problem.
Some companies provide systematic procedures for dealing with sporadic changes. See, for
example, Sandorf and Bassett (1993).
For added discussion on troubleshooting, see Section 22.
An essential activity within the feedback loop is the collection and analysis of data. This activity falls
within the scientific discipline known as “statistics.” The methods and tools used are often called
“statistical methods.” These methods have long been used to aid in data collection and analysis in
many fields: biology, government, economics, finance, management, etc. Section 44 contains a thorough
discussion of the basic statistical methods, while Section 45 contains a good discussion on
those methods used in statistical process control.
During this century, much has happened to apply statistical methodology to quality-oriented
problems. This has included development of special tools such as the Shewhart control chart. An
early wave of such application took place during the 1920s, largely within the Bell System. A second
and broader wave was generated during the 1940s and 1950s. It came to be known as statistical
quality control. A third wave, broader still, emerged during the 1980s, and came to be widely known
as statistical process control. This is covered in Section 45.
Statistical Process Control (SPC). The term has multiple meanings, but in most companies
it is considered to include basic data collection; analysis through such tools as frequency distributions,
Pareto principle, Ishikawa (fish bone) diagram, Shewhart control chart, etc.; and application
of the concept of process capability.
Advanced tools, such as design of experiments and analysis of variance (see Section 47), are a
part of statistical methods but are not normally considered to be a part of statistical process control.
The Merits. These statistical methods and tools have contributed in an important way to quality
control and also to the other processes of the Juran trilogy—quality improvement and quality planning.
For some types of quality problems the statistical tools are more than useful—the problems cannot be
solved at all without using the appropriate statistical tools.
The SPC movement has succeeded in training a great many supervisors and workers in basic statistical
tools. The resulting increase in statistical literacy has made it possible for them to improve
their grasp of the behavior of processes and products. In addition, many have learned that decisions
based on data collection and analysis yield superior results.
The Risks. There is danger in taking a tool-oriented approach to quality instead of a problemoriented
or results-oriented approach. During the 1950s this preoccupation became so extensive that
the entire statistical quality control movement collapsed; the word “statistical” had to be eliminated
from the names of the departments.
The proper sequence in managing is first to establish goals and then to plan how to meet those
goals, including choice of the appropriate tools. Similarly, when dealing with problems—threats or
opportunities—experienced managers start by first identifying the problems. They then try to solve
those problems by various means, including choice of the proper tools.
During the 1980s, numerous companies did, in fact, try a tool-oriented approach by training large
numbers of their personnel in the use of statistical tools. However, there was no significant effect on
the “bottom line.” The reason was that no infrastructure had been created to identify which projects
to tackle, to assign clear responsibility for tackling those projects, to provide needed resources, to
review progress, etc.
Managers should ensure that training in statistical tools does not become an end in itself. One form
of such assurance is through measures of progress. These measures should be designed to evaluate the
effect on operations, such as improvement in customer satisfaction or product performance, reduction
in cost of poor quality, etc. Measures such as numbers of courses held, or numbers of people trained,
do not evaluate the effect on operations and hence should be regarded as subsidiary in nature.
Information for Decision Making. Quality control requires extensive decision-making.
These decisions cover a wide variety of subject matter and take place at all levels of the hierarchy.
The planning for quality control should provide an information network which can serve all
decision makers. At some levels of the hierarchy, a major need is for real-time information to permit
prompt detection and correction of nonconformance to goals. At other levels, the emphasis is
on summaries which enable managers to exercise control over the vital few control subjects (see
Sections 9 and 34). In addition the network should provide information as needed to detect major
trends, identify threats and opportunities, and evaluate performance of organization units and
In some companies the quality information system is designed to go beyond control of product
features and process features; the system is also used to control the quality performance of organizations
and individuals, for example, departments and department heads. For example, many companies
prepare and regularly publish scoreboards showing summarized quality performance data for
various market areas, product lines, operating functions, etc. These performance data are often used
as indicators of the quality performance of the personnel in charge.
To provide information which can serve all those purposes requires planning which is directed
specifically to the information system. Such planning is best done by a multifunctional team whose
mission is focused on the quality information system. That team properly includes the customers as
well as the suppliers of information. The management audit of the quality control system should
include assurance that the quality information system meets the needs of the various customers. (For
additional discussion relating to the quality information system, see Section 9, Measurement,
Information, and Decision Making.)
A great deal of quality planning is done through “procedures” which are really repetitive-use plans.
Such procedures are thought out, written out, and approved formally. Once published, they become
the authorized ways of conducting the company’s affairs. It is quite common for the procedures relating
to managing for quality to be published collectively in a “quality manual” (or similar title). A
significant part of the manual relates to quality control.
Quality manuals add to the usefulness of procedures in several ways:
Legitimacy: The manuals are approved at the highest levels of organization.
Readily findable: The procedures are assembled into a well-known reference source rather than
being scattered among many memoranda, oral agreements, reports, minutes, etc.
Stable: The procedures survive despite lapses in memory and employee turnover.
Study of company quality manuals shows that most of them contain a core content which is quite
similar from company to company. Relative to quality control, this core content includes procedures for:
Application of the feedback loop to process and product control
Ensuring that operating processes are capable of meeting the quality goals
Maintenance of facilities and calibration of measuring instruments
Relations with suppliers on quality matters
Collection and analysis of the data required for the quality information system
Training the personnel to carry out the provisions of the manual
Audit to ensure adherence to procedures
The need for repetitive-use quality control systems has led to evolution of standards at industry,
national, and international levels. For elaboration, see Section 11, The ISO 9000 Family of
International Standards. For an example of developing standard operating procedures, including the use
of videocassettes, see Murphy and McNealey (1990). Work-force participation during preparation of
procedures helps to ensure that the procedures will be followed. See, in this connection, Gass (1993).
Format of Quality Manuals. Here again, there is much commonality. The general sections
of the manual include:
1. An official statement by the general manager. It includes the signatures which confer legitimacy.
2. The purpose of the manual and how to use it.
3 The pertinent company (or divisional, etc.) quality policies.
4 The organizational charts and tables of responsibility relative to the quality function.
5. Provision for audit of performance against the mandates of the manual.
Additional sections of the manual deal with applications to functional departments, technological
products and processes, business processes, etc. For elaboration, see Juran (1988, pp. 6.40–6.47).
Managers are able to influence the adequacy of the Quality Control manual in several ways:
Participate in defining the criteria to be met by the manual.
Approve the final draft of the manual to make it official.
Periodically audit the up-to-dateness of the manual as well as conformance to the manual.
An important influence on Quality Control is the extent to which the reward system (merit rating,
etc.) emphasizes quality in relation to other parameters. This aspect of quality control is discussed
throughout Section 15. See also Section 40, under Motivation, Recognition, and Reward.
Experience has shown that control systems are subject to “slippage” of all sorts. Personnel turnover
may result in loss of essential knowledge. Entry of unanticipated changes may result in obsolescence.
Shortcuts and misuse may gradually undermine the system until it is no longer effective.
The major tool for guarding against deterioration of a control system has been the audit. Under
the audit concept a periodic, independent review is established to provide answers to the following
questions: Is the control system still adequate for the job? Is the system is being followed?
The answers are obviously useful to the operating managers. However, that is not the only purpose
of the audit. A further purpose is to provide those answers to people who, though not directly
involved in operations, nevertheless have a need to know. If quality is to have top priority, those who
have a need to know include the upper managers.
It follows that one of the responsibilities of managers is to mandate establishment of a periodic
audit of the quality control system.
Recent decades have witnessed a growing trend to improve the effectiveness of quality control by
formal adoption of modern concepts, methodologies, and tools. These have included:
Systematic planning for quality control, with extensive participation by the operating personnel
Formal application of the feedback loop, and establishment of clear responsibility for the associated
decisions and actions
Delegation of decisions to the work force through self-control and self-inspection
Wide application of statistical process control and the associated training of the operating personnel
A structured information network to provide a factual basis for decision making
A systematic process for corrective action in the event of sporadic adverse change
Formal company manuals for quality control, with periodic audits to ensure up-to-dateness and
The quality control process is a universal managerial process for conducting operations so as to provide
stability—to prevent adverse change and to “maintain the status quo.” Quality control takes
place by use of the feedback loop. Each feature of the product or process becomes a control subject—
a center around which the feedback loop is built. As much as possible, human control should
be done by the work force—the office clerical force, factory workers, salespersons, etc. The flow diagram
is widely used during the planning of quality controls. The weakest link in facilities control has
been adherence to schedule. To ensure strict adherence to schedule requires an independent audit.
Knowing which process variable is dominant helps planners during allocation of resources and priorities.
The work of the planners is usually summarized on a control spreadsheet. This spreadsheet
is a major planning tool.
The question “Who is responsible for quality?” is inherently unanswerable. However, if the question
is restated in terms of decisions and actions, the way is open to agree on the answers. The design
for process control should provide the tools needed to help the operating forces distinguish between
real changes and false alarms. It is most desirable to provide umpires with tools which can help to
distinguish between special causes and common causes. An elegant tool for this purpose is the
Shewhart control chart (or just control chart). The criteria for self-control are applicable to processes
in all functions, and all levels, from general manager to nonsupervisory worker. Responsibility for
results should be keyed to controllability. Ideally the decision of whether the process conforms to
process quality goals should be made by the work force. There is no shorter feedback loop.
To make use of self-inspection requires meeting several essential criteria: quality is number one;
mutual confidence, self-control, training, and certification are the others. Personnel who are assigned
to make product conformance decisions should be provided with clear definitions of responsibility
as well as guidelines for decision making. The proper sequence in managing is first to establish goals
and then to plan how to meet those goals, including the choice of the appropriate tools. The planning
for quality control should provide an information network which can serve all decision makers.
Managers should avoid getting deeply into making decisions on quality control. They should make
the vital few decisions, provide criteria to distinguish the vital few from the rest, and delegate the
rest under a decision-making process.
To eliminate the confusion relative to control limits and product quality tolerance, managers
should clarify the responsibility for corrective action on points outside the control limits and establish
guidelines on action to be taken when points are outside the statistical control limits but the product
still meets the quality tolerances.
Managers should, as part of their audit, ensure that the processes for making product conformance
decisions are appropriate to company needs. They should also ensure that training in statistical
tools does not become an end in itself. The management audit of the quality control
system should include assurance that the quality information system meets the needs of the various
Managers are able to influence the adequacy of the quality control manual in several ways: participate
in defining the criteria to be met, approve the final draft to make it official, and periodically
audit the up-to-dateness of the manual as well as the state of conformance.
Burgam, Patrick M. (1985). “Application: Reducing Foundry Waste.” Manufacturing Engineering, March.
Bylinsky, Gene (1991). “How Companies Spy on Employees.” Fortune, November, pp. 131–140.
Carr, Wendell E. (1989). “Modified Control Limits.” Quality Progress, January, pp. 44–48.
Deming, W. Edwards (1986). Out of the Crisis, MIT Center for Advanced Engineering Study. Cambridge, MA.
Deming,W. Edwards (1950). Elementary Principles of the Statistical Control of Quality, Nippon Kagaku Gijutsu
Renmei (Japanese Union of Scientists and Engineers), Tokyo.
Duyck, T. O. (1989). “Product Control Through Process Control and Process Design.” 1989 ASQC Quality
Congress Transactions, pp. 676–681.
Gass, Kenneth C. (1993). “Getting the Most Out of Procedures.” Quality Engineering, June.
Goble, Joann (1987). “A Systematic Approach to Implementing SPC,” 1987 ASQC Quality Congress
Transactions, pp. 154–164.
Juran, J. M. (1992). Juran on Quality by Design, The Free Press, A Division of Macmillan, Inc., New York.
Juran, J. M., ed. (1988). Juran’s Quality Control Handbook, McGraw-Hill, New York.
Juran, J. M. (1964). Managerial Breakthrough, McGraw-Hill, New York.
Koura, Kozo (1991). “Deming Cycle to Management Cycle.” Societas Qualitas, Japanese Union of Scientists and
Engineers, Tokyo, May–June.
Lenehan, Michael (1982). “The Quality of the Instrument, Building Steinway Grand Piano K 2571.” The Atlantic
Monthly, August, pp. 32–58.
Murphy, Robert W., and McNealey, James E. (1990). “A Technique for Developing Standard Operating
Procedures to Provide Consistent Quality.” 1990 Juran IMPRO Conference Proceedings, pp. 3D1–3D6.
Nowlan, F. Stanley, and Heap, Howard F. Reliability Centered Maintenance. Catalog number ADA 066579,
National Technical Information Service, Springfield, Va.
Nowak, Major General John M. (1991). “Transitioning from Correction to Prevention.” 1991 Juran IMPRO
Conference., pp. 1-15–1-20.
Pearl, Daniel H. (1988). “Operator Owned SPC, How to Get it and Keep It.” 1988 ASQC Congress Transactions,
pp. 817–821.
Pyzdek, Thomas (1990). “There’s No Such Thing as a Common Cause.” ASQC Quality Congress Transactions,
pp. 102–108.
Radford, G. S. (1922). The Control of Quality of Manufacturing, Ronald Press Company, New York.
Radford, G. S. (1917). “The Control of Quality.” Industrial Management, vol. 54, p. 100.
Radman, Miroslav, and Wagner, Robert (1988). “The High Fidelity of DNA Duplication.” Scientific American,
August, pp. 40–46.
Sandorf, John P., and Bassett, A. Thomas III (1993). “The OCAP: Predetermined Responses to Out-of-Control
Conditions.” Quality Progress, May, pp. 91–95.
Shainin, Dorian, and Shainin, Peter D. (1988). “PRE-Control.” Juran’s Quality Control Handbook, 4th ed.,
McGraw-Hill, New York.
Siff, Walter C. (1984). “The Strategic Plan of Control—A Tool For Participative Management.” 1984 ASQC
Quality Congress Transactions, pp. 384–390.
Shewhart,W. A. (1931). Economic Control of Quality of Manufactured Product, Van Nostrand, New York, 1931.
Reprinted by ASQC, Milwaukee, 1980.
Umscheid, Thomas E. (1991). “Production Monitoring Enhances SPC.” Quality Progress, December, pp. 58–61.
White, Joseph B. (1988). “Auto Mechanics Struggle to Cope With Technology in Today’s Cars.” The Wall Street
Journal, July 26, p. 37.
J. M. Juran1
Two Kinds of Beneficial Change 5.3
Structured Product Development 5.3
Unstructured Reduction of Chronic Waste
Lessons Learned 5.5
The Rate of Improvement Is Decisive 5.5
The Emerging Consensus 5.7
Improvement Distinguished from Control
All Improvement Takes Place Project by
Project 5.8
Quality Improvement Is Applicable
Universally 5.8
Quality Improvement Extends to All
Parameters 5.9
The Backlog of Improvement Projects Is
Huge 5.10
Quality Improvement Does Not Come
Free 5.10
Reduction in Chronic Waste Is Not
Capital-Intensive 5.10
The Return on Investment Is Among the
Highest 5.10
The Major Gains Come from the Vital
Few Projects 5.11
Disillusioned by the Failures 5.11
“Higher Quality Costs More” 5.11
The Illusion of Delegation 5.11
Employee Apprehensions 5.11
Awareness: Proof of the Need 5.12
The Size of the Chronic Waste 5.13
The Potential Return on Investment 5.13
Use of Bellwether Projects 5.14
Getting the Cost Figures 5.14
Languages in the Hierarchy 5.14
Presentations to Upper Managers 5.15
The Need for Formality 5.16
Membership 5.17
Responsibilities 5.17
Anticipating the Questions 5.18
Apprehensions about Elimination of Jobs
Assistance from the Quality Department
Deployment of Goals 5.20
The Project Concept 5.20
Use of the Pareto Principle 5.20
The Useful Many Projects 5.24
Sources of Nominations 5.24
Effect of the Big Q Concept 5.24
The Nomination Processes 5.25
Nominations from the Work Force 5.25
Joint Projects with Suppliers and
Customers 5.26
Criteria for the First Projects 5.27
Criteria for Projects Thereafter 5.27
Vital Few and Useful Many 5.27
Cost Figures for Projects 5.28
Costs versus Percent Deficiencies 5.28
Elephant-Sized and Bite-Sized Projects
Cloning 5.29
Purpose of Mission Statements 5.30
The Numerical Goal 5.30
Perfection as a Goal 5.30
Publication of Mission Statements 5.31
Revision of Mission Statements 5.31
Why a Team? 5.32
Appointment of Teams; Sponsors 5.32
Responsibilities and Rights 5.33
Membership 5.33
1In the fourth edition, the section on quality improvement was prepared by Frank M. Gryna.
Membership from the Work Force 5.34
Upper Managers on Teams 5.34
Model of the Infrastructure 5.35
The Team Leader 5.36
The Team Secretary 5.36
The Team Members 5.36
Finding the Time to Work on Projects
The Roles 5.37
The Qualifications 5.38
Sources and Tenure 5.38
The Two Journeys 5.39
Definition of Key Words 5.39
Diagnosis Should Precede Remedy 5.40
Understanding the Symptoms 5.40
Autopsies 5.41
Generating Theories 5.41
Arranging Theories 5.41
Choosing Theories to Be Tested 5.42
The Factual Approach 5.43
Flow Diagrams 5.44
Process Capability Analysis 5.44
Process Dissection 5.44
Simultaneous Dissection 5.45
Defect Concentration Analysis 5.45
Association Searches 5.47
Cutting New Windows 5.48
Design of Experiments 5.49
Measurement for Diagnosis 5.50
Responsibility for Diagnosis 5.50
The Santayana Review 5.51
The Influence of Cycle Time and
Frequency 5.51
Application to High-Frequency Cycles
Application to Intermediate-Frequency
Cycles 5.52
Application to Low-Frequency Cycles
Some Famous Case Examples 5.53
The Potential for Long-Cycle Events 5.54
Choice of Alternatives 5.55
Remedies: Removing the Causes 5.55
Remedies: Optimizing the Costs 5.56
Remedies: Acceptability 5.56
The Remedy for Rare but Critical Defects
Remedy through Replication 5.57
Test under Operating Conditions 5.57
Control at the New Level; Holding the
Gains 5.57
Extent of Work Force Errors 5.58
Species of Work Force Errors 5.58
Inadvertent Errors 5.59
Remedies for Inadvertent Errors 5.60
Technique Errors 5.60
Remedies for Technique Errors 5.62
Conscious Errors 5.62
Remedies for Conscious Errors 5.62
Communication Errors 5.63
Remedies for Communication Errors
Cultural Patterns 5.65
Rules of the Road 5.66
Resolving Differences 5.66
Review of Results 5.68
Inputs to Progress Review 5.69
Evaluation of Performance 5.69
Some Accounts of Mobilizing for Quality
Improvement 5.73
The purpose of this section is to explain the nature of quality improvement and its relation to
managing for quality, show how to establish quality improvement as a continuing process that
goes on year after year, and define the action plan and the roles to be played, including those of
upper management.
As used here, “improvement” means “the organized creation of beneficial change; the attainment of
unprecedented levels of performance.” A synonym is “breakthrough.”
Two Kinds of Beneficial Change. Better quality is a form of beneficial change. It is applicable
to both the kinds of quality that are summarized in Section 2, Figure 2.1:
Product features: These can increase customer satisfaction. To the producing company, they are
Freedom from deficiencies: These can create customer dissatisfaction and chronic waste. To the
producing company, they are cost-oriented.
Quality improvement to increase income may consist of such actions as
Product development to create new features that provide greater customer satisfaction and hence
may increase income
Business process improvement to reduce the cycle time for providing better service to customers
Creation of “one-stop shopping” to reduce customer frustration over having to deal with multiple
personnel to get service
Quality improvement to reduce deficiencies that create chronic waste may consist of such actions as
Increase of the yield of factory processes
Reduction of the error rates in offices
Reduction of field failures
The end results in both cases are called “quality improvement.” However, the processes used to
secure these results are fundamentally different, and for a subtle reason.
Quality improvement to increase income starts by setting new goals, such as new product features,
shorter cycle times, and one-stop shopping. Meeting such new goals requires several kinds
of planning, including quality planning. This quality planning is done through a universal series of
steps: identify the “customers” who will be affected if the goal is met, determine the needs of those
customers, develop the product features required to meet those needs, and so on. Collectively, this
series of steps is the “quality planning roadmap,” which is the subject matter of Section 3, The
Quality Planning Process.
In the case of chronic waste, the product goals are already in place; so are the processes for meeting
those goals. However, the resulting products (goods and services) do not all meet the goals. Some
do and some do not. As a consequence, the approach to reducing chronic waste is different from the
quality planning roadmap. Instead, the approach consists of (1) discovering the causes—why do
some products meet the goal and others do not—and (2) applying remedies to remove the causes. It
is this approach to quality improvement that is the subject of this section.
Continuing improvement is needed for both kinds of quality, since competitive pressures apply
to each. Customer needs are a moving target. Competitive costs are also a moving target. However,
improvement for these two kinds of quality has in the past progressed at very different rates. The
chief reason is that many upper managers, perhaps most, give higher priority to increasing sales than
to reducing costs. This difference in priority is usually reflected in the respective organization structures.
An example is seen in the approach to new product development.
Structured Product Development. Many companies maintain an organized approach for
evolving new models of products, year after year. Under this organized approach:
Product development projects are a part of the business plan.
A New Products Committee maintains a business surveillance over these projects.
Full-time product and process development departments are equipped with personnel, laboratories,
and other resources to carry out the technological work.
There is clear responsibility for carrying out the essential technological work.
A structured procedure is used to progress the new developments through the functional departments.
The continuing existence of this structure favors new product development on a year-to-year
This special organization structure, while necessary, is not sufficient to ensure good results. In
some companies, the cycle time for getting new products to market is lengthy, the new models compete
poorly in the market, or new chronic wastes are created. Such weaknesses usually are traceable
to weaknesses in the quality planning process, as discussed in Section 3, The Quality Planning
Unstructured Reduction of Chronic Waste. In most companies, the urge to reduce
chronic waste has been much lower than the urge to increase sales. As a result:
The business plan has not included goals for reduction of chronic waste.
Responsibility for such quality improvement has been vague. It has been left to volunteers to initiate
The needed resources have not been provided, since such improvement has not been a part of the
business plan.
The lack of priority by upper managers is traceable in large part to two factors that influence the
thinking processes of many upper managers:
1. Not only do many upper managers give top priority to increasing sales, but some of them even
regard cost reduction as a form of lower-caste work that is not worthy of the time of upper managers.
This is especially the case in high-tech industries.
2. Upper managers have not been aware of the size of the chronic waste, nor of the associated potential
for high return on investment. The “instrument panel” available to upper managers has
stressed performance measures such as sales, profit, cash flow, and so on but not the size of chronic
waste and the associated opportunities. The quality managers have contributed to this unawareness
by presenting their reports in the language of quality specialists rather than in the language
of management—the language of money.
The major focus of this section of the handbook is to show how companies can mobilize their
resources to deal with this neglected opportunity.
Chronic waste does not seem to have been a major problem during the early centuries of artisanship.
The artisan typically carried out many tasks to complete a unit of product. During each of these
tasks, he was his own customer. His client lived in the same village, so the feedback loops were short
and prompt.
The Industrial Revolution of the mid-eighteenth century greatly reduced the role of artisans while
creating large factories and complex organizational structures that became breeding grounds for
chronic waste. The Taylor system of the early twentieth century improved productivity but had a negative
effect on quality. To minimize the damage, the companies expanded product inspection. This
helped to shield customers from receiving defective products but encouraged the resulting chronic
waste, which became huge.
The widespread practice of relying on inspection was shattered by the Japanese quality revolution
that followed World War II. That revolution greatly reduced chronic waste, improved product
features, and contributed to making Japan an economic superpower. In addition, it greatly intensified
international competition in quality. This competition soon created a growing crisis in Western countries,
reaching alarming proportions by the 1980s.
In response to the crisis, many companies, especially in the United States, undertook initiatives to
improve their quality. For various reasons, most of these initiatives fell far short of their goals.
However, a relatively few companies made stunning improvements in quality and thereby became
the role models. The methods used by these role models have been analyzed and have become the
lessons learned—what actions are needed to attain quality leadership and what processes must be
devised to enable those actions to be taken.
Lessons Learned. Analysis of the actions taken by the successful companies shows that most
of them carried out many or all of the strategies set out below:
They enlarged the business plan at all levels to include goals for quality improvement.
They designed a process for making improvements and set up special organizational machinery
to carry out that process.
They adopted the big Q concept—they applied the improvement process to business processes as
well as to manufacturing processes.
They trained all levels of personnel, including upper management, in how to carry out their
respective missions of managing for quality.
They empowered the work force to participate in making improvements.
They established measures to evaluate progress against the improvement goals.
The managers, including the upper managers, reviewed progress against the improvement goals.
They expanded use of recognition for superior quality performance.
They revised the reward system to recognize the changes in job responsibilities.
The Rate of Improvement Is Decisive. The central lesson learned was that the annual rate
of quality improvement determines which companies emerge as quality leaders. Figure 5.1 shows the
effect of differing rates of quality improvement.
In this figure, the vertical scale represents product saleability, so what goes up is good. The
upper line shows the performance of company A, which at the outset was the industry quality
leader. Company A kept getting better, year after year. In addition, company A was profitable.
Seemingly, Company A faced a bright future.
The lower line shows that company B, a competitor, was at the outset not the quality leader.
However, company B has improved at a rate much faster than that of company A. Company A is now
threatened with loss of its quality leadership. The lesson is clear:
The most decisive factor in the competition for quality leadership is the rate of quality improvement.
The sloping lines of Figure 5.1 help to explain why Japanese goods attained quality leadership in
so many product lines. The major reason is that the Japanese rate of quality improvement was for
decades revolutionary when compared with the evolutionary rate of the West.
Figure 5.2 shows my estimate of the rates of quality improvement in the automobile industry. [For
elaboration, see Juran (1993).] There are also lessons to be learned from the numerous initiatives
during the 1980s that failed to produce useful results. These have not been well analyzed, but one
lesson does stand out. Collectively, those failed initiatives show that attaining a revolutionary rate of
quality improvement is not simple at all. There are numerous obstacles and much cultural resistance,
as will be discussed throughout this section.
By the late 1990s, the efforts to meet competition in quality were proceeding along two lines based
on two very different philosophies:
FIGURE 5.1 Two contrasting rates of improvement. (From Making Quality
Happen, 1988, Juran Institute, Wilton, CT, p. D4.)
1950 1960
1970 1980
Quality (salability of product)
1990 2000
FIGURE 5.2 Estimate of rates of quality improvement in the automobile
industry. (From Making Quality Happen, 1988, Juran Institute, Wilton, CT,
p. D5.)
1. Political leaders focused on traditional political solutions—import quotas, tariffs, legislation on
“fair trade,” and so on.
2. Industrial leaders increasingly became convinced that the necessary response to competition was
to become more competitive. This approach required applying the lessons learned from the role
models across the entire national economy. Such a massive scaling up likely would extend well
into the twenty-first century.
The Emerging Consensus. The experience of recent decades and the lessons learned have
led to an emerging consensus as to the status of managing for quality, the resulting threats and opportunities,
and the actions that need to be taken. As related to quality improvement, the high points of
this consensus include the following:
Competition in quality has intensified and has become a permanent fact of life. A major needed
response is a high rate of quality improvement, year after year.
Customers are increasingly demanding improved quality from their suppliers. These demands are
then transmitted through the entire supplier chain. The demands may go beyond product improvement
and extend to improving the system of managing for quality. (For an example, see Krantz
1989. In that case, a company used product inspection to shield its customers from receiving
defective products. Nevertheless, a large customer required the company to revise its system of
managing for quality as a condition for continuing to be a supplier.)
The chronic wastes are known to be huge. In the United States during the early 1980s, about a
third of what was done consisted of redoing what was done previously, due to quality deficiencies
(estimate by the author). The emerging consensus is that such wastes should not continue on
and on, since they reduce competitiveness in costs.
Quality improvement should be directed at all areas that influence company performance—
business processes as well as factory processes.
Quality improvement should not be left solely to voluntary initiatives; it should be built into the
Attainment of quality leadership requires that the upper managers personally take charge of managing
for quality. In companies that did attain quality leadership, the upper managers personally
guided the initiative. I am not aware of any exceptions.
The remainder of this section focuses on “how to do it.” The division of the subject includes
The basic concepts that underlie quality improvement
How to mobilize a company’s resources so as to make quality improvement an integral part of
managing the company
The improvement process itself—the universal sequence of steps for making any improvement
How to “institutionalize” improvement so that it goes on and on, year after year
The quality improvement process rests on a base of certain fundamental concepts. For most companies
and managers, annual quality improvement is not only a new responsibility, it is also a radical
change in the style of management—a change in company culture. Therefore, it is important to grasp
the basic concepts before getting into the improvement process itself.
Improvement Distinguished from Control. Improvement differs from control. The trilogy
diagram (Figure 5.3) shows this difference. (Note that Figure 5.3 is identical with Figure 2.4
in Section 2.) In this figure, the chronic waste level (the cost of poor quality) was originally about
23 percent of the amount produced. This chronic waste was built into the process—“It was planned
that way.” Later, a quality improvement project reduced this waste to about 5 percent. Under my
definition, this reduction in chronic waste is an improvement—it attained an unprecedented level
of performance.
Figure 5.3 also shows a “sporadic spike”—a sudden increase in waste to about 40 percent. Such
spikes are unplanned—they arise from various unexpected sources. The personnel promptly got rid
of that spike and restored the previous chronic level of about 23 percent. This action did not meet the
definition of an improvement—it did not attain an unprecedented level of performance. Usual names
for such actions are “troubleshooting”, “corrective action”, or “firefighting.”
All Improvement Takes Place Project by Project. There is no such thing as improvement
generally. All improvement takes place project by project and in no other way.
As used here, “improvement project” means “a chronic problem scheduled for solution.” Since
improvement project has multiple meanings, the company glossary and training manuals should
define it. The definition is helped by including some case examples that were carried out successfully
in that company.
Quality Improvement Is Applicable Universally. The huge numbers of projects carried
out during the 1980s and 1990s demonstrated that quality improvement is applicable to
FIGURE 5.3 The Juran trilogy diagram. (Adapted from Juran, J. M, 1989, The Quality Trilogy: A Universal Approach to
Managing for Quality, Juran Institute, Inc., Wilton, CT.)
Service industries as well as manufacturing industries
Business processes as well as manufacturing processes
Support activities as well as operations
Software as well as hardware
During the 1980s and 1990s, quality improvement was applied to virtually all industries, including
government, education, and health. [For a seminal book related to the health industry, see
Berwick et al. (1990).]
In addition, quality improvement has been applied successfully to the entire spectrum of company
functions: finance, product development, marketing, legal, and so on.
In one company, the legal vice-president doubted that quality improvement could be applied to
legal work. Yet within 2 years he reduced by more that 50 percent the cycle time of filing for a
patent. (Private communication to the author.)
(For elaboration and many case examples, see the Proceedings of the Juran Institute’s Annual
IMPRO Conferences on Quality Management.)
Quality Improvement Extends to All Parameters. Published reports of quality
improvements show that the effects have extended to all parameters:
Productivity: The output per person-hour
Cycle time: The time required to carry out processes, especially those which involve many steps
performed sequentially in various departments. Section 6, Process Management, elaborates on
improvement as applied to such processes.
Human safety: Many projects improve human safety through errorproofing, fail-safe designs,
and so on.
The environment: Similarly, many projects have been directed at protecting the environment by
reducing toxic emissions and so on.
Some projects provide benefits across multiple parameters. A classic example was the color television
set (Juran 1979). The Japanese Matsushita Company had purchased an American color television
factory (Quasar). Matsushita then made various improvements, including
Product redesign to reduce field failures
Process redesign to reduce internal defect rates
Joint action with suppliers to improve quality of purchased components
The results of these and other changes are set out in the before and after data:
1974 1977
Fall-off rate, i.e., defects (on assembled sets) requiring repair 150 per 100 sets 4 per 100 sets
Number of repair and inspection personnel 120 15
Failure rate during the warranty period 70% 10%
Cost of service calls $22 million $4 million
The manufacturer benefited in multiple ways: lower costs, higher productivity, more reliable
deliveries, and greater saleability. The ultimate users also benefited—the field failure rate was
reduced by over 80 percent.
The Backlog of Improvement Projects Is Huge. The existence of a huge backlog is evident
from the numbers of improvements actually made by companies that carried out successful initiatives
during the 1980s and 1990s. Some reported making improvements by the thousands, year
after year. In very large companies, the numbers are higher still, by orders of magnitude.
The backlog of improvement projects exists in part because the planning of new products and
processes has long been deficient. In effect, the planning process has been a dual hatchery. It hatched
out new plans. It also hatched out new chronic wastes, and these accumulated year after year. Each
such chronic waste then became a potential improvement project.
A further reason for a huge backlog is the nature of human ingenuity—it seems to have no limit.
Toyota Motor Company has reported that its 80,000 employees offered 4 million suggestions for
improvement during a single year—an average of 50 suggestions per person per year (Sakai 1994).
Quality Improvement Does Not Come Free. Reduction of chronic waste does not
come free—it requires expenditure of effort in several forms. It is necessary to create an infrastructure
to mobilize the company’s resources toward the end of annual quality improvement. This
involves setting specific goals to be reached, choosing projects to be tackled, assigning responsibilities,
following progress, and so on.
There is also a need to conduct extensive training in the nature of the improvement process, how
to serve on improvement teams, how to use the tools, and so on.
In addition to all this preparatory effort, each improvement project requires added effort to conduct
diagnoses to discover the causes of the chronic waste and provide remedies to eliminate the
The preceding adds up to a significant front-end outlay, but the results can be stunning. They
have been stunning in the successful companies—the role models. Detailed accounts of such
results have been widely published, notably in the proceedings of the annual conferences held by
the U.S. National Institute for Standards and Technology (NIST), which administers the Malcolm
Baldrige National Quality Award.
Reduction in Chronic Waste Is Not Capital-Intensive. Reduction in chronic waste seldom
requires capital expenditures. Diagnosis to discover the causes usually consists of the time of the
quality improvement project teams. Remedies to remove the causes usually involve fine-tuning
the process. In most cases, a process that is already producing over 80 percent good work can be
raised to the high 90s without capital investment. Such avoidance of capital investment is a major
reason why reduction of chronic waste has a high return on investment (ROI).
In contrast, projects for product development to increase sales involve outlays to discover customer
needs, design products and processes, build facilities, and so on. Such outlays are largely
classified as capital expenditures and thereby lower the ROI estimates.
The Return on Investment Is Among the Highest. This is evident from results publicly
reported by national award winners in Japan (Deming Prize), the United States (Baldrige
Award), and elsewhere. More and more companies have been publishing reports describing their
quality improvements, including the gains made. [For examples, see the Proceedings of the Juran
Institute’s Annual IMPRO Conferences on Quality Management for 1983 and subsequent years. See
especially, Kearns and Nadler (1995).]
While these and other published case examples abound, the actual return on investment from
quality improvement projects has not been well researched. I once examined 18 papers published by
companies and found that the average quality improvement project had yielded about $100,000 of
cost reduction (Juran 1985). The companies were large—sales in the range of over $1 billion (milliard)
per year.
I have also estimated that for projects at the $100,000 level, the investment in diagnosis and remedy
combined runs to about $15,000. The resulting ROI is among the highest available to managers.
It has caused some managers to quip: “The best business to be in is quality improvement.”
The Major Gains Come from the Vital Few Projects. The bulk of the measurable
gains comes from a minority of the quality improvement projects—the “vital few.” These are
multifunctional in nature, so they need multifunctional teams to carry them out. In contrast, the
majority of the projects are in the “useful many” category and are carried out by local departmental
teams. Such projects typically produce results that are orders of magnitude smaller than
those of the vital few.
While the useful many projects contribute only a minor part of the measurable gains, they provide
an opportunity for the lower levels of the hierarchy, including the work force, to participate in
quality improvement. In the minds of many managers, the resulting gain in quality of work life is
quite as important as the tangible gains in operating performance.
While the role-model companies achieved stunning results through quality improvement, most companies
did not. Some of these failures were due to honest ignorance of how to mobilize for improvement,
but there are also some inherent inhibitors to establishing improvement on a year-to-year basis.
It is useful to understand the nature of some of the principal inhibitors before setting out.
Disillusioned by the Failures. The lack of results mentioned earlier has led some influential
journals to conclude that improvement initiatives are inherently doomed to failure. Such
conclusions ignore the stunning results achieved by the role-model companies. (Their results
prove that such results are achievable.) In addition, the role models have explained how they got
those results, thereby providing lessons learned for other companies to follow. Nevertheless,
the conclusions of the media have made some upper managers wary about going into quality
“Higher Quality Costs More.” Some managers hold to a mindset that “higher quality costs
more.” This mindset may be based on the outmoded belief that the way to improve quality is to
increase inspection so that fewer defects escape to the customer. It also may be based on the confusion
caused by the two meanings of the word “quality.”
Higher quality in the sense of improved product features (through product development) usually
requires capital investment. In this sense, it does cost more. However, higher quality in the sense of
lower chronic waste usually costs less—a lot less. Those who are responsible for preparing proposals
for management approval should be careful to define the key words—Which kind of quality are
they talking about?
The Illusion of Delegation. Managers are busy people, yet they are constantly bombarded
with new demands on their time. They try to keep their workload in balance through delegation. The
principle that “a good manager is a good delegator” has wide application, but it has been overdone
as applied to quality improvement. The lessons learned from the role-model companies show that
going into annual quality improvement adds minimally about 10 percent to the workload of the entire
management team, including the upper managers.
Most upper managers have tried to avoid this added workload through sweeping delegation.
Some established vague goals and then exhorted everyone to do better—“Do it right the first time.”
In the role-model companies, it was different. In every such company, the upper managers took
charge of the initiative and personally carried out certain nondelegable roles. (See below, under The
Nondelegable Roles of Upper Managers.)
Employee Apprehensions. Going into quality improvement involves profound changes in a
company’s way of life—far more than is evident on the surface. It adds new roles to the job descriptions
and more work to the job holders. It requires accepting the concept of teams for tackling projects—a
concept that is alien to many companies and which invades the jurisdictions of the functional departments.
It raises the priority of quality, with damaging effects on other priorities. It requires training on
how to do all this. Collectively, it is a megachange that disturbs the peace and breeds many unwanted
side effects.
To the employees, the most frightening effect of this profound set of changes is the threat to jobs
and/or status. Reduction of chronic waste reduces the need for redoing prior work and hence the
jobs of people engaged in such redoing. Elimination of such jobs then becomes a threat to the status
and/or jobs of the associated supervision. It should come as no surprise if the efforts to reduce
waste are resisted by the work force, the union, the supervision, and others.
Nevertheless, quality improvement is essential to remaining competitive. Failure to go forward
puts all jobs at risk. Therefore, the company should go into improvement while realizing that
employee apprehension is a very logical reaction of worried people to worrisome proposals. The
need is to open a communication link to explain the why, understand the worries, and search for optimal
solutions. In the absence of forthright communication, the informal channels take over, breeding
suspicions and rumors. For added discussion, see below, under the Quality Council: Anticipating
the Questions.
Additional apprehension has its origin in cultural patterns. (See below, under Resistance to
Change, Cultural Patterns.) (The preceding apprehensions do not apply to improvement of product
features to increase sales. These are welcomed as having the potential to provide new opportunities
and greater job security.)
The lessons learned during the 1980s and 1990s include a major finding: Personal participation by
upper managers is indispensable to getting a high rate of annual quality improvement. This finding
suggests that advocates for quality initiatives should take positive steps to convince the upper
managers of
The merits of annual quality improvement
The need for active upper management participation
The precise nature of the needed upper management participation
Awareness: Proof of the Need. Upper managers respond best when they are shown a major
threat or opportunity. An example of a major threat is seen in the case of company G, a maker of
household appliances. Company G and its competitors R and T were all suppliers to a major customer
involving four models of appliances. (See Table 5.1.) This table shows that in 1980, company
G was a supplier for two of the four models. Company G was competitive in price, on-time delivery,
and product features, but it was definitely inferior in the customer’s perception of quality, the chief
problem being field failures. By 1982, lack of response had cost company G the business on model
number 1. By 1983, company G also had lost the business on model number 3.
TABLE 5.1 Suppliers to a Major Customer
Model number 1980 1981 1982 1983
1 G G R R
2 R R R R
3 G G G R
4 T R R R
Awareness also can be created by showing upper managers other opportunities, such as cost
reduction through cutting chronic waste.
The Size of the Chronic Waste. A widespread major opportunity for upper managers is
to reduce the cost of poor quality. In most cases, this cost is greater than the company’s annual profit,
often much greater. Quantifying this cost can go far toward proving the need for a radical change
in the approach to quality improvement. An example is shown in Table 5.2. This table shows the
estimated cost of poor quality for a company in a process industry using the traditional accounting
classifications. The table brings out several matters of importance to upper managers:
The order of magnitude: The total of the costs is estimated at $9.2 million per year. For this company,
this sum represented a major opportunity. (When such costs have never before been brought
together, the total is usually much larger than anyone would have expected.)
The areas of concentration: The table is dominated by the costs of internal failures—they are
79.4 percent of the total. Clearly, any major cost reduction must come from the internal failures.
The limited efforts for prevention: The figure of 1.9 percent for prevention suggests that greater
investment in prevention would be cost-effective.
(For elaboration, see Section 8, Quality and Costs.)
The Potential Return on Investment. A major responsibility of upper managers is to
make the best use of the company’s assets. A key measure of judging what is best is return on investment
(ROI). In general terms, ROI is the ratio of (1) the estimated gain to (2) the estimated resources
needed. Computing ROI for projects to reduce chronic waste requires assembling estimates such as
The costs of chronic waste associated with the projects
The potential cost reductions if the projects are successful
The costs of the needed diagnosis and remedy
Many proposals to go into quality improvement have failed to gain management support because
no one has quantified the ROI. Such an omission is a handicap to the upper managers—they are
unable to compare (1) the potential ROI from quality improvement with (2) the potential ROI from
other opportunities for investment.
Quality managers and others who prepare such proposals are well advised to prepare the information
on ROI in collaboration with those who have expertise in the intricacies of ROI. Computation
of ROI gets complicated because two kinds of money are involved—capital and expenses. Each is
money, but in some countries (including the United States) they are taxed differently. Capital expenditures
are made from after-tax money, whereas expenses are paid out of pretax money.
This difference in taxation is reflected in the rules of accounting. Expenses are written off
promptly, thereby reducing the stated earnings and hence the income taxes on earnings. Capital
expenditures are written off gradually—usually over a period of years. This increases the stated
TABLE 5.2 Analysis of Cost of Poor Quality
Category Amount, $ Percent of total
Internal failures 7,279,000 79.4
External failures 283,000 3.1
Appraisal 1,430,000 15.6
Prevention 170,000 1.9
9,162,000 100.0
earnings and hence the income taxes on those earnings. All this is advantageous to proposals to go
into quality improvement because quality improvement is seldom capital intensive. (Some upper
managers tend to use the word investment as applying only to capital investment.)
Use of Bellwether Projects. Presentation of the cost figures becomes even more effective if
it is accompanied by a “bellwether project”—a case example of a successful quality improvement
actually carried out within the company. Such was the approach used in the ABC company, a large
maker of electronic instruments.
Historically, ABC’s cost of poor quality ran to about $200 million annually. A notorious part of
this was the $9 million of annual cost of scrap for a certain electronic component. The principal
defect type was defect X. It had been costing about $3 million per year.
The company had launched a project to reduce the frequency of defect X. The project was a stunning
success—it had cut the cost of defect X from $3 million to $1 million—an annual improvement
of $2 million. The investment needed to make this improvement was modest—about one-fourth of a
million—to fine-tune the process and its controls. The gain during the first year of application had
been eight times the investment.
This bellwether project was then used to convince the upper managers that expansion of quality
improvement could greatly reduce the company’s cost of poor quality and do so at a high return on
the investment.
In most companies, the previously successful quality improvements can serve collectively as a
bellwether project. The methodology is as follows:
Identify the quality improvement projects completed within the last year or two.
For each such project, estimate (1) what was gained and (2) what was the associated expenditure.
Summarize, and determine the composite ROI.
Compare this composite with the returns being earned from other company activities. (Such comparisons
usually show that quality improvement provides the highest rate of return.)
Getting the Cost Figures. Company accounting systems typically quantify only a minority
of the costs of poor quality. The majority are scattered throughout the various overheads. As a result,
quality specialists have looked for ways to supply what is missing. Their main efforts toward solution
have been as follows:
1. Make estimates: This is the “quick and dirty” approach. It is usually done by sampling and
involves only a modest amount of effort. It can, in a few days or weeks, provide (a) an evaluation
of the approximate cost of chronic waste and (b) indicate where this is concentrated.
2. Expand the accounting system: This is much more elaborate. It requires a lot of work from various
departments, especially Accounting and Quality. It runs into a lot of calendar time, often two
or three years. (For elaboration, see Section 8, Quality and Costs.)
In my experience, estimates involve much less work, can be prepared in far less time, and yet are
adequate for managerial decision making.
Note that the demand for “accuracy” of the cost figures depends on the use to which the figures
will be put. Balancing the books demands a high degree of accuracy. Making managerial decisions
sometimes can tolerate a margin of error. For example, a potential improvement project has been
estimated to incur about $300,000 in annual cost of poor quality. This figure is challenged. The contesting
estimates range from $240,000 to $360,000—quite a wide range. Then someone makes an
incisive observation: “It doesn’t matter which estimate is correct. Even at the lowest figure, this is a
good opportunity for improvement, so let’s tackle it.” In other words, the managerial decision to
tackle the project is identical despite a wide range of estimate.
Languages in the Hierarchy. A subtle aspect of securing upper management approval is
choice of language. Industrial companies make use of two standard languages—the language of
money and the language of things. (There are also local dialects, each peculiar to a specific function.)
However, as seen in Figure 5.4, use of the standard languages is not uniform.
Figure 5.4 shows the use of standard languages
in different levels of a typical hierarchy.
At the apex, the principal language of the top
management team is the language of money. At
the base, the principal language of the first-line
supervisors and the work force is the language
of things. In between, the middle managers and
the specialists need to understand both the principal
languages—the middle managers should
be bilingual.
It is quite common for chronic waste to be
measured in the language of things: percent
errors, process yields, hours of rework, and so
on. Converting these measures into the language
of money enables upper managers to relate them
to the financial measures that have long dominated
the management “instrument panel.”
Years ago, I was invited to visit a major
British manufacturer to study its approach to
managing for quality, and to provide a critique.
I found that the company’s cost of poor
quality was huge, that it was feasible to cut
this in two in 5 years, and that the resulting
return on investment would be much greater than that of making and selling the company’s products.
When I explained this to the managing director, he was most impressed—it was the first
time that the problem of chronic waste had been explained to him in the language of return on
investment. He promptly convened his directors (vice presidents) to discuss what to do about this
Presentations to Upper Managers. Presentations to upper managers should focus on the
goals of the upper managers, not on the goals of the advocates. Upper managers are faced with meeting
the needs of various stakeholders: customers, owners, employees, suppliers, the public (e.g.,
safety, health, the environment), and so on. It helps if the proposals identify specific problems of
stakeholders and estimate the benefits to be gained.
Upper managers receive numerous proposals for allocating the company’s resources: invade foreign
markets, develop new products, buy new equipment to increase productivity, make acquisitions,
enter joint ventures, and so on. These proposals compete with each other for priority, and a major
test is return on investment (ROI). It helps if the proposal to go into quality improvement includes
estimates of ROI.
Explanation of proposals is sometimes helped by converting the supporting data into units of
measure that are already familiar to upper managers. For example:
Last year’s cost of poor quality was five times last year’s profit of $1.5 million.
Cutting the cost of poor quality in half would increase earnings by 13 cents per share of stock.
Thirteen percent of last year’s sales orders were canceled due to poor quality.
Thirty-two percent of engineering time was spent in finding and correcting design weaknesses.
Twenty-five percent of manufacturing capacity is devoted to correcting quality problems.
Seventy percent of the inventory carried is traceable to poor quality.
Twenty-five percent of all manufacturing hours were spent in finding and correcting defects.
FIGURE 5.4 Common languages in the hierarchy.
Last year’s cost of poor quality was the equivalent of the X factory making 100 percent defective
work during the entire year.
Experience in making presentations to upper management has evolved some useful do’s and don’ts.
Do summarize the total of the estimated costs of poor quality. The total will be big enough to
command upper management attention.
Do show where these costs are concentrated. A common grouping is in the form of Table 5.2.
Typically (as in that case), most of the costs are associated with failures, internal and external.
Table 5.2 also shows the fallacy of trying to start by reducing inspection and test. The failure costs
should be reduced first. After the defect levels come down, inspection costs can be reduced as well.
Do describe the principal projects that are at the heart of the proposal.
Do estimate the potential gains, as well as the return on investment. If the company has never
before undertaken an organized approach to reducing quality-related costs, then a reasonable
goal is to cut these costs in two within a space of 5 years.
Do have the figures reviewed in advance by those people in finance (and elsewhere) to whom
upper management looks for checking the validity of financial figures.
Don’t inflate the present costs by including debatable or borderline items. The risk is that the
decisive review meetings will get bogged down in debating the validity of the figures without ever
getting to discuss the merits of the proposals.
Don’t imply that the total costs will be reduced to zero. Any such implication will likewise divert
attention from the merits of the proposals.
Don’t force the first few projects on managers who are not really sold on them or on unions who
are strongly opposed. Instead, start in areas that show a climate of receptivity. The results
obtained in these areas will determine whether the overall initiative will expand or die out.
The needs for quality improvement go beyond satisfying customers or making cost reductions.
New forces keep coming over the horizon. Recent examples have included growth in product liability,
the consumerism movement, foreign competition, and legislation of all sorts. Quality improvement
has provided much of the response to such forces.
Similarly, the means of convincing upper managers of the need for quality improvement go
beyond reports from advocates. Conviction also may be supplied by visits to successful companies,
hearing papers presented at conferences, reading reports published by successful companies, and listening
to the experts, both internal and external. However, none of these is as persuasive as results
achieved within one’s own company.
A final element of presentation to upper managers is to explain their personal responsibilities in
launching and perpetuating quality improvement. (See below, under The Nondelegable Roles of
Upper Managers.)
Until the 1980s, quality improvement in the West was not mandated—it was not a part of the business
plan or a part of the job descriptions. Some quality improvement did take place, but on a voluntary
basis. Here and there a manager or a nonmanager, for whatever reason, elected to tackle some
improvement project. He or she might persuade others to join an informal team. The result might be
favorable, or it might not. This voluntary, informal approach yielded few improvements. The emphasis
remained on inspection, control, and firefighting.
The Need for Formality. The quality crisis that followed the Japanese quality revolution
called for new strategies, one of which was a much higher rate of quality improvement. It then
became evident that an informal approach would not produce thousands (or more) improvements
year after year. This led to experiments with structured approaches that in due course helped some
companies to become the role models.
Some upper managers protested the need for formality. “Why don’t we just do it?” The answer
depends on how many improvements are needed. For just a few projects each year, informality is
adequate; there is no need to mobilize. However, making improvements by the hundreds or the thousands
does require a formal structure. (For some published accounts of company experiences in
mobilizing for quality improvement, see under References, Some Accounts of Mobilizing for
Quality Improvement.)
As it has turned out, mobilizing for improvement requires two levels of activity, as shown in
Figure 5.5. The figure shows the two levels of activity. One of these mobilizes the company’s
resources to deal with the improvement projects collectively. This becomes the responsibility of
management. The other activity is needed to carry out the projects individually. This becomes the
responsibility of the quality improvement teams.
The first step in mobilizing for quality improvement is to establish the company’s quality council (or
similar name). The basic responsibility of this council is to launch, coordinate, and “institutionalize”
annual quality improvement. Such councils have been established in many companies. Their experiences
provide useful guide lines.
Membership. Council membership is typically drawn from the ranks of senior managers. Often
the senior management committee is also the quality council. Experience has shown that quality
councils are most effective when upper managers are personally the leaders and members of the
senior quality councils.
In large companies, it is common to establish councils at the divisional level as well as at the corporate
level. In addition, some individual facilities may be so large as to warrant establishing a local
quality council. When multiple councils are established, they are usually linked together—members
of high-level councils serve as chairpersons of lower-level councils. Figure 5.6 is an example of such
Experience has shown that organizing quality councils solely in the lower levels of management
is ineffective. Such organization limits quality improvement projects to the “useful many” while
neglecting the “vital few” projects—those which can produce the greatest results. In addition, quality
councils solely at lower levels send a message to all: “Quality improvement is not high on upper
management’s agenda.”
Responsibilities. It is important for each council to define and publish its responsibilities so
that (1) the members agree on what is their mission, and (2) the rest of the organization can become
informed relative to upcoming events.
Activities by management Activities by teams
Establish quality councils Analyze symptoms
Select projects; write mission statements Theorize as to causes
Assign teams Test theories
Review progress Establish causes
Provide recognition and rewards Stimulate remedies and controls
FIGURE 5.5 Mobilizing for quality improvement.
Many quality councils have published their statements of responsibility. Major common elements
have included the following:
Formulate the quality policies, such as focus on the customer, quality has top priority, quality
improvement must go on year after year, participation should be universal, or the reward system
should reflect performance on improvement.
Estimate the major dimensions, such as status of quality compared with competitors, extent of
chronic waste, adequacy of major business processes, or results achieved by prior improvements.
Establish processes for selecting projects, such as soliciting and screening nominations, choosing
projects to be tackled, preparing mission statements, or creating a favorable climate for quality
Establish processes for carrying out the projects, such as selecting team leaders and members or
defining the role of project teams.
Provide support for the project teams, such as training (see Section 16, Training for Quality), time
for working on projects, diagnostic support, facilitator support, or access to facilities for tests and
Establish measures of progress, such as effect on customer satisfaction, effect on financial performance,
or extent of participation by teams.
Review progress, assist teams in the event of obstacles, and ensure that remedies are implemented.
Provide for public recognition of teams.
Revise the reward system to reflect the changes demanded by introducing annual quality
Anticipating the Questions. Announcement of a company’s intention to go into annual
quality improvement always stimulates questions from subordinate levels, questions such as
What is the purpose of this new activity?
How does it relate to other ongoing efforts to make improvements?
FIGURE 5.6 How quality councils are linked together. (From
Making Quality Happen, 1988, Juran Institute, Wilton, CT, p. D17.)
How will it affect other quality-oriented activities?
What jobs will be affected, and how?
What actions will be taken, and in what sequence?
In view of this new activity, what should we do that is different from what we have been doing?
Quality councils should anticipate the troublesome questions and, to the extent feasible, provide
answers at the time of announcing the intention to go into annual quality improvement. Some senior
managers have gone to the extent of creating a videotape to enable a wide audience to hear the identical
message from a source of undoubted authority.
Apprehensions about Elimination of Jobs. Employees not only want answers to such
questions, they also want assurance relative to their apprehensions, notably the risk of job loss due
to quality improvement. Most upper managers have been reluctant to face up to these apprehensions.
Such reluctance is understandable. It is risky to provide assurances when the future is uncertain.
Nevertheless, some managers have estimated in some depth the two pertinent rates of change:
1. The rate of creation of job openings due to attrition: retirements, offers of early retirement, resignation,
and so on. This rate can be estimated with a fair degree of accuracy.
2. The rate of elimination of jobs due to reduction of chronic waste. This estimate is more speculative—
it is difficult to predict how soon the improvement rate will get up to speed. In practice, companies
have been overly optimistic in their estimates.
Analysis of these estimates can help managers to judge what assurances they can provide, if any.
It also can shed light on choice of alternatives for action: retrain for jobs that have opened up, reassign
to areas that do have job openings, offer early retirement, assist in finding jobs in other companies,
and/or provide assistance in the event of termination.
Assistance from the Quality Department. Many quality councils secure the assistance
of the Quality Department to
Provide inputs needed by the council for planning to introduce quality improvement
Draft proposals and procedures
Carry out essential details such as screening nominations for projects
Develop training materials
Develop new measures for quality
Prepare reports on progress
It is also usual, but not invariable, for the quality manager to serve as secretary of the quality council.
Companies that have become the quality leaders—the role models—all adopted the practice of
enlarging their business plan to include quality-oriented goals. In effect, they translated the threats
and opportunities faced by their companies into quality goals such as
Increase on-time deliveries from 83 to 98 percent over the next 2 years.
Reduce the cost of poor quality by 50 percent over the next 5 years.
Such goals are clear—each is quantified, and each has a timetable. Convincing upper managers to
establish such goals is a big step, but it is only the first step.
Deployment of Goals. Goals are merely a wish list until they are deployed—until they are
broken down into specific projects to be carried out and assigned to specific individuals or teams
who are then provided with the resources needed to take action. Figure 5.7 shows the anatomy of the
deployment process. In the figure, the broad (strategic) quality goals are established by the quality
council and become a part of the company business plan. These goals are then divided and allocated
to lower levels to be translated into action. In large organizations there may be further subdivision
before the action levels are reached. The final action level may consist of individuals or teams.
In response, the action levels select improvement projects that collectively will meet the goals.
These projects are then proposed to the upper levels along with estimates of the resources needed.
The proposals and estimates are discussed and revised until final decisions are reached. The end
result is an agreement on which projects to tackle, what resources to provide, and who will be
responsible for carrying out the projects.
This approach of starting at the top with strategic quality goals may seem like purely a top-down
activity. However, the deployment process aims to provide open discussion in both directions before
final decisions are made, and such is the way it usually works out.
The concept of strategic quality goals involves the vital few matters, but it is not limited to the
corporate level. Quality goals also may be included in the business plans of divisions, profit centers,
field offices, and still other facilities. The deployment process is applicable to all of these. (For added
discussion of the deployment process, see Section 13, Strategic Planning.)
The Project Concept. As used here, a project is a chronic problem scheduled for solution. The
project is the focus of actions for quality improvement. All improvement takes place project by project
and in no other way.
Some projects are derived from the quality goals that are in the company business plan. These are
relatively few in number, but each is quite important. Collectively, these are among the vital few projects
(see below, under Use of the Pareto principle). However, most projects are derived not from the
company business plan but from the nomination-selection process, as discussed below.
Use of the Pareto Principle. A valuable aid to selection of projects during the deployment
process is the Pareto Principle. This principle states that in any population that contributes to a com-
FIGURE 5.7 Anatomy of the deployment process. (From Visual OPQ9-2, Juran Institute, Inc.,
Wilton, CT.)
mon effect, a relative few of the contributors—the vital few—account for the bulk of the effect. The
principle applies widely in human affairs. Relatively small percentages of the individuals write most
of the books, commit most of the crimes, own most of the wealth, and so on.
An example of using the Pareto principle to select projects is seen in a paper mill’s goal of reducing
its cost of poor quality. The estimated total was $9,070,000 per year, divided among seven
accounting categories. (See Table 5.3.) One of these seven categories is called “broke.” It amounts
to $5,560,000, or 61 percent of total. Clearly, there will be no major improvement in the total unless
there is a successful attack on broke—this is where the money is concentrated. (Broke is paper mill
dialect for paper so defective that it must be returned to the beaters for reprocessing.)
This paper mill makes 53 types of paper. When the broke is analyzed by type of paper, the Pareto
principle is again in evidence. (See Table 5.4.) Six of the 53 product types account for $4,480,000,
which is 80 percent of the $5,560,000. There will be no major improvement in broke unless there is
a successful attack on these six types of paper. Note that studying 12 percent of the product types
results in attacking 80 percent of the cost of broke.
Finally, the analysis is extended to the defect types that result in the major cost of broke. There
are numerous defect types, but five of them dominate. (See Table 5.5.) The largest number is
$612,000 for “tear” on paper type B. Next comes $430,000 for “porosity” on paper type A, and so
on. Each such large number has a high likelihood of being nominated for an improvement project.
Identification of the vital few (in this case, accounting categories, product types, and defect types) is
made easier when the tabular data are presented in graphic form. Figures 5.8, 5.9, and 5.10 present the
paper mill data graphically. Like their tabular counterparts, each of these graphs contains three elements:
1. The contributors to the total effect, ranked by the magnitude of their contribution
2. The magnitude of the contribution of each expressed numerically and as a percentage of total
3. The cumulative percentage of total contribution of the ranked contributors
TABLE 5.3 Pareto Analysis by Accounts
Percent of total quality loss
Accounting category Annual quality loss,* This category Cumulative
Broke 5560 61 61
Customer claim 1220 14 75
Odd lot 780 9 84
High material cost 670 7 91
Downtime 370 4 95
Excess inspection 280 3 98
High testing cost 1190 2 100
TOTAL 9070
*Adjusted for estimated inflation since time of original study.
TABLE 5.4 Pareto Analysis by Products
Product type Annual broke loss,* Percent of Cumulative percent
$thousands broke loss broke loss
A 1320 24 24
B 960 17 41
C 720 13 54
D 680 12 66
E 470 8 74
F 330 (4480) 6 80
47 other types 1080 220 100
TOTAL 53 types 5560 100
*Adjusted for estimated inflation since time of original study.
TABLE 5.5 Matrix of Quality Costs*
Type Trim, Visual defects,† Caliper, Tear, Porosity, All other causes, Total,
$thousands $thousands $thousands $thousands $thousands $thousands $thousands
A 270 94 None‡ 162 430 364 1320
B 120 33 None‡ 612 58 137 960
C 95 78 380 31 74 62 720
D 82 103 None‡ 90 297 108 680
E 54 108 None‡ 246 None‡ 62 470
F 51 49 39 16 33 142 330
TOTAL 672 465 419 1157 892 875 4480
*Adjusted for estimated inflation since time of original study.
†Slime spots, holes, wrinkles, etc.
‡Not a specified requirement for this type.
FIGURE 5.8 Pareto analysis: annual loss by category.
Annual Loss, $ Millions
Percent of Total
Paper Type
E F Other
Cum Loss
Key: Loss Type Cum. Loss % Loss Cum. %
Annual Loss, $ Millions
Percent of Total
Paper Type – Defect Type
Cum Loss
FIGURE 5.9 Pareto analysis: annual loss by paper type.
FIGURE 5.10 Pareto analysis: annual loss by defect type.
In addition to facilitating analysis, presentation of the data in the form of a Pareto diagram greatly
enhances communication of the information, most notably in convincing upper management of the
source of a problem and gaining support for a proposed course of action to remedy the problem. [For
an account of how I came to misname the Pareto principle, see Juran (1975).]
The Useful Many Projects. Under the Pareto principle, the vital few projects provide the
bulk of the improvement, so they receive top priority. Beyond the vital few are the useful many projects.
Collectively they contribute only a minority of the improvement, but they provide most of the
opportunity for employee participation. Choice of these projects is made through the nominationselection
Most projects are chosen through the nomination and selection process, involving several steps:
Project nomination
Project screening and selection
Preparation and publication of project mission statements
Sources of Nominations. Nominations for projects can come from all levels of the organization.
At the higher levels, the nominations tend to be extensive in size (the vital few) and multifunctional
in their scope. At lower levels, the nominations are smaller in size (the useful many) and
tend to be limited in scope to the boundaries of a single department.
Nominations come from many sources. These include
Formal data systems such as field reports on product performance, customer complaints, claims,
returns, and so on; accounting reports on warranty charges and on internal costs of poor quality;
and service call reports. (Some of these data systems provide for analyzing the data to identify
problem areas.) [For an example of project nomination based on customer complaints, see Rassi
Special studies such as customer surveys, employee surveys, audits, assessments, benchmarking
against competitive quality, and so on.
Reactions from customers who have run into product dissatisfactions are often vocal and insistent.
In contrast, customers who judge product features to be not competitive may simply (and
quietly) become ex-customers.
Field intelligence derived from visits to customers, suppliers, and others; actions taken by competitors;
and stories published in the media (as reported by sales, customer service, technical service,
and others).
The impact of quality on society, such as new legislation, extension of government regulation, and
growth of product liability lawsuits.
The managerial hierarchy, such as the quality council, managers, supervisors, professional specialists,
and project teams.
The work force through informal ideas presented to supervisors, formal suggestions, ideas from
quality circles, and so on.
Proposals relating to business processes.
Effect of the Big Q Concept. Beginning in the 1980s, the scope of nominations for projects
broadened considerably under the big Q concept. (For details relative to the big Q concept, see
Section 2, Figure 2.1.)
The breadth of the big Q concept is evident from the wide variety of projects that have already
been tackled:
Improve the precision of the sales forecast.
Reduce the cycle time for developing new products.
Increase the success rate in bidding for business.
Reduce the time required to fill customers’ orders.
Reduce the number of sales cancellations.
Reduce the errors in invoices.
Reduce the number of delinquent accounts.
Reduce the time required to recruit new employees.
Improve the on-time arrival rate (for transportation services).
Reduce the time required to file for patents.
(For examples from many industries, see proceedings of IMPRO conferences. See also The Juran
Report.) (For elaboration on projects in business processes, see Section 6, Process Management.)
The Nomination Processes. Nominations must come from human beings. Data systems are
impersonal—they make no nominations. Various means are used to stimulate nominations for quality
improvement projects:
Call for nominations: Letters or bulletin boards are used to invite all personnel to submit nominations,
either through the chain of command or to a designated recipient such as the secretary
of the quality council.
Make the rounds: In this approach, specialists (such as quality engineers) are assigned to
visit the various departments, talk with the key people, and secure their views and nominations.
The council members themselves: They become a focal point for extensive data analyses and
Brainstorming meetings: These are organized for the specific purpose of making nominations.
Whatever the method used, it will produce the most nominations if it urges use of the big Q concept—
the entire spectrum of activities, products, and processes.
Nominations from the Work Force. The work force is potentially a source of numerous
nominations. Workers have extensive residence in the workplace. They are exposed to many local
cycles of activity. Through this exposure, they are well poised to identify the existence of quality
problems and to theorize about their causes. As to the details of goings on in the workplace, no
one is better informed than the work force. “That machine hasn’t seen a maintenance man for the
last 6 months.” In addition, many workers are well poised to identify opportunities and to propose
new ways.
Work force nominations consist mainly of local useful many projects along with proposals of a
human relations nature. For such nominations, workers can supply useful theories of causes as well
as practical proposals for remedies. For projects of a multifunctional nature, most workers are handicapped
by their limited knowledge of the overall process and of the interactions among the steps that
collectively make up the overall.
In some companies, the solicitation of nominations from the work force has implied that such
nominations would receive top priority. The effect was that the work force was deciding which projects
the managers should tackle first. It should have been made clear that workers’ nominations must
compete for priority with nominations from other sources.
Joint Projects with Suppliers and Customers. All companies buy goods and services
from suppliers; over half the content of the finished product may come from suppliers. In
earlier decades, it was common for customers to contend that “the supplier should solve his quality
problems.” Now there is growing awareness that these problems require a partnership approach
based on
Establishing mutual trust
Defining quality in terms of customer needs as well as specifications
Exchanging essential data
Direct communication at the technical level as well as the commercial level
This approach gains momentum from joint projects between suppliers and customers. Published
examples include
Alcoa and Kodak, involving photographic plates (Kegarise and Miller 1985).
Alcoa and Nalco, involving lubricants for rolling mills (Boley and Petska 1990).
Alcoa and Phifer, involving aluminum wire (Kelly et al. 1990).
NCR and its customers, establishing a universal code for tracking product failures as they
progress through the customer chain (Daughton 1987).
Efforts to serve customers are sometimes delayed by actions of the customers themselves.
A maker of technological instruments encountered delays when installing the instruments in customers’
premises, due to lack of site preparation. When the installers arrived at the site, the foundation
was not yet in place, supply lines such as compressed air were not yet in place, and so on.
The company analyzed a number of these delays and then created a videotape on site preparation.
The company sent this videotape to customers at the time of signing the contract. Once the site
was ready, the customers sent back a certificate to this effect. The result was a sharp drop in
installation time, improved delivery to customers, as well as a cost reduction (communication to
the author).
For further information on Quality Councils, see Section 13, Strategic Planning, and Section 14,
Total Quality Management.
A call for nominations can produce large numbers of responses—numbers that are beyond the digestive
capacity of the organization. In such cases, an essential further step is screening to identify those
nominations which promise the most benefits for the effort expended.
To start with a long list of nominations and end up with a list of agreed projects requires an organized
approach—an infrastructure and a methodology. The screening process is time-consuming, so
the quality council usually delegates it to a secretariat, often the Quality Department. The secretariat
screens the nominations—it judges the extent to which the nominations meet the criteria set out
below. These judgments result in some preliminary decision making. Some nominations are rejected.
Others are deferred. The remainder are analyzed in greater depth to estimate potential benefits,
resources needed, and so on.
The quality councils and/or the secretariats have found it useful to establish criteria to be used
during the screening process. Experience has shown that there is need for two sets of criteria:
1. Criteria for choosing the first projects to be tackled by any of the project teams
2. Criteria for choosing projects thereafter
Criteria for the First Projects. During the beginning stages of project-by-project improvement,
everyone is in a learning state. Projects are assigned to project teams who are in training.
Completing a project is a part of that training. Experience with such teams has evolved a broad criterion:
The first project should be a winner. More specifically:
The project should deal with a chronic problem—one that has been awaiting solution for a long
The project should be feasible. There should be a good likelihood of completing it within a few
months. Feedback from companies suggests that the most frequent reason for failure of the first
project has been failure to meet the criterion of feasibility.
The project should be significant. The end result should be sufficiently useful to merit attention
and recognition.
The results should be measurable, whether in money or in other significant terms.
Criteria for Projects Thereafter. These criteria aim to select projects that will do the company
the most good:
Return on investment: This factor has great weight and is decisive, all other things being equal.
Projects that do not lend themselves to computing return on investment must rely for their priority
on managerial judgment.
The amount of potential improvement: One large project will take priority over several small ones.
Urgency: There may be a need to respond promptly to pressures associated with product safety,
employee morale, and customer service.
Ease of technological solution: Projects for which the technology is well developed will take
precedence over projects that require research to discover the needed technology.
Health of the product line: Projects involving thriving product lines will take precedence over
projects involving obsolescent product lines.
Probable resistance to change: Projects that will meet a favorable reception take precedence
over projects that may meet strong resistance, such as from the labor union or from a manager set
in his or her ways.
Some companies use a systematic approach to evaluate nominations relative to these criteria. This
yields a composite evaluation that then becomes an indication of the relative priorities of the nominations.
[For an example, see Hartman (1983); also see DeWollf et al. (1987).]
The end result of the screening process is a list of recommended projects in their order of priority.
Each recommendation is supported by the available information on compatibility with the criteria
and potential benefits, resources required, and so on. This list is commonly limited to matters in
which the quality council has a direct interest.
The quality council reviews the recommendations and makes the final determination on which
projects are to be tackled. These projects then become an official part of the company’s business.
Other recommended projects are outside the scope of the direct interest of the quality council. Such
projects are recommended to appropriate subcouncils, managers, and so on. None of the preceding
prevents projects from being undertaken at local levels by supervisors or by the work force.
Vital Few and Useful Many. During the 1980s, some companies completed many quality
improvement projects. Then, when questions were raised—“What have we gotten for all this
effort?”—they were dismayed to learn that there was no noticeable effect on the “bottom line.”
Investigation then showed that the reason was traceable to the process used for project selection. The
projects actually selected had consisted of
Firefighting projects: These are special projects for getting rid of sporadic “spikes.” Such projects
did not attack the chronic waste and hence could not improve financial performance.
Useful many projects: By definition, these have only a minor effect on financial performance.
Projects for improving human relations: These can be quite effective in their field, but the
financial results are usually not measurable.
To achieve a significant effect on the bottom line requires selecting the “vital few” projects as well
as the “useful many.” It is feasible to work on both, since different people are assigned to each.
There is a school of thought that contends that the key to quality leadership is “tiny improvements
in a thousand places”—in other words, the useful many (Gross 1989). Another school urges focus
on the vital few. In my experience, neither of these schools has the complete answer.
The vital few projects are the major contributors to quality leadership and to the bottom line. The
useful many projects are the major contributors to employee participation and to the quality of work
life. Each is necessary; neither is sufficient.
The vital few and useful many projects can
be carried out simultaneously. Successful companies
have done just that. They did so by recognizing
that while there are these two types of
projects, they require the time of different categories
of company personnel.
The interrelation of these two types of projects
is shown in Figure 5.11. In this figure, the
horizontal scale is time. The vertical scale is
chronic waste. What goes up is bad. The useful
many improvements collectively create a gradually
sloping line. The vital few improvements,
though less frequent, contribute the bulk of the
total improvement.
Cost Figures for Projects. To meet the
preceding criteria (especially that of return on
investment) requires information on various costs:
The cost of chronic waste associated with a given nomination
The potential cost reduction if the project is successful
The cost of the needed diagnosis and remedy
For the methodology of providing the cost figures, see above, under Getting the Cost Figures.
Costs versus Percent Deficiencies. It is risky to judge priorities based solely on the percentage
of deficiencies (errors, defects, and so on). On the face of it, when this percentage is low, the
priority of the nomination also should be low. In some cases this is true, but in others it can be seriously
In a large electronics company the percentage of invoices protested by customers was 2.4 percent.
While this was uncomfortable, it was below the average for similar processes in the industry.
Then a study in-depth showed that nearly half the time of the sales force was spent placating the
protesting customers and getting the invoices straightened out. During that time, the sales people
were not selling anything (communication to the author).
FIGURE 5.11 Interrelation of projects, vital few and
useful many.
Even more dramatic was the case of the invoices in Florida Power and Light Company. Protested
invoices ran to about 60,000 per year, which figured to about 0.2 percent of all invoices. The cost
to straighten these out came to about $2.1 million annually. A quality improvement project then
cut the percent errors to about 0.05 percent, at an annual saving of over $1 million. Even more
important was the improvement in customer relations and the related reduction of complaints to
the Public Utility Commission (Florida Power and Light Company 1984).
Elephant-Sized and Bite-Sized Projects. Some projects are “elephant-sized”; i.e., they
cover so broad an area of activity that they must be subdivided into multiple “bite-sized” projects. In
such cases, one project team can be assigned to “cut up the elephant.” Other teams are then assigned to
tackle the resulting bite-sized projects. This approach shortens the time to complete the project, since
the teams work concurrently. In contrast, use of a single team stretches the time out to several years.
Frustration sets in, team membership changes due to attrition, the project drags, and morale declines.
In the Florida Power and Light Company invoice case, the project required several teams, each
assigned to a segment of the invoicing process.
In Honeywell, Inc., a project to improve the information security system required creation of
seven teams, involving 50 team members. (See Parvey 1990.)
A most useful tool for cutting up the elephant is the Pareto analysis. For an application, see the paper
mill example earlier, under Use of the Pareto Principle.
For elephant-sized projects, separate mission statements (see below) are prepared for the broad
coordinating team and for each team assigned to a bite-sized project.
Cloning. Some companies consist of multiple autonomous units that exhibit much commonality.
A widespread example is the chains of retail stores, repair shops, hospitals, and so on. In such companies,
a quality improvement project that is carried out successfully in one operating unit logically
becomes a nomination for application to other units. This is called cloning the project.
It is quite common for the other units to resist applying the improvement to their operation. Some
of this resistance is cultural in nature (not invented here, and so on). Other resistance may be due to
real differences in operating conditions. For example, telephone exchanges perform similar functions
for their customers. However, some serve mainly industrial customers, whereas others serve mainly
residential customers.
Upper managers are wary of ordering autonomous units to clone improvements that originated
elsewhere. Yet cloning has advantages. Where feasible, it provides additional quality improvements
without the need to duplicate the prior work of diagnosis and design of remedy. What has emerged
is a process as follows:
Project teams are asked to include in their final report their suggestions as to sites that may be
opportunities for cloning.
Copies of such final reports go to those sites.
The decision of whether to clone is made by the sites.
However, the sites are required to make a response as to their disposition of the matter. This
response is typically in one of three forms:
1. We have adopted the improvement.
2. We will adopt the improvement, but we must first adapt it to our conditions.
3. We are not able to adopt the improvement for the following reasons.
In effect, this process requires the units to adopt the improvement or give reasons for not doing so.
The units cannot just quietly ignore the recommendation.
A more subtle but familiar form of cloning is done through projects that have repetitive application
over a wide variety of subject matter.
A project team develops computer software to find errors in spelling. Another team evolves an
improved procedure for processing customer orders through the company. A third team works up
a procedure for conducting design reviews.
What is common about such projects is that the end result permits repetitive application of the
same process to a wide variety of subject matter: many different misspelled words, many different
customer orders, and many different designs.
Each project selected should be accompanied by a written mission statement that sets out the intended
end result of the project. On approval, this statement defines the mission of the team assigned to
carry out the project.
Purpose of Mission Statements. The mission statement serves a number of essential
It defines the intended end result and so helps the team to know when it has completed the project.
It establishes clear responsibility—the mission becomes an addition to each team member’s job
It provides legitimacy—the project becomes official company business. The team members are
authorized to spend the time needed to carry out the mission.
It confers rights—the team has the right to hold meetings, to ask people to attend and assist the
team, and to request data and other services germane to the project.
The Numerical Goal. The ideal mission statement quantifies two critical elements: (1) the
intended amount of improvement and (2) the timetable.
Examples of such mission statements follow:
During the coming fiscal year, reduce the time to fill customer orders to an average of 1.5 days.
Reduce the field failure rate of product X by 50 percent over the next 3 years.
The numbers that enter the goals have their origin in various sources. They may originate in
Demands from customers who have their own goals to meet.
Actions taken by competitors, with associated threats to share of market.
Benchmarking to find the best results now being achieved. (The fact that they are being achieved
proves that they are achievable.)
In some cases, the available information is not enough to support a scientific approach to goal setting.
Hence the goal is set by consensus—by a “jury of opinion.”
Perfection as a Goal. There is universal agreement that perfection is the ideal goal—complete
freedom from errors, defects, failures, and so on. The reality is that the absence of perfection is due
to many kinds of such deficiencies and that each requires its own improvement project. If a company
tries to eliminate all of them, the Pareto principle applies:
The vital few kinds of deficiencies cause most of the trouble but also readily justify the resources
needed to root them out. Hence they receive high priority during the screening process and
become projects to be tackled.
The remaining many types of deficiencies cause only a small minority of the trouble. As one
comes closer and closer to perfection, each remaining kind of deficiency becomes rarer and rarer
and hence receives lower and lower priority during the screening process.
All companies tackle those rare types of failure which threaten human life or which risk significant
economic loss. In addition, companies that make improvements by the thousands year after year
tackle even the mild, rare kinds of deficiency. To do so they enlist the creativity of the work force
through such means as quality circles.
Some critics contend that publication of any goal other than perfection is proof of a misguided
policy—a willingness to tolerate defects. Such contentions arise from lack of experience with the
realities. It is easy to set goals that demand perfection now. Such goals, however, require companies
to tackle failure types so rare that they do not survive the screening process.
Nevertheless, there has been progress. During the twentieth century there was a remarkable
revision in the unit of measure for deficiencies. In the first half of the century, the usual measure
was in percent defective, or defects per hundred units. By the 1990s, many industries had adopted
a measure of defects per million units. The leading companies now do make thousands of
improvements year after year. They keep coming closer to perfection, but it is a never-ending
While many nominated projects cannot be justified solely on their return on investment, they may
provide the means for employee participation in the improvement process, which has value in its own
Publication of Mission Statements. Publication of the mission statement makes a project
an official part of company business. However, the quality council cannot predict precisely what the
project team will encounter as it tackles the project. Experience with numerous projects has provided
guidelines as to what to include (and exclude) from mission statements.
What to include: A mission statement may include information about the importance of the
problem. It may include data about the present level of performance as well as stating the intended
goal. It may include other factual information such as known symptoms of the problem.
What not to include: The mission statement should not include anything that may bias the
approach of the project team, such as theories of causes of the problem or leading questions. The
statement also should avoid use of broad terms (people problems, communication, and so on) for
which there are no agreed definitions.
(The preceding are derived from the training materials of Aluminum Company of America.)
Some companies separate the statement of the problem from the mission statement. In one Dutch
company, the quality council published a problem statement as follows:
The lead time of project-related components, from arrival to availability in the production departments,
is too long and leads to delays and interruptions in production.
The subsequent mission statement was as follows:
Investigate the causes of this problem and recommend remedies that would lead to a 50 percent
reduction in production delays within 3 months after implementation. A preliminary calculation
estimated the cost savings potential to be approximately 800,000 Dutch guilders ($400,000).
(Smidt and Doesema 1991)
Revision of Mission Statements. As work on the project progresses, the emerging new
information may suggest needed changes in the mission statement, changes such as the following:
The project is bigger than anticipated; it should be subdivided.
The project should be deferred because there is a prerequisite to be carried out first.
The project should change direction because an alternative is more attractive.
The project should be aborted because any remedy will be blocked.
Project teams generally have been reluctant to come back to the quality council for a revision of
the mission statement. There seems to be a fear that such action may be interpreted as a failure to
carry out the mission or as an admission of defeat. The result can be a dogged pursuit of a mission
that is doomed to failure.
The quality council should make clear to all project teams that they have the duty as well as the
right to recommend revision of mission statements if revision is needed. This same point also should
be emphasized during the training of project teams.
For each selected project, a team is assigned. This team then becomes responsible for completing the
Why a Team? The most important projects are the vital few, and they are almost invariably multifunctional
in nature. The symptoms typically show up in one department, but there is no agreement
on where the causes lie, what the causes are, or what the remedies should be. Experience has shown
that the most effective organizational mechanisms for dealing with such multifunctional problems
are multifunctional teams.
Some managers prefer to assign problems to individuals rather than to teams. (“A camel is a horse
designed by a committee.”) The concept of individual responsibility is in fact quite appropriate if
applied to quality control. (“The best form of control is self-control.”) However, improvement, certainly
for multifunctional problems, inherently requires teams. For such problems, assignment to
individuals runs severe risks of departmental biases in the diagnosis and remedy.
A process engineer was assigned to reduce the number of defects coming from a wave soldering
process. His diagnosis concluded that a new process was needed. Management rejected this conclusion,
on the ground of excess investment. A multifunctional team was then appointed to
restudy the problem. The team found a way to solve the problem by refining the existing process
(Betker 1983).
Individual biases also show up as cultural resistance to proposed remedies. However, such resistance
is minimal if the remedial department has been represented on the project team.
Appointment of Teams; Sponsors. Project teams are not attached to the chain of command
on the organization chart. This can be a handicap in the event that teams encounter an impasse.
For this reason, some companies assign council members or other upper managers to be sponsors (or
“champions”) for specific projects. These sponsors follow team progress (or lack of progress). If the
team does run into an impasse, the sponsor may be able to help the team get access to the proper person
in the hierarchy.
Teams are appointed by sponsors of the projects, by process owners, by local managers, or by
others. In some companies, work force members are authorized to form teams (quality circles, and
so on) to work on improvement projects. Whatever the origin, the team is empowered to make the
improvement as defined in the mission statement.
Most teams are organized for a specific project and are disbanded on completion of the project.
Such teams are called ad hoc, meaning “for this purpose.” During their next project, the members will
be scattered among several different teams. There are also “standing” teams that have continuity—
the members remain together as a team and tackle project after project.
Responsibilities and Rights. A project team has rights and responsibilities that are coextensive
with the mission statement. The basic responsibilities are to carry out the assigned mission
and to follow the universal improvement process (see below). In addition, the responsibilities include
Proposing revisions to the mission statement
Developing measurement as needed
Communicating progress and results to all who have a need to know
The rights of the teams were set out earlier, under Purpose of Mission Statements: convene meetings,
ask people for assistance, and request data and other services needed for the project.
Membership. The team is selected by the sponsor after consulting with the managers who are
affected. The selection process includes consideration of (1) which departments should be represented
on the team, (2) what level in the hierarchy team members should come from, and (3) which
individuals in that level.
The departments to be represented should include
The ailing department: The symptoms show up in this department, and it endures the effects.
Suspect departments: They are suspected of harboring the causes. (They do not necessarily
agree that they are suspect.)
Remedial departments: They will likely provide the remedies. This is speculative, since in
many cases the causes and remedies come as surprises.
Diagnostic departments: They are needed in projects that require extensive data collection and
On-call departments: They are invited in as needed to provide special knowledge or other services
required by the team (Black and Stump 1987).
This list includes the usual sources of members. However, there is need for flexibility.
In one company, once the team had gotten under way, it was realized that the internal customer—
a “sister facility”—was not represented. Steps were taken to invite the facility in, to avoid an “us
versus them” relationship (Black and Stump 1987).
Choice of level in the hierarchy depends on the subject matter of the project. Some projects relate
strongly to the technological and procedural aspects of the products and processes. Such projects
require team membership from the lower levels of the hierarchy. Other projects relate to broad business
and managerial matters. For such projects, the team members should have appropriate business
and managerial experience.
Finally comes the selection of individuals. This is negotiated with the respective supervisors, giving
due consideration to workloads, competing priorities, and so on. The focus is on the individual’s
ability to contribute to the team project. The individuals need
Time to attend the team meetings and to carry out assignments outside the meetings—“the
A knowledge base that enables the individual to contribute theories, insights, and ideas, as well
as job information based on his or her hands-on experience.
Training in the quality improvement process and the associated tools. During the first projects,
this training can and should be done concurrently with carrying out the projects.
Most teams consist of six to eight members. Larger numbers tend to make the team unwieldy as well
as costly. (A convoy travels only as fast as the slowest ship.)
Should team members all come from the same level in the hierarchy? Behind this question is the
fear that the biases of high-ranking members will dominate the meeting. Some of this no doubt takes
place, especially during the first few meetings. However, it declines as the group dynamics take over
and as members learn to distinguish between theory and fact.
Once the team is selected, the members’ names are published, along with their project mission.
The act of publication officially assigns responsibility to the individuals as well as to the team. In
effect, serving on the project team becomes a part of the individuals’ job descriptions. This same
publication also gives the team the legitimacy and rights discussed earlier.
Membership from the Work Force. During the early years of using quality improvement
teams, companies tended to maintain a strict separation of team membership. Teams for multifunctional
projects consisted exclusively of members from the managerial hierarchy plus professional
specialists. Teams for local departmental projects (such as quality circles) consisted exclusively of
members from the work force. Figure 5.12 compares the usual features of quality circles with those
of multifunctional teams.
Experience then showed that as to the details of operating conditions, no one is better informed
than the work force. Through residence in the workplace, workers can observe local changes and
recall the chronology of events. This has led to a growing practice of securing such information by
interviewing the workers. The workers become “on call” team members.
These same interviews have disclosed that many workers can contribute much more than knowledge
of workplace conditions. They can theorize about causes. They have ideas for remedies. In
addition, it has become evident that such participation improves human relations by contributing to
job satisfaction.
One result of all this experience has been a growing interest in broadening worker participation generally.
This has led to experimenting with project teams that make no distinction as to rank in the hierarchy.
These teams may become the rule rather than the exception. (For further discussion on the trends
in work force participation, see Section 15, Human Resources and Quality.)
Upper Managers on Teams. Some projects by their nature require that the team include
members from the ranks of upper management. Here are some examples of quality improvement
projects actually tackled by teams that included upper managers:
Shorten the time to put new products on the market.
Improve the accuracy of the sales forecast.
Reduce the carryover of prior failure-prone features into new product models.
Establish a teamwork relationship with suppliers.
Feature Quality circles Project teams
Primary purpose To improve human relations To improve quality
Secondary purpose To improve quality To improve participation
Scope of project Within a single department Multidepartmental
Size of project One of the useful many One of the vital few
Membership From a single department From multiple departments
Basis of membership Voluntary Mandatory
Hierarchical status of Typically in the workforce Typically managerial or professional
Continuity Circle remains intact, project Team is ad hoc, disbands after project
after project is completed.
FIGURE 5.12 Contrast, quality circles, and multifunctional teams. (From Making Quality Happen, 1988, Juran
Institute, Wilton, CT, p. D30.)
Develop the new measures of quality needed for strategic quality planning.
Revise the system of recognition and rewards for quality improvement.
There are some persuasive reasons urging that all upper managers personally serve on some
project teams. Personal participation on project teams is an act of leadership by example. This is
the highest form of leadership. Personal participation on project teams also enables upper managers
to understand what they are asking their subordinates to do, what kind of training is needed,
how many hours per week are demanded, how many months does it take to complete the
project, and what kinds of resources are needed. Lack of upper management understanding of such
realities has contributed to the failure of some well-intentioned efforts to establish annual quality
In one company, out of 150 quality improvement projects tackled, 12 involved teams composed
of senior directors (Egan 1985).
[For one upper manager’s account of his experience when serving on a project team, see Pelletier
Model of the Infrastructure. There are several ways to show in graphic form the infrastructure
for quality improvement—the elements of the organization, how they relate to each other,
and the flow of events. Figure 5.13 shows the elements of infrastructure in pyramid form. The pyramid
depicts a hierarchy consisting of top management, the autonomous operating units, and the
major staff functions. At the top of the pyramid is the corporate quality council and the subsidiary
councils, if any. Below these levels are the multifunctional quality improvement teams. (There may
be a committee structure between the quality councils and the teams).
At the intradepartment level are teams from the work force—quality circles or other forms. This
infrastructure permits employees in all levels of organization to participate in quality improvement
projects, the useful many as well as the vital few.
Quality improvement teams do not appear on the organization chart. Each “floats”—it has no personal
boss. Instead, the team is supervised impersonally by its mission statement and by the quality
improvement roadmap.
The team does have its own internal organizational structure. This structure invariably
includes a team leader (chairperson and so on) and a team secretary. In addition, there is usually
a facilitator.
FIGURE 5.13 Model of the infrastructure for quality improvement. (From
Visual GMQH15, Juran Institute, Inc., Wilton, CT.)
The Team Leader. The leader is usually appointed by the sponsor—the quality council or
other supervising group. Alternatively, the team may be authorized to elect its leader.
The leader has several responsibilities. As a team member, the leader shares in the responsibility
for completing the team’s mission. In addition, the leader has administrative duties. These are
unshared and include
Ensuring that meetings start and finish on time
Helping the members to attend the team meetings
Ensuring that the agendas, minutes, reports, and so on are prepared and published
Maintaining contact with the sponsoring body
Finally, the leader has the responsibility of oversight. This is met not through the power of
command—the leader is not the boss of the team. It is met through the power of leadership. The
responsibilities include
Orchestrating the team activities
Stimulating all members to contribute
Helping to resolve conflicts among members
Assigning the homework to be done between meetings
To meet such responsibilities requires multiple skills, which include
A trained capability for leading people
Familiarity with the subject matter of the mission
A firm grasp of the quality improvement process and the associated tools
The Team Secretary. The team secretary is appointed by the sponsor or, more usually, by the
team leader. Either way, the secretary is usually a member of the project team. As such, he or she
shares in the responsibility for carrying out the team mission.
In addition, the secretary has unshared administrative responsibilities, chiefly preparing the agendas,
minutes, reports, and so on. These documents are important. They are the team’s chief means of
communication with the rest of the organization. They also become the chief reference source for
team members and others. All of which suggests that a major qualification for appointment to the job
of secretary is the ability to write with precision.
The Team Members. “Team members” as used here includes the team leader and secretary.
The responsibilities of any team member consist mainly of the following:
Arranging to attend the team meetings
Representing his or her department
Contributing job knowledge and expertise
Proposing theories of causes and ideas for remedy
Constructively challenging the theories and ideas of other team members
Volunteering for or accepting assignments for homework
Finding the Time to Work on Projects. Work on project teams is time-consuming.
Assigning someone to a project team adds about 10 percent to that person’s workload. This added
time is needed to attend team meetings, perform the assigned homework, and so on. Finding the time
to do all this is a problem to be solved, since this added work is thrust on people who are already fully
No upper manager known to me has been willing to solve the problem by hiring new people to
make up for the time demanded by the improvement projects. Instead, it has been left to each team
member to solve the problem in his or her own way. In turn, the team members have adopted such
strategies as
Delegating more activities to subordinates
Slowing down the work on lower-priority activities
Improving time management on the traditional responsibilities
Looking for ongoing activities that can be terminated. (In several companies, there has been a
specific drive to clear out unneeded work to provide time for improvement projects.)
As projects begin to demonstrate high returns on investment, the climate changes. Upper managers
become more receptive to providing resources. In addition, the successful projects begin to
reduce workloads that previously were inflated by the presence of chronic wastes. [Relative to team
organization, see AT&T Quality Library, Quality Improvement Cycle (1988, pp. 7–12). Relative to
team meetings, see also AT&T Quality Improvement Team Helper (1990, pp. 17–21).]
Most companies make use of internal consultants, usually called “facilitators”, to assist quality
improvement teams, mainly teams that are working on their first projects. A facilitator is not a member
of the team and has no responsibility for carrying out the team mission. (The literal meaning of
the word facilitate is “to make things easy.”) The prime role of the facilitator is to help the team to
carry out its mission.
The Roles. The usual roles of facilitators consist of a selection from the following:
Explain the company’s intentions: The facilitator usually has attended briefing sessions that
explain what the company is trying to accomplish. Much of this briefing is of interest to the project
Assist in team building: The facilitator helps the team members to learn to contribute to the
team effort: propose theories, challenge theories of others, and/or propose lines of investigation.
Where the team concept is new to a company, this role may require working directly with individuals
to stimulate those who are unsure about how to contribute and to restrain the overenthusiastic
ones. The facilitator also may evaluate the progress in team building and provide feedback
to the team.
Assist in training: Most facilitators have undergone training in team building and in the quality
improvement process. They usually have served as facilitators for other teams. Such experiences
qualify them to help train project teams in several areas: team building, the quality improvement
roadmap, and/or use of the tools.
Relate experiences from other projects: Facilitators have multiple sources of such experiences:
 Project teams previously served
 Meetings with other facilitators to share experiences in facilitating project teams
 Final published reports of project teams
 Projects reported in the literature
Assist in redirecting the project: The facilitator maintains a detached view that helps to sense
when the team is getting bogged down. As the team gets into the project, it may find itself getting
deeper and deeper into a swamp. The project mission may turn out to be too broad, vaguely
defined, or not doable. The facilitator usually can sense such situations earlier than the team and
can help guide it to a redirection of the project.
Assist the team leader: Facilitators provide such assistance in various ways:
 Assist in planning the team meetings. This may be done with the team leader before each meeting.
 Stimulate attendance. Most nonattendance is due to conflicting demands made on a team member’s
time. The remedy often must come from the member’s boss.
 Improve human relations. Some teams include members who have not been on good terms with
each other or who develop friction as the project moves along. As an “outsider,” the facilitator
can help to direct the energies of such members into constructive channels. Such action usually
takes place outside the team meetings. (Sometimes the leader is part of the problem. In such
cases the facilitator may be in the best position to help out.)
 Assist on matters outside the team’s sphere of activity. Projects sometimes require decisions or
actions from sources that are outside the easy reach of the team. Facilitators may be helpful due
to their wider range of contacts.
Support the team members: Such support is provided in multiple ways:
 Keep the team focused on the mission by raising questions when the focus drifts.
 Challenge opinionated assertions by questions such as “Are there facts to support that theory?”
 Provide feedback to the team based on perceptions from seeing the team in action.
Report progress to the councils: In this role the facilitator is a part of the process of reporting
on progress of the projects collectively. Each project team issues minutes of its meetings. In due
course each also issues its final report, often including an oral presentation to the council.
However, reports on the projects collectively require an added process. The facilitators are often
a part of this added reporting network.
The Qualifications. Facilitators undergo special training to qualify them for the preceding
roles. The training includes skills in team building, resolving conflicts, communication, and management
of change; knowledge relative to the quality improvement processes, e.g., the improvement
roadmap and the tools and techniques; and knowledge of the relationship of quality improvement to
the company’s policies and goals. In addition, facilitators acquire maturity through having served on
project teams and having provided facilitation to teams.
This prerequisite training and experience are essential assets to the facilitator. Without them, he
or she has great difficulty winning the respect and confidence of the project’s team.
Sources and Tenure. Most companies are aware that to go into a high rate of quality improvement
requires extensive facilitation. In turn, this requires a buildup of trained facilitators. However, facilitation
is needed mainly during the startup phase. Then, as team leaders and members acquire training
and experience, there is less need for facilitator support. The buildup job becomes a maintenance job.
This phased rise and decline has caused most companies to avoid creating full-time facilitators or
a facilitator career concept. Facilitation is done on a part-time basis. Facilitators spend most of their
time on their regular job. [For an interesting example of a company’s thinking process on the question
of full-time versus part-time facilitators, see Kinosz and Ice (1991). See also Sterett (1987).]
A major source of facilitators is line supervisors. There is a growing awareness that service as a
facilitator provides a breadth of experience that becomes an aid on the regular job. In some companies,
this concept is put to deliberate use. Assignment to facilitation serves also as a source of training
in managing for quality. A second major source of facilitators is specialists. These are drawn
from the Human Relations Department or from the Quality Department. All undergo the needed
training discussed earlier.
A minority of large companies use a category of full-time specialists called “quality improvement
manager” (or similar title). Following intensive training in the quality improvement process, these
managers devote full time to the quality improvement activity. Their responsibilities go beyond facilitating
project teams and may include
Assisting in project nomination and screening
Conducting training courses in the quality improvement process
Coordinating the activities of the project team with those of other activities in the company
Assisting in the preparation of summarized reports for upper managers
(For elaboration on facilitators and their roles, see “Quality Improvement Team Helper,” a part of
AT&T’s Quality Library.)
A quality improvement team has no personal boss. Instead, the team is supervised impersonally. Its
responsibilities are defined in
The project mission statement: This mission statement is unique to each team.
The universal sequence2 (or roadmap) for quality improvement: This is identical for all teams.
It defines the actions to be taken by the team to accomplish its mission.
Some of the steps in the universal sequence have already been discussed in this section: proof of
the need, project nomination and selection, and appointment of project teams. The project team has
the principal responsibility for the steps that now follow—taking the two “journeys.”
The Two Journeys. The universal sequence includes a series of steps that are grouped into
two journeys:
1. The diagnostic journey from symptom to cause. It includes analyzing the symptoms, theorizing
as to the causes, testing the theories, and establishing the causes.
2. The remedial journey from cause to remedy. It includes developing the remedies, testing and
proving the remedies under operating conditions, dealing with resistance to change, and establishing
controls to hold the gains.
Diagnosis is based on the factual approach and requires a firm grasp of the meanings of key
words. It is helpful to define some of these key words at the outset.
Definition of Key Words
A “defect” is any state of unfitness for use or nonconformance to specification. Examples are
illegible invoice, oversizing, and low mean time between failures. Other names include “error”,
“discrepancy”, and “nonconformance.”
A “symptom” is the outward evidence of a defect. A defect may have multiple symptoms. The
same word may serve as a description of both defect and symptom.
A “theory” is an unproved assertion as to reasons for the existence of defects and symptoms.
Usually, multiple theories are advanced to explain the presence of defects.
A “cause” is a proved reason for the existence of a defect. Often there are multiple causes, in
which case they follow the Pareto principle—the vital few causes will dominate all the rest.
2 The concept of a universal sequence evolved from my experience first in Western Electric Company (1924–1941) and later
during my years as an independent consultant, starting in 1945. Following a few preliminary published papers, a universal
sequence was published in book form (Juran 1964). This sequence then continued to evolve based on experience gained from
applications by operating managers.
The creation of the Juran Institute (1979) led to the publication of the videocassette series Juran on Quality Improvement
(Juran 1981). This series was widely received and became influential in launching quality improvement initiatives in many companies.
These companies then developed internal training programs and spelled out their own versions of a universal sequence.
All these have much in common with the original sequence published in 1964. In some cases, the companies have come up with
welcome revisions or additions.
A “dominant cause” is a major contributor to the existence of defects and one that must be remedied
before there can be an adequate improvement.
“Diagnosis” is the process of studying symptoms, theorizing as to causes, testing theories, and
discovering causes.
A “remedy” is a change that can eliminate or neutralize a cause of defects.
Diagnosis Should Precede Remedy. It may seem obvious that diagnosis should precede
remedy, yet biases or outdated beliefs can get in the way.
For example, during the twentieth century many upper managers held deep-seated beliefs that
most defects were due to work force errors. The facts seldom bore this out, but the belief persisted.
As a result, during the 1980s, many of these managers tried to solve their quality problems
by exhorting the work force to make no defects. (In fact, defects are generally over 80 percent
management-controllable and under 20 percent worker-controllable.)
Untrained teams often try to apply remedies before the causes are known. (“Ready, fire, aim.”)
For example:
An insistent team member “knows” the cause and pressures the team to apply a remedy for that
The team is briefed as to the technology by an acknowledged expert. The expert has a firm opinion
about what is the cause of the symptom, and the team does not question the expert’s opinion.
As team members acquire experience, they also acquire confidence in their diagnostic skills. This
confidence then enables them to challenge unproved assertions.
Where deep-seated beliefs are widespread, special research may be needed.
In a classic study, Greenridge (1953) examined 850 failures of electronic products supplied by
various companies. The data showed that 43 percent of the failures were traceable to product
design, 30 percent to field operation conditions, 20 percent to manufacture, and the rest to miscellaneous
The diagnostic journey starts with analyzing the symptoms of the chronic quality problem. Evidence
of defects and errors comes in two forms:
The words used in written or oral descriptions
The autopsies conducted to examine the defects in-depth
Understanding the Symptoms. Symptoms are often communicated in words such as
incorrect invoices, machine produces poor copies, or “I don’t feel well.” Understanding such expressions
is often hindered because key words have multiple or vague meanings. In such cases, the person
who prepared the report becomes an essential source of information.
An inspection report persistently showed a high percentage of defects due to “contamination.”
Various remedies were tried to reduce contamination. All were unsuccessful. In desperation, the
investigators spoke with the inspectors to learn about the meaning of contamination. The inspectors
explained that there were 12 defect categories on the inspection form. If the observed defect
did not fit any of the categories, they would report the defect as contamination.
A frequent source of misunderstanding is the use of generic words to describe multiple subspecies
of defects.
In a plant making rubber products by the latex dip process, the word tears was used on the data
sheets to describe torn products. One important manager regarded tears as due to workers’ errors
and urged a remedy through motivational and disciplinary measures. Actually, there were three
species of tears: strip tears from a stripping operation, click tears from a press operation, and assembly
tears from an assembly operation. Only strip tears were due to worker errors, and their frequency
was only 15 percent. Revising the manager’s belief became possible only after clearing up
the meaning of the terminology and quantifying the relative frequencies of the subspecies of tears.
A useful tool for reducing semantic confusion is the “glossary.” A team is assigned to think out
the meanings of key words. The resulting agreements are then published as part of the official company
Autopsies. An important aid to understanding the meanings behind the words is the “autopsy”
(to see with one’s own eyes). Scientific autopsies can furnish extensive objective knowledge about
symptoms and thereby can supplement or override the information contained in the written reports.
The report on tests of a product may include a category of “electrical” defects. Autopsies of a
sample of such defects may show that there are multiple subspecies: open circuit, short circuit,
dead battery, and so on.
[For a case example of using autopsies, see Black and Stump (1987).]
All progress in diagnosis is made theory by theory—by affirming or denying the validity of the theories
about causes. The process consists of three steps: generating theories, arranging theories in
some order, and choosing theories to be tested.
Generating Theories. Securing theories should be done systematically. Theories should be
sought from all potential contributors—line managers and supervisors, technologists, the work force,
customers, suppliers, and so on. Normally, the list of theories is extensive, 20 or more. If only 3 or
4 theories have emerged, it usually means that the theorizing has been inadequate.
One systematic way of generating theories is called “brainstorming.” Potential contributors are
assembled for the specific purpose of generating theories. Creative thinking is encouraged by asking
each person, in turn, to propose a theory. No criticism or discussion is allowed until all theories are
recorded. The end result is a list of theories that are then subjected to discussion.
Experience has shown that brainstorming can have a useful effect on team members who carry
strong opinions. Such members may feel that their views should be accepted as facts. “I know this
is so.” However, other members regard these views as theories—unproved assertions. It all leads to
a growing awareness of the difference between theory and fact.
Another systematic approach—“nominal group technique”—is similar to brainstorming.
Participants generate their theories silently, in writing. Each then offers one theory at a time, in rotation.
After all ideas have been recorded, they are discussed and then prioritized by vote.
Theories should not be limited to those which relate to errors on specific products or processes.
In some cases, the cause may lie in some broader system that affects multiple products.
A manager observes, “In the last 6 weeks, we have lost four needed batches of unrelated products
due to four different instruments being out of calibration. This shows that we should review
our system for maintaining the accuracy of instruments.”
Arranging Theories. The brainstorming process provides a helter-skelter list of theories.
Orderly arrangement of such a list helps the improvement team to visualize the interrelation of the
theories. In addition, an orderly arrangement is an essential aid to choosing which theories to test.
The orderly arrangement can be made in several ways:
Storyboarding: A supplement to brainstorming, this is a form of orderly arrangement of theories.
As each theory is proposed, it is recorded on an index card. The cards are then appropriately
arranged on a board to form a visual display of the theories. [See Betker (1985) for an example
of use of storyboarding in an electronics company.]
Tabular arrangement: Another form of arrangement is a table showing a logical hierarchy: theories,
subtheories, sub-subtheories, and so on. Table 5.6 is an example as applied to yield of fine
powder chemicals.
Cause-and-effect diagram: This popular diagram (also known as an Ishikawa diagram or fishbone
diagram was developed in 1950 by the late Professor Kaoru Ishikawa. An example is shown
in Figure 5.14.
To create the diagram, the effect (symptom) is written at the head of the arrow. Potential causes
(theories) are then added to complete the diagram. A common set of major categories of causes consists
of personnel, work methods, materials, and equipment. Figure 5.14 shows the cause-and-effect
diagram as prepared for the same list of theories as was arranged in Table 5.6. Note how the diagram
aids in identifying interrelationships among theories.
Cause-and-effect diagrams were first applied to manufacturing problems. They have since
demonstrated that they are applicable to all manner of industries, processes, and problems. As a
result, they are now in universal use in every conceivable application.
A cause-and-effect diagram can be combined with a force-field analysis. The team identifies the
situations and events that contribute to the problem (these are the “restraining forces”). The actions
necessary to counter the restraining forces are then identified (these actions are the “driving forces”).
Finally, a diagram combining the restraining and driving forces is prepared to assist in diagnosis.
[For example, see Stratton (1987).]
Choosing Theories to Be Tested. Theories are numerous, yet most turn out to be invalid.
As a result, project teams have learned to discuss priorities for testing theories and to arrive at a con-
TABLE 5.6 Orderly Arrangement of Theories
Raw material
Shortage of weight
Method of discharge
Solution and concentration
B solution temperature
Solution and pouring speed
Stirrer, rpm
Mother crystal
Moisture content
Charging speed of wet powder
Dryer, rpm
Steam pressure
Steam flow
Overweight of package
Type of balance
Accuracy of balance
Maintenance of balance
Method of weighing
sensus. This approach has proved to be effective in reducing the teams’ time and effort, as well as in
minimizing the frustration of pursuing invalid theories.
Here and there companies have evolved structured matrixes for arriving at a quantitative score
for each theory. A simple method is to ask each team member to rank all theories in his or her
order of importance. The totals of the rank numbers then become an input to the final consensus
on priorities.
There are many strategies for testing theories, and they follow the Pareto principle—a relative few
of them are applicable to most problems. What follows is a brief description of some vital few strategies
along with their principal areas of application.
A critical question is whether to test one theory at a time, one group of interrelated theories at a
time, or all theories simultaneously. To make a proper choice requires an understanding of the methods
of data collection and analysis (see below). The team may be need to secure the advice of specialists
in data analysis.
The Factual Approach. The basic concept behind diagnosis is the factual approach—to make
decisions based on fact rather than on theory. This concept enables amateurs in the technology nevertheless
to contribute usefully to the project. Thus the teams must learn to distinguish theory from
fact. Facts are supported by suitable evidence. Theories are unsupported assertions. Sometimes the
distinction is subtle.
FIGURE 5.14 Ishikawa cause-and-effect diagram.
In one team, the engineering member asserted that changing the temperature of the solder bath
would reduce the frequency of the defect under study but would create a new defect that would make
matters worse. His belief was based on data collected over 10 years earlier on different equipment.
The team challenged his assertion, conducted a new trial, and found that the higher temperature
caused no such adverse effect (Betker 1983).
Flow Diagrams. For many products, the anatomy of the producing process is a “procession”—
a sequential series of steps, each performing a specific task. Most team members are familiar with
some of the steps, but few are familiar with the entire procession. Note that the steps in the procession
may include those within the external supplier chain as well as those taken during marketing,
use, and customer service.
Preparing a flow diagram helps all members to better understand the progression and the relation
of each step to the whole. [See, for example, Engle and Ball (1985).] (For details on constructing
flow diagrams, see Section 3, The Quality Planning Process.)
Process Capability Analysis. One of the most frequent questions raised by improvement
team members refers to “process capability.” Some members contend that “this process is inherently
unable to meet the specifications.” The opposing contention is that “the process is capable but it isn’t
being run right.” In recent decades, tools have been devised to test these assertions, especially as
applied to manufacturing processes.
A common test of process capability uses the “Shewart control chart.” Data are take from the
process at (usually) equal chronological intervals. Having established by control chart analysis that
the process is inherently stable, the data are then compared with the terms of the specification. This
comparison provides a measure of the ability of the process to consistently produce output within
specified limits. (For elaboration on the Shewart control chart, see Section 45.)
While evaluation of process capability originally was applied to manufacturing processes, it has
since been applied increasingly to administrative and business processes in all industries. A common
example has been the application to cycle time of such processes.
Many of these processes consist of a procession in which the work is performed in a sequence of
steps as it moves from department to department. It may take days (weeks, or even months) to complete
a cycle, yet the time required to do the work has taken only a few hours. The remaining time
has consisted of waiting for its turn at each step, redoing, and so on.
For such processes, the theoretical process capability is the cumulative work time. A person who is
trained to perform all the steps and has access to all the database might meet this theoretical number.
Some companies have set a target of cutting the cycle time to about twice the theoretical capability.
Process Dissection. A common test of why a capable process isn’t being run right is “process
dissection.” This strategy tries to trace defects back to their origins in the process. There are multiple
forms of such process dissection.
Test at Intermediate Stages. When defects are found at the end of a procession, it is not known
which operational step did the damage. In such cases, a useful strategy may be to inspect or test the
product at intermediate steps to discover at which step the defect first appears. Such discovery, if successful,
can drastically reduce the effort of testing theories.
Stream-to-Stream Analysis. High-volume products often require multiple sources (“streams”)
of production—multiple suppliers, machines, shifts, workers, and so on. The streams may seem to
be identical, but the resulting products may not be. Stream-to-stream analysis consists of separating
the production into streams of origin and testing for stream-to-stream differences in an effort to
find the guilty stream, if any.
Time-to-Time Analysis. Another form of process dissection is time-to-time analysis. The purpose
is to discover if production of defects is concentrated in specific spans of time. This type of analysis
has been used to study time between abnormalities, effect of change of work shifts, influence of the
seasons of the year, and many other such potential causes.
A frequent example of time-to-time analysis is the Shewhart control chart, which also can show
whether the variability in a process is at random or is due to assignable causes. (See Section 45.)
A special case of time-to-time changes is drift—a continuing deterioration of some aspect of the
process. For example, in factory operations, the chemical solution gradually may become more
dilute, the tools gradually may wear, or the workers may become fatigued.
In time-to-time analysis, the process (or product) is measured (usually) at equal time intervals.
Graphic presentation of the data is an aid to interpretation. Presentation in cumulative form (cumulative
sum charts) is an aid to detecting drift.
There are also “piece-to-piece” and “within-piece” variations.
An example of piece-to-piece variation is seen in foundry processes that produce castings in
“stacks.” In such cases, the quality of the castings may depend on their location in the stack. An
example of within-piece variation is in lathe operations, where the diameter of a cylindrical piece
is not uniform.
Simultaneous Dissection. Some forms of process dissection can test multiple theories
simultaneously. A classic example is the Multi-Vari3 chart. See Figure 5.15. In this figure, a vertical
line depicts the range of variation within a single unit of product, as compared with specification
tolerance limits. In the left-hand example, the within-piece variation alone is too great in relation to
the tolerance. Hence no improvement is possible unless within-piece variation is reduced. The middle
example is one in which within-piece variation is comfortable; the problem is piece-to-piece variation.
In the right-hand example, the problem is excess time-to-time variability. Traver (1983)
presents additional examples of Multi-Vari charts.
Defect Concentration Analysis. In “defect concentration analysis”, the purpose is to discover
concentrations that may point to causes. This method has been used in widely varied applications.
During one of the London cholera epidemics of the mid-nineteenth century, Dr. John Snow
secured the addresses of those in the Soho district who had died of cholera. He then plotted the
3The name Multi-Vari was given to this form of analysis by L. A. Seder in his classic paper, “Diagnosis with Diagrams,” in
Industrial Quality Control, (January 1950 and March 1950). The concept of the vertical line had been used by J. M. Juran, who
derived it from the method used in financial papers to show the ranges of stock prices.
FIGURE 5.15 Multi-Vari chart.
addresses on a map of that district. (See Figure 5.16.) The addresses were concentrated around
the Broad Street pump, which supplied drinking water for the Soho district. In those days, no one
knew what caused cholera, but a remedy was provided by removing the handle from the pump.
In the case of manufactured products, it is common to plot defect locations on a drawing of the
product. See Figure 5.17. This concentration diagram shows the location of defects on an office copier.
The circled numbers show various locations on the equipment. The numbers adjacent to the circles
show how many defects were found in the sample of machines under study. It is seen that locations
24 and 2 account for about 40 percent of the defects.
Concentration analysis has been applied to military operations.
During World War II, the United States Air Force studied the damage done to aircraft returning from
combat missions. One form of analysis was to prepare diagrams to show where enemy bullet holes
FIGURE 5.16 Dr. John Snow’s concentration analysis.
and other forms of damage were concentrated. The diagrams also seemed to show that some areas
of the aircraft never received damage. The conclusion was that damage to those areas had destroyed
the planes and that redesign was needed to reduce the vulnerability of those areas.
Association Searches. Some diagnosis consists of relating data on symptoms to some theory
of causation such as design, process, worker, and so on. Possible relationships are examined using
various statistical tools such as correlation, ranking, and matrixes.
Correlation: In this approach, data on frequency of symptoms are plotted against data on the
suspected cause. Figure 5.18 is an example in which the frequency of pitted castings was related
to the “choke” thickness in the molds.
Ranking: In this approach, the data on defects are ranked in their order of frequency. This ranking
is then compared with the incidence of the suspected cause.
Table 5.7 shows the frequency of the defect “dynamic unbalance” for 23 types of automotive
torque tubes. The suspected cause was a swaging operation that was performed on some of the product
types. The table shows which types had undergone swaging. It is clear that swaged product types
were much worse than the unswaged types.
In some cases, it is feasible to study data on multiple variables using a structured cookbook
method of analysis. An early published example is the SPAN plan (Seder and Cowan 1956). This
FIGURE 5.17 Concentration diagram: defects on copiers.
approach uses standardized data collection and analysis forms to permit successive separation of
observed total product variability into five stages: lot-to-lot, stream-to-stream, time-to-time, withinpiece
(or positional), and error of measurement. Other forms of search for association are set out in
the statistical group of Sections 44 to 48 of this handbook.
Cutting New Windows. In some cases, the data available from operations are not able to test
certain of the theories. In such cases, it may be necessary to create new data specifically for the purpose
of testing theories. This is called “cutting new windows” and takes several forms.
Measurement at Intermediate Stages. A common example is seen in products made by a procession
of steps but tested only after completion of all steps. (See preceding, under Process Dissection.)
In such cases, cutting new windows may consist of making measurements at intermediate stages of
the procession.
In a project to reduce the time required to recruit new employees, data were available on the total
time elapsed. Test of the theories required cutting new windows by measuring the time elapsed
for each of the six steps in the recruitment process.
FIGURE 5.18 Test of theories by correlation.
TABLE 5.7 Test of Theories by Ranking
Type % defective Swaged (marked X) Type % defective Swaged (marked X)
A 52.3 X M 19.2 X
B 36.7 X N 18.0 X
C 30.8 X O 17.3
D 29.9 X P 16.9 X
E 25.3 X Q 15.8
F 23.3 X R 15.3
G 23.1 X S 14.9
H 22.5 T 14.7
I 21.8 X U 14.2
J 21.7 X V 13.5
K 20.7 X W 12.3
L 20.3
In a process for welding of large joints in critical pressure vessels, all finished joints were x-rayed
to find any voids in the welds. The process could be dissected to study some sources of variation:
worker-to-worker, time-to-time, and joint-to-joint. However, data were not available to study
other sources of variations: layer-to-layer, bead-to-bead, and within bead. Cutting new windows
involved x-raying some welds after each bead was laid down.
Creation of New Measuring Devices. Some theories cannot be tested with the measuring devices
used during operations. In such cases, it may be necessary to create new devices.
In a project to reduce defects in automotive radiators, some theories focused on the heat-treating
and drying operations that occurred inside a closed brazing oven. To measure what was happening
inside the oven, an insulated box—about the size of a radiator—was equipped with thermocouples
and designed to log time and temperatures within the oven. The box was placed on the
assembly line along with the radiators and sent through the oven on a normal brazing cycle. The
resulting data were used to modify the temperature profile inside the oven. Down went the failure
rate (Mizell and Strattner 1981).
Nondissectable Features. A “dissectable” product feature is one that can be measured during various
stages of processing. A “nondissectible” feature cannot be measured during processing; many
nondissectible features do not even come into existence until all steps in the process have been completed.
A common example is the performance of a television set. In such cases, a major form of test
of theories is through design of experiments (see below).
Design of Experiments. Test of theories through experiment usually involves producing trial
samples of product under specially selected conditions. The experiment may be conducted either in
a laboratory or in the real world of offices, factories, warehouses, users’ premises, and so on.
It is easy enough to state the “minimal criteria” to be met by an experiment. It should
Test the theories under study without being confused by extraneous variables
Discover the existence of major causes even if these were not advanced as theories
Be economic in relation to the amounts at stake
Provide reliable answers
To meet these criteria requires inputs from several sources:
The managers identify the questions to which answers are needed.
The technologists select and set priorities on the proper variables to be investigated.
The diagnosticians provide the statistical methods for planning the experimental design and analyzing
the resulting data.
Designs of experiments range from simple rifleshot cases to the complex unbridled cases, and
most of them are not matters to be left to amateurs. In its simplest form, the “rifleshot experiment”
uses a split-lot method to identify which of two suspects is the cause. For example, if processes A
and B are suspects, a batch of homogeneous material is split. Half goes through process A; half goes
through process B. If two types of material are also suspects, each is sent through both processes, A
and B, creating a two-by-two design of experiment. As more variables get involved, more combinations
are needed, but now the science of design of experiments enters to simplify matters.
In the “unbridled experiment”, a sample (or samples) of product are followed through the various
processes under a plan that provides for measuring values of the selected suspects at each stage.
The resulting product features are also measured. The hope is that analysis of the resulting data will
find the significant relationships between causes and effects.
The unbridled experiment should be defined in writing to ensure that it is understood and that
it represents a meeting of the minds. Carefully planned experiments have a high probability of
identifying the guilty suspects. The disadvantage is the associated cost and the time interval needed
to get answers.
Statisticians have developed remarkably useful tools: to get rid of unwanted variables through
“randomizing”; to minimize the amount of experimentation through skillful use of factorial, blocked,
nested, and other designs; to read the meaning out of complex data. (See Section 47, Design and
Analysis of Experiments.)
Measurement for Diagnosis. A frequent roadblock to diagnosis is the use of shop instruments
to make the measurements. These instruments were never intended to be used for diagnosis. They were
provided for other purposes such as process regulation and product testing. There are several principal
categories of cases in which measurement for diagnosis differs from measurement for operations:
Measurement by variables instead of attributes.
Process capability studies usually demand
variables measurements.
Measurement with a precision superior to
that of the shop instruments. In some cases,
the instruments provided for operations lack
adequate precision and hence are a dominant
cause of the quality problem.
Creation of new instruments to cut new windows
or to deal with nondissectible processes.
Measurement to test suspected variables
that are not controlled or even mentioned
by the specifications.
Responsibility for Diagnosis. Some of
the work of diagnosis consists of the discussions
that take place during project team meetings:
analyzing symptoms, theorizing about
causes, selecting theories for test, and so on. In
addition, the work of diagnosis involves test of
theories, which consists mainly of data collection
and analysis and is done largely as homework
outside of team meetings.
For some projects, the homework consists of
data collection and analysis on a small scale. In
such cases, the project team members themselves
may be able to do the homework. Other
projects require extensive data collection and
analysis. In such cases, the project team may
delegate or subcontract much or all of the work
to diagnosticians—persons who have the needed
time, skills, and objectivity. Despite such delegation,
the project team remains responsible for
getting the work done.
In large organizations working on many
improvement projects, the work of diagnosis
occupies the full-time equivalent of numerous
diagnosticians. In response, many companies
create full-time job categories for diagnosis,
under such titles as quality engineer. Where to
locate these on the organizational chart has led
to several alternatives. (See Figure 5.19.)
FIGURE 5.19 Alternatives for organization of diagnosticians.
1. The diagnosticians are assigned to line managers in proportion to the needs of their departments.
(See Figure 5.19a). This arrangement is preferred by line managers. In practice, these arrangements
tend to end up with the diagnosticians being assigned to help the line managers meet current
goals, fight fires, and so on. Such assignments then take priority over the chronic problems.
2. The diagnosticians are assigned to the various line managers (as above) but with a “dotted line”
running to a central diagnostic department such as Quality Engineering. (See Figure 5.19b). This
arrangement is better from the standpoint of training diagnosticians, offering them an obvious
career path and providing them with consulting assistance. However, the arrangement runs into
conflicts on the problem of priorities—on which projects should the diagnosticians be working.
3. The diagnosticians are assigned to a central diagnostic department such as Quality Engineering.
(See Figure 5.19c). This arrangement increases the likelihood that chronic projects will have adequate
priority. In addition, it simplifies the job of providing training and consulting assistance for
diagnosticians. However, it makes no specific provision for line manager participation in choice
of projects or in setting priorities. Such an omission can be fatal to results.
4. The diagnosticians are assigned to a central department but with a structured participation by the
line managers. (See Figure 5.19d). In effect, the line managers choose the projects and establish
priorities. The diagnostic department assigns the diagnosticians in response to these priorities.
It also provides training, consulting services, and other assistance to the diagnosticians. This
arrangement is used widely and has demonstrated its ability to adapt to a wide variety of company
The choice among these (and other) alternatives depends on many factors that differ from one company
to another.
Lessons learned are based on experience that is derived from prior historical events. These events
become lessons learned only after analysis—“retrospective analysis.”
An enormous amount of diagnosis is done by analysis of historical events. A common example
is seen in quality control of an industrial process. It is done by measuring a sample of units of product
as they emerge from the process. Production of each unit is a historical event. Production of multiple
units becomes multiple historical events. Analysis of the measurements is analysis of historical
events and thereby an example of retrospective analysis.
The Santayana Review. A short name is needed as a convenient label for this process of retrospective
analysis. I have proposed calling it the Santayana review. The philosopher George
Santayana once observed that “Those who cannot remember the past are condemned to repeat it.”
This is a terse and accurate expression of the concept of lessons learned through retrospective analysis.
The definition becomes:
The Santayana review is the process of deriving lessons learned from retrospective analysis of
historical events.
The Influence of Cycle Time and Frequency. Use of the Santayana review has
depended largely on
The cycle time of the historical events
The frequency of these same events, which is closely correlated with their cycle time
The influence of these two factors, cycle time and frequency, is best understood by looking at a few
Application to High-Frequency Cycles. High-frequency events abound in companies
of all kinds. The associated processes are of a mass production nature, and they process various
Industry Mass Processing of
Utilities Invoices
Factories Goods
All industries Payroll checks
The resulting cycles can number millions and even billions annually. Nevertheless, many companies
manage to run these processes at extremely low levels of error. They do so by analysis of samples
from the processes—by analyzing data from historical events.
It is fairly easy to apply the Santayana review in such mass production cases. The data are available
in large numbers—sampling is a necessity to avoid drowning in data. The data analysis is often
simple enough to be done locally by personnel trained in basic statistics. The effort involved is modest,
so there is seldom any need to secure prior approval from higher levels. As a result, the
Santayana review is widely applied. Of course, those who make such applications seldom consider
that they are engaged in a study of prior historical events. Yet this is precisely what they are doing.
Application to Intermediate-Frequency Cycles. As used here, “intermediate frequency”
is an order of magnitude of tens or hundreds of cycles per year—a few per month or week. Compared
with mass production, these cycles are longer, each involves more functions, each requires more effort,
and more is at stake. Examples within this range of frequency include recruitment of employees or bids
for business.
Applications of the Santayana review to intermediate-frequency cycles have been comparatively
few in number, but the opportunities abound. It is obviously desirable to reduce the time needed to
recruit employees. It is also desirable to increase the percentage of successful bids. (In some industries,
the percentage is below 10 percent). The low level of retrospective analysis is traceable to some
realities of the Santayana review as it applies to intermediate-frequency cycles :
The application is to a multifunctional process, usually requiring a team effort.
It can require a lot of work now, for benefits to come later, and with no ready way of computing
return on investment.
There is rarely a clear responsibility for doing the work.
The urge to volunteer to do the work is minimal, since the improvement will benefit the organization
generally but not necessarily the volunteer’s department.
(These realities do not preclude application of the Santayana review to high-frequency cycles,
since usually the application is to departmental processes, the amount of work is small, and the urge
to volunteer is present because the results will benefit the volunteer’s department).
Application to Low-Frequency Cycles. As used here, “low frequency” refers to a range
of several cycles per year down to one cycle in several years. Examples on an annual schedule
include the sales forecast and the budget. Examples on an irregular schedule include new product
launches, major construction projects, and acquisitions.
Application of the Santayana review to low-frequency cycles has been rare. Each such cycle is a
sizable event; some are massive. A review of multiple cycles becomes a correspondingly sizable
An example is the historical reviews conducted by a team of historians in British Petroleum
Company. This team reviews large business undertakings: joint ventures, acquisitions, and major
construction projects. The reviews concern matters of business strategy rather than conformance
to functional goals. Each review consumes months of time and requires about 40 interviews to
supply what is not in the documented history. The conclusions and recommendations are presented
to the highest levels (Gulliver 1987).
A widespread low-frequency process that desperately needs application of the Santayana review
is the launching of new products. Such launchings are carried out through an elaborate multifunctional
process. Each product launched has a degree of uniqueness, but the overall process is quite
similar from one cycle to another. Such being the case, it is entirely feasible to apply the Santayana
Much of the time required during the launch cycle consists of redoing what was done previously.
Extra work is imposed on internal and external customers. The extent and cost of these delays can
be estimated from a study of prior cycles. Retrospective analysis can shed light on what worked and
what did not and thereby can improve decision making.
Note that the bulk of this delay and cost does not take place within the product development
department. An example is seen in the launch of product X that incurred expenses as follows (in
Market research 0.5
Product development 6.0
Manufacturing facilities 22.0
Marketing planning 22.0
Total 30.5
All this was lost because a competitor captured the market by introducing a similar product 2
years before the launch of product X. The bulk of the loss—80 percent—took place outside the product
development department.
Some Famous Case Examples. The potential of the Santayana review can best be seen
from some famous historical case examples.
Sky watchers and calendars: One of the astounding achievements of ancient civilizations was
the development of precise calendars. These calendars were derived from numerous observations
of the motions of celestial bodies, cycle after cycle. Some of these cycles were many years in
length. The calendars derived from the data analysis were vital to the survival of ancient societies.
For example, they told when to plant crops.
Prince Henry’s think tank: During the voyages of discovery in the fifteenth and sixteenth centuries,
Portuguese navigators were regarded as leaders in guiding ships to their destinations and
bringing them back safely. As a result, Portuguese navigators were preferred and demanded by
ship owners, governments, and insurers. The source of this leadership was an initiative by a
Portuguese prince—Prince Henry the Navigator (1394–1460.) In the early 1400s, Prince Henry
established (at Sagres, Portugal) a center for marine navigation—a unique, unprecedented think
tank. The facilities included an astronomical observatory, a fortress, a school for navigators, living
quarters, a hospital, and a chapel. To this center, Prince Henry brought cartographers, instrument
makers, astronomers, mathematicians, shipwrights, and drafters. He also established a data bank—
a depository of logs of marine voyages describing prevailing winds, ocean currents, landmarks,
and so on. Lessons learned from these logs contributed to Portuguese successes during the voyages
of discovery around the coast of Africa, through the Indian Ocean, and across the Atlantic.
Mathew Maury’s navigation charts: In the mid-nineteenth century, Mathew Maury, a U.S.
Navy lieutenant, analyzed the logs of thousands of naval voyages. He then entered the findings
(current speeds, wind directions, and so on) on the navigation charts using standardized graphics
and terminology. One of the first ships to use Maury’s charts was the famous Flying Cloud. In
1851 it sailed from New York to San Francisco in 89 days. The previous record was 119 days
(Whipple 1984). The new record then endured for 138 years!
Research on recurring disasters: Some individual disasters are so notorious that the resulting
glare of publicity forces the creation of a formal board of inquiry. However, the most damage is
done by repetitive disasters that, although less than notorious individually, are notorious collectively.
Some institutions exist to study these disasters collectively. At their best, these institutions
have contributed mightily to the wars against diseases, to reduction of accidents, and to making
buildings fireproof. A fascinating example is a multinational study to shed light on the relation of
diet to cancer. Figure 5.20 shows the resulting correlation (Cohen 1987).
The Potential for Long-Cycle Events. The usefulness of the Santayana review has been
amply demonstrated in the case of short-cycle, high-frequency activities. As a result, the Santayana
review is widely applied to such cases and with good effect. The opportunities for application to
long-cycle, low-frequency activities are enormous. However, the actual applications have been comparatively
rare due to some severe realities.
Sponsorship requires a consensus among multiple managers rather than an initiative by one
The associated work of the diagnostician is usually extensive and intrudes on the time of others.
The resulting lessons learned do not benefit current operations. The benefits apply to future operations.
The results do not necessarily benefit the departmental performances of participating managers.
There is no ready way to compute return on investment.
It is understandable that projects facing such realities have trouble in securing priorities. As matters
stand, an initiative by upper managers is needed to apply the Santayana review to long-cycle
FIGURE 5.20 Correlation, diet and cancer.
activities. To date, such initiatives have been few, and published papers have been rare. The paper
relative to the experience at British Petroleum is decidedly an exception (Gulliver 1987).
Will the pace of application accelerate? I doubt it. My prognosis is that the pace will remain evolutionary
until some spectacular result is achieved and widely publicized. This is a discouraging forecast,
the more so in the light of the quotation from Santayana: “Those who cannot remember the past
are condemned to repeat it.” (For extensive additional discussion and case examples, see Juran 1992.)
Once the causes are established, the diagnostic journey is over, and the remedial journey begins.
While each remedy is unique to its project, the managerial approach to selecting and applying remedies
is common to all projects.
Choice of Alternatives. For most projects, there are multiple proposals for remedy. Choice
of remedy then depends on the extent to which the proposals meet certain essential criteria. The proposed
remedies should
Remove or neutralize the cause(s)
Optimize the costs
Be acceptable to those who have the last word
Remedies: Removing the Causes. Proposed remedies typically must clear three hurdles
before becoming effective:
1. The project team accepts the proposal based on logical reasoning—on its belief that the proposed
remedy will meet the preceding criteria.
2. The proposal is tested out on a small scale, whether in operations or in the laboratory.
3. The proposal is tested full scale during operations.
In many companies a fourth hurdle has existed; the responsibility of project teams is vague, or
limited to recommending remedies, with no responsibility to follow through. In such cases, many
recommendations are simply not acted on. Results are much better in companies that make the teams
responsible for ensuring that the remedies are in fact applied and that they are effective under operating
Many remedies consist of technological changes. These encounter the biases of some remedial
departments, such as favoring remedies that involve buying new facilities. In many cases, however,
the optimal remedy is through making better use of existing facilities. [For examples, see Black and
Stump (1987); also see Bigelow and Floyd (1990).]
Actually, the remedies with the highest return on investment have involved managerial changes
rather than technological changes. Dramatic evidence of this was seen when teams from the United
States visited their Japanese counterparts to learn why Japanese quality was superior. Such visits
were made to plants making steel, rubber tires, die castings, large-scale integrated circuits, automobiles,
and so on. The Americans were astonished to find that the Japanese facilities (machinery, tools,
instruments, and so on) were identical to those used in the American plants—they had even been
bought from the same suppliers. The difference in quality had resulted from making better use of the
existing facilities (personal experience of the author).
Still other remedies consist of revising matters of a broad managerial nature—policies, plans,
organization, standards, procedures. Such remedies have effects that extend well beyond the specific
project under study. Getting such remedies accepted requires special skills in dealing with cultural
resistance (see below).
Occasionally, remedies can be remarkably imaginative. In a plant making chips for integrated circuits,
a vibration problem caused by a nearby railroad was solved by constructing a swimming pool
between the plant and the railroad. Another problem was due to cement dust from an adjacent concrete
mixing plant. The remedy: Buy the plant and demolish it.
For some chronic quality problems, the remedy consists of replanning some aspect of the process
or product in question. (For the methodology, see Section 3, The Quality Planning Process.)
Remedies: Optimizing the Costs. In complex processes it is easy to apply a remedy that
reduces costs in department A, only to learn that this cost reduction is more than offset by increased
costs in department B. The cure can be worse than the disease. The project team should check out
the side effects of the remedy to ensure that the costs are optimal for the company. This same check
should extend to the effect on external customers’ costs.
A well-chosen project team can do much to optimize costs because the membership is multifunctional.
However, the team should look beyond the functions of its members. It also should enlist the aid
of staff personnel from departments such as finance to assist in reviewing the figures and estimates.
(For details on quantifying quality-related costs, see Section 8, Quality and Costs.)
Remedies: Acceptability. Any remedy involves a change of some kind—redesign the product
or process, revise the tool, and/or retrain the worker. Each such change falls within the jurisdiction of
some functional department that then becomes the remedial department for the project in question.
Normally the jurisdictional lines are respected, so the responsibility for making the change lies with the
remedial department, not with the project team.
All this is simplified if someone from the remedial department is a member of the project team,
which is usually the case. Such a member keeps his or her superiors informed and thereby helps to
ensure that the proposed remedy will be adopted forthwith.
Matters are more complex if the remedial department has not been represented on the project
team. Now the team must recommend that the remedial department adopt the remedy. This recommendation
may encounter resistance for cultural reasons, including possible resentment at not having
been represented. The project team is then faced with trying to convince the remedial department
of the merits of the change. In the event of an impasse, the team may appeal through its sponsor or
in other ways, such as through channels in the hierarchy.
Ideally, the remedial department is represented on the team from the outset. This is not always
feasible—at the outset it is not known what will turn out to be the causes and hence the remedy.
However, once the nature of the remedy becomes evident, the corresponding remedial department
should be invited to join the team.
The concept of anticipating resistance applies to other sources as well—the union, the local community,
and so on. The team is well advised to look for ways to establish a dialogue with those who
are potentially serious opponents of the remedy. (For discussion of cultural resistance, see below
under Resistance to Change.)
The Remedy for Rare but Critical Defects. Some defects, while rare, can result in catastrophic
damage to life or property. For such defects, there are special remedies.
Increase the factor of safety through additional structural material, use of exotic materials, design
for misuse as well as intended use, fail-safe design, and so on. Virtually all of these involve an
increase in costs.
Increase the amount and severity of test. Correlation of data on severe tests versus normal tests
then provides a prediction of failure rates.
Reduce the process variability. This applies when the defects have their origin in manufacture.
Use automated 100 percent test. This concept has been supported recently by a remarkable
growth in the technology: nondestructive test methods, automated testing devices, and computerized
Use redundant 100 percent inspection. Inspection by human beings can be notoriously fallible.
To find rare but critical defects, use can be made of multiple 100 percent inspections.
Remedy through Replication. One form of replication of remedies is cloning, as discussed
earlier in this section under Project Selection, Cloning. Through cloning, a remedy developed in one
project may have application elsewhere in the same company. Replication also may be achieved
through a generic remedy that applies to an assortment of error types.
Office work has long had the annoying problem of misspelled words. These misspellings are scattered
among numerous different words. Now, word processing programs include a dictionary in
their memory as a means of detecting misspelled words. The planners found a way to deal with
numerous error types, each of which is comparatively rare.
Test under Operating Conditions. Remedies are often tested in the laboratory before
being adopted. A common approach is to develop a theoretical model and then construct and test
some prototypes. This is a valuable step that can screen out inadequate remedies. Yet it is limited as
a predictor of results in the real world of operations.
The theoretical model is based on assumptions that are never fully met.
The prototypes are constructed in a laboratory environment rather than in the operating environment.
The testing is done on a small sample size and under closely controlled test conditions.
The testing is done by trained technicians under the guidance of supervisors and engineers.
These and other limitations create the risk that the remedy, despite having passed its laboratory
examination with flying colors, will not prove adequate under operating conditions. This has led
some companies to require that the project team remain attached to the project until the remedy has
been proved under operating conditions.
Control at the New Level; Holding the Gains. To enable the operating forces to hold
the gains requires (1) a successful transfer of the remedy from the laboratory to operations and (2) a
systematic means of holding the gains—the control process. Ideally, the remedial change should be
irreversible. Failing this, it may be necessary to conduct periodic audits to ensure that the change
remains in place.
In a famous foundry project, one change involved the replacement of old ladle spouts with largerdiameter
spouts. To make the change irreversible, the old spouts were destroyed. A different
remedy required the melters to use scales to weigh accurately the amount of metal to be poured.
This change could be reversed—some melters did not use the scales; they went right back to estimating
by eye and feel.
Transfer to operations should include the revisions in operating standards, procedures, and so on
needed to serve as a basis for training, control, and audit. These matters tend to be well defined with
respect to the technology. In contrast, standards and procedures are often vague or silent on matters
such as why the criteria should be met, what can happen if they are not met, equipment maintenance,
and work methods. Failure to deal with these latter areas can be a threat to holding the gains.
Transfer to operations should include transfer of information related to the change. This transfer
may require formal training in the use of the new processes and methods. It helps if the training also
extends to the reasons behind the change, the resulting new responsibilities for decisions and actions,
and the significant findings that emerged during the project.
The final step is establishing controls to hold the gains. This is done through the feedback loop—a
cyclic process of evaluating actual performance, comparing this with the standard, and taking action on
the difference. (Various aspects of the control process are discussed in Section 4, The Quality Control
Process; Section 45, Statistical Process Control; and Section 11, ISO 9000 Family of Standards.)
In some projects, the contributing causes include human error. Such errors are committed by all
human beings—managers, supervisors, professional specialists, and the work force. Except for work
force errors, the subject has received very little research, so the database is small. In view of this,
what follows focuses on work force errors.
Extent of Work Force Errors. Most errors are controllable by management. Errors are controllable
by workers only if the criteria for self-control have all been met—if the worker has the
means of
Knowing what he or she is supposed to do
Knowing what is his or her actual performance
Regulating his or her performance
Investigators in many countries have conducted studies on controllability. As reported to me,
these generally confirm my own conclusion that in industry, by and large, controllability prevails as
Management-controllable: over 80 percent
Worker-controllable: under 20 percent
Species of Work Force Error. It has long been a widely held belief by managers that work
force errors are due to lack of motivation. However, recent research has shown that there are multiple
species of work force errors and that only a minority of such errors have their origin in lack of
Table 5.8 shows the distribution of 80 errors made by six office workers engaged in preparing
insurance policy contracts. There are 29 types of errors, and they are of multiple origins.
Error type 3 was made 19 times, but worker B made 16 of them. Yet, except for error type 3,
worker B makes few errors. Seemingly, there is nothing wrong with worker B, except on defect type
3. Seemingly also, there is nothing wrong with the job instructions, since no one else had trouble
with error type 3. It appears that worker B and no one else is misinterpreting some instruction, resulting
in that clump of 16 errors of type 3.
TABLE 5.8 Matrix of Errors by Insurance Policy Writers
Policy writer
Error type A B C D E F Total
1 0 0 1 0 2 1 4
2 1 0 0 0 1 0 2
3 0 16 1 0 2 0 19
4 0 0 0 0 1 0 1
5 2 1 3 1 4 2 13
6 0 0 0 0 3 0 3
• • • • • • • •
• • • • • • • •
• • • • • • • •
TOTAL 6 20 8 3 36 7 80
Error type 5 is of a different species. There are 13 of these, and every worker makes this error,
more or less uniformly. This suggests a difference in approach between all the workers on the one
hand and the inspector on the other. Such a difference is usually of management origin, but the realities
can be established by interviews with the respective employees.
A third phenomenon is the column of numbers associated with worker E. The total is 36 errors—
worker E made nearly half the errors, and he or she made them in virtually all error type categories.
Why did worker E make so many errors? It might be any of several reasons, such as inadequate training,
lack of capability to do exacting work, and so on. Further study is needed, but some managers
might prefer to go from symptom directly to remedy—find a less demanding job for that worker.
This single table of data demonstrates the existence of multiple species of worker error. The remedy
is not as simplistic as “motivate the worker.” Analysis of many such tables, plus discovery of the
causes, has identified four principal species of work force error: inadvertent, technique, conscious,
and communication. Table 5.9 shows the interrelations among the error patterns, the likely subspecies,
and the likely remedies. The error species are examined below.
Inadvertent Errors. “Inadvertent” means “caused by lack of attention.” Inadvertent errors are
made because of human inability to maintain attention. (Ancient generals and admirals limited the
length of the sentry’s watch because of the risk of lack of attention.) (If not paying attention is deliberate,
then the resulting errors are conscious rather than inadvertent.)
Diagnosis to identify errors as inadvertent is aided by understanding their distinguishing features.
They are
Unintentional: The worker does not want to make errors.
Unwitting: At the time of making an error, the worker is unaware of having made it.
Unpredictable: There is nothing systematic as to when the next error will be made, what type
of error will be made, or which worker will make the error. Due to this unpredictability, the
error pattern exhibits randomness. Conversely, data that show a random pattern of worker error
TABLE 5.9 Interrelation among Human Error Patterns
Pattern disclosed by analysis Likely subspecies of error
of worker error causing this pattern Likely solution
On certain defects, no one is
error-prone; defect pattern
is random.
On certain defects, some
workers are consistently
error-prone, while others are
consistently “good.”
Some workers are consistently
error-prone over a wide
range of defects.
On certain defects, all workers
are error-prone.
Errors are due to inadvertence.
Errors are due to lack of technique
(ability, know-how,
etc.). Lack of technique may
take the form of secret ignorance.
Technique may consist
of known knack or of
secret knowledge.
There are several potential
Conscious failure to comply
to standards.
Inherent incapacity to perform
this task.
Lack of training.
Errors are management
Error-proof the process.
Discovery and propagation of
knack. Discovery and elimination
of secret ignorance.
Solution follows the cause:
Transfer worker.
Supply training.
Meet the criteria for self-control.
Standardize the language;
provide translation, glossaries.
suggest that the errors are due to inadvertence. The randomness may apply to the types of error,
to the workers who make the errors, and to the time when the errors are made.
The cause of inadvertent errors is inattention. But what causes inattention? The search for an
answer leads into the complexities of psychological (e.g., monotony) and physiologic (e.g., fatigue)
phenomena. These are not fully understood, even by experts. To explore these complexities in-depth
means going deeper and deeper into an endless swamp. Practical managers prefer to go around the
swamp—to go directly from symptom to remedy.
Remedies for Inadvertent Errors. Remedies for inadvertent errors involve two main
1. Reduce the dependence on human attention through error-proofing: fail-safe designs, countdowns,
redundant verification, cutoffs, interlocks, alarm signals, automation, and robots. (Use of
bar codes has greatly reduced errors in identifying goods).
2. Make it easier for workers to remain attentive. Reorganize work to reduce fatigue and monotony
by use of job rotation, sense multipliers, templates, masks, overlays, and so on.
[For an uncommonly useful paper on error-proofing, with numerous examples, especially as applied
to service industries, see Chase and Stewart (1994).]
Technique Errors. Technique errors are made because workers lack some “knack”—some
essential technique, skill, or knowledge needed to prevent errors from happening. Technique errors
exhibit certain outward features. They are
Unintentional: The worker does not want to make errors.
Specific: Technique errors are unique to certain defect types—those types for which the missing
technique is essential.
Consistent: Workers who lack the essential technique consistently make more defects than workers
who possess the technique. This consistency is readily evident from data on worker errors.
Unavoidable: The inferior workers are unable to match the performance of the superior workers
because they (the inferior workers) do not know “what to do different.”
An example of technique errors is seen in the gun assembly case. Guns were assembled by 22
skilled artisans, each of whom assembled complete guns from bits and pieces. After the safety test,
about 10 percent of the guns could not be opened up to remove the spent cartridge—a defect known
as “open hard after fire.” For this defect it was necessary to disassemble the gun and then reassemble,
requiring about 2 hours per defective gun—a significant chronic waste.
Following much discussion, a table like Table 5.10 was prepared to show the performance of the
assemblers. This table shows the frequency of “open hard after fire” by assembler and by month over
a 6-month period. Analysis of the table brings out some significant findings.
The departmental defect rate varied widely from month to month, ranging from a low of 1.8 percent
in January to a high of 22.6 percent in February. Since all workers seemed to be affected,
this variation had its cause outside the department. (Subsequent analysis confirmed this.)
The ratio of the five best worker performances to the five worst showed a stunning consistency.
In each of the 6 months, the five worst performances add up to an error rate that is at least 10
times as great as the sum of the five best performances. There must be a reason for such a consistent
difference, and it can be found by studying the work methods—the techniques used by the
respective workers.
The knack: The study of work methods showed that the superior performers used a file to cut
down one of the dimensions on a complex component; the inferior performers did not file the file.
This filing constituted the knack—a small difference in method that accounts for a large differ-
ence in results. (Until the diagnosis was made, the superior assemblers did not realize that the filing
greatly reduced the incidence of “open hard after fire.”)
Usually the difference in worker performance is traceable to some superior knack used by the
successful performers to benefit the product. In the case of the gun assemblers, the knack consisted
of filing the appropriate component. In other cases, the difference in worker performance is due to
unwitting damage done to the product by the inferior performers—sort of “negative knack.”
There is a useful rule for predicting whether the difference in worker performance is due to a beneficial
knack or to a negative knack. If the superior performers are in the minority, the difference is
probably due to a beneficial knack. If the inferior performers are in the minority, then the difference
in performance is likely due to a negative knack.
In an aircraft assembly operation, data analysis by individual workers revealed that one worker
met the production quota consistently, whereas the others did not. The worker explained that he
had taken his powered screwdriver home and rebuilt the motor. The company replaced all the
motors, with a resulting increase in quality and productivity.
Analysis of data on damage to crankshafts showed that only one worker’s product was damaged.
Study in the shop then revealed that this worker sometimes bumped a crankshaft into a nearby
conveyor. Why? Because the worker was left-handed and the workplace layout was too inconvenient
for a left-handed person.
The gun assembly case shows the dangers of assuming that differences in worker performance
are due to a lack of motivation. Such an assumption is invalid as applied to technique errors.
Technique errors are doomed to go on and on until ways are found to provide the inferior workers
with an answer to the question, “What should I do different than I am doing now?”
How are such questions to be answered? Worker improvement teams sometimes can provide
answers. Failing this, they will keep on doing what they have been doing (and keep on making the
same defects) until the answers are provided by management.
TABLE 5.10 Matrix Analysis to Identify Technique Errors
operator rank Nov. Dec. Jan. Feb. Mar. Apr. Total
1 4 1 0 0 0 0 5
2 1 2 0 5 1 0 9
3 3 1 0 3 0 3 10
4 1 1 0 2 2 4 10
5 0 1 0 10 2 1 14
6 2 1 0 2 2 15 22
• • • • • • • •
• • • • • • • •
• • • • • • • •
17 18 8 3 37 9 23 98
18 16 17 0 22 36 11 102
19 27 13 4 62 4 14 124
20 6 5 2 61 22 29 125
21 39 10 2 45 20 14 130
22 26 17 4 75 31 35 188
TOTAL 234 146 34 496 239 241 1390
% Defective 10.6 6.6 1.8 22.6 10.9 11.0 10.5
5 best 9 6 0 20 5 8 48
5 worst 114 62 12 265 113 103 669
Ratio 13 10 ? 13 23 13 14
Remedies for Technique Errors. Solution of numerous cases of technique errors has yielded
a structured generic approach:
1. Collect data on individual worker performances.
2. Analyze the data for consistent worker-to-worker differences.
3. For cases of consistent differences, study the work methods used by the best and worst performers
to identify their differences in technique.
4. Study these differences further to discover the beneficial knack that produces superior results (or
the negative knack that damages the product).
5. Bring everyone up to the level of the best through appropriate remedial actions such as
 Train the inferior performers in use of the knack or in avoidance of damage.
 Change the technology so that the process embodies the knack.
 Error-proof the process in ways that force use of the knack or that prohibit use of the negative
 Institute controls and audits to hold the gains.
Conscious Errors. Conscious errors involve distinctive psychological elements. Conscious
errors are
Witting: At the time of making an error, the worker is aware of it.
Intentional: The error is the result of a deliberate decision on the part of the worker.
Persistent: The worker who makes the error usually intends to keep it up.
Conscious errors also exhibit some unique outward evidences. Whereas inadvertent errors exhibit
randomness, conscious errors exhibit consistency—some workers consistently make more errors
than others. However, whereas technique errors typically are restricted to one or a few defect types,
conscious errors tend to cover a wide spectrum of defect types.
On the face of it, workers who commit conscious errors deserve to be disciplined, but this principle
has only partial validity. Many such errors are actually initiated by management.
A major source of conscious errors is an atmosphere of blame. In such an atmosphere, workers
defend themselves by violating company rules. They omit making out the rework tickets, they hide
the scrap, and so on.
Another widespread source of conscious errors is conflict in priorities. For example, in a sellers’
market, priority on delivery schedules can prevail over some quality standards. The pressures on the
managers are transmitted down through the hierarchy and can result in conscious violation of quality
standards to meet the schedules.
In addition, some well-intentioned actions by management can have a negative effect. For example,
the managers launch a poster campaign to urge everyone to do better work. However, the campaign
makes no provision to solve some quality problems well known to the workers: poor quality from suppliers,
incapable processes, inadequate maintenance of facilities, and so on. Thus management loses
credibility—the workers conclude that the real message of the managers is “Do as we say, not as we do.”
Some conscious errors are initiated by the workers. Workers may have real or imagined grievances
against the boss or the company. They may get revenge by not meeting standards. Some
become rebels against the whole social system and use sabotage to show their resentment. Some of
the instances encountered are so obviously antisocial that no one—not the fellow employees, not the
union—will defend the actions.
To a degree, conscious worker errors can be dealt with through the disciplinary process. However,
managers also have access to a wide range of constructive remedies for conscious worker errors.
Remedies for Conscious Errors. Generally, the remedies listed here emphasize securing
changes in behavior but without necessarily changing attitudes. The remedies may be directed
toward the persons or the “system”—the managerial and technological processes.
Depersonalize the order: In one textile plant, the spinners were failing to tie the correct knots
(“weaver’s knots”) when joining two ends of yarn together. The pleas and threats of the supervisor
were of no avail. The spinners disliked the supervisor, and they resented the company’s poor
responses to their grievances. The problem was solved when the personnel manager took the
informal leader of the spinners to the Weaving Department to show her how the weavers were
having trouble due to wrong knots. Despite their unsolved grievances, once they learned about
the events in the weaving room, the spinners were unwilling to continue making trouble for their
fellow workers. The principle involved here is the law of the situation—one person should not
give orders to another person; both should take their orders from the situation. The law of the situation
is a phrase coined by Mary Parker Follett. [See Metcalf and Urwick (1941).] The situation
in the Weaving Department requires that weaver’s knots be tied. Hence this situation is binding
on the president, the managers, the supervisors, and the spinners.
Establish accountability: To illustrate, in one company the final product was packaged in bulky
bales that were transported by conventional forklift trucks. Periodically, a prong of a fork would
pierce a bale and do a lot of damage. Yet there was no way of knowing which trucker moved which
bale. When the company introduced a simple means of identifying which trucker moved which bale,
the amount of damage dropped dramatically.
Provide balanced emphasis: Workers discover the company’s real priorities on multiple standards
(quality, productivity, delivery) from the behavior of management. For example, scoreboards
on productivity and delivery rates should be supplemented with a scoreboard on quality
to provide evidence of balanced emphasis.
Conduct periodic quality audits: Systems of continuing traceability or scorekeeping are not
always cost-effective. Quality audits can be designed to provide, on a sampling basis, information
of an accountability and scorekeeping nature.
Provide assistance to workers: Visual aids to help prevent defects can be useful. Some companies
have used wall posters listing the four or five principal defects in the department, along with
a narrative and graphic description of the knack that can be used to avoid each defect.
Create competition, incentives: These devices have potential value if they are not misused.
Competition among workers and teams should be designed to be in good humor and on a friendly
level, such as prevails among departmental sports teams. Financial incentives are deceptively
attractive. They look good while pay is going up—during that part of the cycle there are “bonuses”
for good work. However, during a spate of poor work, removal of the bonuses converts the
incentives into penalties, with all the associated arguments about who is responsible.
Nonfinancial incentives avoid the pitfall of bonuses becoming penalties, but they should be kept
above the gimmickry level.
Error-proof the operation: Error-proofing has wide application to conscious errors. (See
Section 22, Operations, under Error-Proofing the Process.)
Reassign the work: An option usually available to managers is selective assignment, i.e., assign
the most demanding work to workers with the best quality record. Application of this remedy
may require redesign of jobs—separation of critical work from the rest so that selective assignment
becomes feasible.
Use the tools of motivation: This subject is discussed in Section 15, Human Resources and
This list of remedies helps to solve many conscious errors. However, prior study of the symptoms
and surrounding circumstances is essential to choosing the most effective remedy.
Communication Errors. A fourth important source of human error is traceable to errors in communication.
There are numerous subspecies of these, but a few of them are especially troublesome.
Communication omitted: Some omissions are by the managers. There are situations in
which managers take actions that on their face seem antagonistic to quality but without
informing the workers why. For example, three product batches fail to conform to quality feature
X. In each case, the inspector places a hold on the batch. In each case, a material review
board concludes that the batch is fit for use and releases it for delivery. However, neither the
production worker nor the inspector is told why. Not knowing the reason, these workers may
(with some logic) conclude that feature X is unimportant. This sets the stage for future unauthorized
actions. In this type of case and in many others, company procedures largely assume
that the workers have no need to know. (The release forms of material review boards contain
no blank to be filled in requiring the members to face the question: “What shall we communicate
to the work force?” Lacking such a provision, the question is rarely faced, so by default
there is no communication.
Communication inhibited: In most hierarchies, the prevailing atmosphere historically has inhibited
communication from the bottom up. The Taylor system of the late nineteenth century made
matters worse by separating planning from execution. More recently, managers have tried to use
this potential source of information through specific concepts such as suggestion systems,
employee improvement teams, and most recently, self-directed teams of workers. The twentieth
century rise in education levels has greatly increased the workers’ potential for participating usefully
in planning and improvement of operations. It is a huge underemployed asset. Managers are
well advised to take steps to make greater use of this asset.
Transmission errors: These errors are not conscious. They arise from limitations in human
communication. Identical words have multiple meanings, so the transmitter may have one meaning
in mind, but the receiver has a different meaning in mind. Dialects differ between companies
and even within companies. (The chief language in the upper levels is money, whereas in the
lower levels it is things.)
A critical category of terminology contains the words used to transmit broad concepts and matters
of a managerial nature: policies, objectives, plans, organization structure, orders (commands),
advice, reviews, incentives, and audits. The recipients (receivers) are mainly internal, across all functions
and all levels. The problem is to ensure that receivers interpret the words in ways intended by
transmitters. There is also the problem of ensuring that the responses are interpreted as intended.
In other cases, the intention of the transmitter is clear, but what reaches the receiver is something
else. A misplaced comma can radically change the meaning of a sentence. In oral communications,
background noise can confuse the receiver.
In an important football game, on a key play near the end of the game, amid the deafening noise
of the crowd, the defensive signal was three, which called for a man-to-man defense. One defensive
player thought he heard green, which called for a zone defense. The error resulted in loss of
the game (Anderson 1982).
Remedies for Communication Errors. Communication errors are sufficiently extensive
and serious to demand remedial action. The variety of error types has required corresponding variety
in the remedies.
Translation: For some errors, the remedy is to create ways to translate the transmitters’ communications
into the receivers’ language. A common example is the Order Editing Department,
which receives orders from clients. Some elements of these orders are in client language. Order
Editing translates these elements into the supplier’s language, through product code numbers,
acronyms, and other means. The translated version is then issued as an internal document within
the supplier’s company. A second example is the specialists in the Technical Service
Department. The specialists in this department are trained to be knowledgeable about their company’s
products. Through their contacts with customers, they learn of customer needs. This
combined knowledge enables them to assist both companies to communicate, including assistance
in translation.
The glossary: This useful remedy requires reaching agreement on definitions for the meanings
of key words and phrases. These definitions are then published in the form of a glossary—a list
of terms and their definitions. The publication may be embellished by other forms of communication:
sketches, photographs, and/or videotapes.
Standardization: As companies and industries mature, they adopt standardization for the mutual
benefit of customers and suppliers. This extends to language, products, processes, and so on.
In the case of physical goods, standardization is very widely used. Without it, a technological
society would be a perpetual tower of Babel. All organizations make use of short designations for
their products: code numbers, acronyms and so on. Such standardized nomenclature makes it easier
to communicate internally. If external customers also adopt the nomenclature, the problem of
multiple dialects is greatly reduced. The Airline Flight Guide publishes flight information for
multiple airlines. This information is well standardized. Some clients learn how to read the flight
guide. For such clients, communication with the airlines is greatly simplified.
Measurement: Saying it in numbers is an effective remedy for some communication problems,
e.g., those in which adjectives (such as roomy, warm, quick, and so on) are used to describe product
features). (For elaboration, see Section 9, Measurement, Information, and Decision-Making.)
A role for upper managers: Companies endure extensive costs and delays due to poor communication.
The remedies are known, but they do not emerge from day-to-day operations. Instead,
they are the result of specific projects set up to create them. In addition, they evolve slowly
because they share the common feature of “invest now for rewards later.” Upper managers are in
a position to speed up this evolution by creating project teams with missions to provide the needed
On the face of it, once a remedy has been determined, all that remains is to apply it. Not so. Instead,
obstacles are raised by various sources. There may be delaying tactics or rejection by a manager, the
work force, or the union. “Resistance to change” is the popular name for these obstacles.
Cultural Patterns. An understanding of resistance to change starts with the realization that
every change actually involves two changes:
1. The intended change
2. The social consequence of the intended change
The social consequence is the troublemaker. It consists of the impact of the intended change
on the cultural pattern of the human beings involved—on their pattern of beliefs, habits, traditions,
practices, status symbols, and so on. This social consequence is the root source of the resistance
to change. Dealing with this resistance requires an understanding of the nature of cultural
Ideally, advocates of change should be aware that all human societies evolve cultural patterns and
that these are fiercely defended as a part of “our way of life.” In addition, the advocates should try
to discover precisely what their proposals will threaten—which habits, whose status, what beliefs.
Unfortunately, too many advocates are not even aware of the existence of cultural patterns, let alone
their detailed makeup.
To make matters more complex, those who resist the change often state their reasons as objections
to the merits of the intended change, whereas their real reasons relate to the social consequences.
As a result, the advocates of the intended change are confused because the stated reasons
are not the real reasons for the resistance.
To illustrate, companies that first tried to introduce computer-aided design (CAD) ran into resistance
from the older designers, who claimed that the new technology was not as effective as
design analysis by a human being. Interviews then found that the real reasons included the fear
of losing status because the younger engineers could adapt more readily to the change.
Rules of the Road. Behavioral scientists have evolved some specific rules of the road for dealing
with cultural resistance (Mead 1951). These rules are widely applicable to industrial and other
organizational entities (Juran 1964).
Provide participation: This is the single most important rule for introducing change. Those who
will be affected by the change should participate in the planning as well as in the execution. Lack
of participation leads to resentment, which can harden into a rock of resistance.
Provide enough time: How long does it take for members of a culture to accept a change? They
need enough time to evaluate the impact of the change. Even if the change seems beneficial, they
need to learn what price they must pay in cultural values.
Start small: Conducting a small-scale tryout before going all out reduces the risks for the advocates
as well as for members of the culture.
Avoid surprises: A major benefit of the cultural pattern is its predictability. A surprise is a shock
to this predictability and a disturber of the peace.
Choose the right year: There are right and wrong years—even decades—for timing a change.
Keep the proposals free of excess baggage: Avoid cluttering the proposals with extraneous matters
not closely related to getting the results. The risk is that the debates will get off the main subject and
into side issues.
Work with the recognized leadership of the culture: The culture is best understood by its members.
They have their own leadership, and this is sometimes informal. Convincing the leadership
is a significant step in getting the change accepted.
Treat the people with dignity: The classic example was the relay assemblers in the Hawthorne
experiments. Their productivity kept rising, under good illumination or poor, because in the laboratory
they were being treated with dignity.
Reverse the positions: Ask the question: “What position would I take if I were a member of the
culture?” It is even useful to go into role playing to stimulate understanding of the other person’s
position. [For a structured approach, see Ackoff (1978).]
Deal directly with the resistance: There are many ways of dealing directly with resistance to
 Try a program of persuasion.
 Offer a quid pro quo—something for something.
 Change the proposals to meet specific objections.
 Change the social climate in ways that will make the change more acceptable.
 Forget it. There are cases in which the correct alternative is to drop the proposal. Human beings
do not know how to plan so as to be 100 percent successful.
[For added discussion, see Schein (1993). See also Stewart (1994) for a discussion of self-rating
one’s resistance to change.]
Resolving Differences. Sometimes resistance to change reaches an impasse. Coonley and
Agnew (1941) once described a structured process used for breaking an impasse on the effort to establish
quality standards on cast iron pipe. Three conditions were imposed on the contesting parties:
1. They must identify their areas of agreement and their areas of disagreement. “That is, they must
first agree on the exact point at which the road began to fork.” When this was done, it was found
that a major point of disagreement concerned the validity of a certain formula.
2. “They must agree on why they disagreed.” They concluded that the known facts were inadequate
to decide whether the formula was valid or not.
3. “They must decide what they were going to do about it.” The decision was to raise a fund to conduct
the research needed to establish the necessary facts. “With the facts at hand, the controversies
The universal sequence for improvement sets up a common pattern for the life cycle of projects.
Following project selection, the project is defined in a mission statement and is assigned to a project
The team then meets, usually once a week for an hour or so. During each meeting, the team
Reviews the progress made since the previous meeting
Agrees on the actions to be taken prior to the next meeting (the homework)
Assigns responsibility for those actions
Gradually, the team works its way through the universal sequence. The diagnostic journey establishes
the causes. The remedial journey provides the remedies and establishes the controls to hold
the gains.
During all this time, the team issues minutes of its meetings as well as periodic progress reports.
These reports are distributed to team members and also to nonmembers who have a need to know.
Such reports form the basis for progress review by upper managers.
The final report contains a summary of the results achieved, along with a narrative of the activities
that led to the results. With experience, the teams learn to identify lessons learned that can be
applied elsewhere in the company. [Relative to the life cycle of a project, see AT&T Quality Library,
Quality Improvement Cycle (1988, pp. 13–17).]
Numerous companies have initiated quality improvement, but few have succeeded in institutionalizing
it so that it goes on year after year. Yet many of these companies have a long history of annually conducting
product development, cost reduction, productivity improvement, and so on. The methods they
used to achieve such annual improvement are well known and can be applied to quality improvement.
Enlarge the annual business plan to include goals for quality improvement.
Make quality improvement a part of the job description. In most companies, the activity of
quality improvement has been regarded as incidental to the regular job of meeting the goals
for quality, cost, delivery, and so on. The need is to make quality improvement a part of the
regular job.
Establish upper management audits that include review of progress on quality improvement.
Revise the merit rating and reward system to include a new parameter—performance on quality
improvement—and give it proper weight.
Create well-publicized occasions to provide recognition for performance on improvement.
The upper managers must participate extensively in the quality initiative. It is not enough to create
awareness, establish goals, and then leave all else to subordinates. This has been tried and has failed
over and over again. I know of no company that became a quality leader without extensive participation
by upper managers.
It is also essential to define just what is meant by “participation.” It consists of a list of roles to be
played by the upper managers, personally. What follows is a list of roles actually played by upper managers
in companies that have become quality leaders. These roles can be regarded as “nondelegable.”
Serve on the quality council: This is fundamental to upper managers’ participation. It also
becomes an indicator of priorities to the rest of the organization.
Acquire training in managing for quality: Sources of such training include visits to successful
companies. Training is also available at courses specially designed for upper managers and
through attending conferences. (Upper managers risk losing credibility if they try to lead while
lacking training in managing for quality.)
Approve the quality vision and policies: A growing number of companies have been defining
their quality vision and policies. Invariably, these require upper management approval before they
may be published.
Approve the major quality goals: The quality goals that enter the business plan must be
deployed to lower levels to identify the deeds to be done and the resources needed. The upper
managers become essential parties to the deployment process.
Establish the infrastructure: The infrastructure includes the means for nominating and selecting
projects, preparing mission statements, appointing team leaders and members, training teams
and facilitators, reporting progress, and so on. Lacking such an infrastructure, quality improvement
will take place only in local areas and with no noticeable effect on the bottom line.
Provide resources: During the 1980s, many upper managers provided extensive resources for
training their personnel, chiefly in awareness and in statistical tools. In contrast, only modest
resources were provided for training in managing for quality and for setting up the infrastructure
for quality improvement.
Review progress: A major shortcoming in personal participation by upper managers has been
the failure to maintain a regular review of progress in making quality improvements. During the
1980s, this failure helped to ensure lack of progress—quality improvement could not compete
with the traditional activities that did receive progress reviews from upper managers.
Give recognition: Recognition usually involves ceremonial events that offer highly visible
opportunities for upper managers to show their support for quality improvement. Upper managers
should seize these opportunities; most upper managers do so. (See below, under Recognition.)
Revise the reward system: Traditional reward systems provide rewards for meeting traditional
goals. These systems must now be opened up to give proper weight to performance on quality
improvement. Upper managers become involved because any changes in the reward system
require their approval. (See below, under Rewards.)
Serve on project teams: There are some persuasive reasons behind this role. See preceding,
under The Project Team, Upper Managers on Teams.
Face up to employee apprehensions: See preceding, under The Quality Council, Apprehensions
about Elimination of Jobs.
Such is a list of the nondelegable roles of upper managers. In companies that have become quality
leaders, the upper managers carry out most, if not all, of these roles. No company known to me
has attained quality leadership without the upper managers carrying out those nondelegable roles.
Scheduled, periodic review of progress by upper managers is an essential part of maintaining annual
quality improvement. Activities that do not receive such review cannot compete for priority with
activities that do receive such review. Subordinates understandably give top priority to matters that
are reviewed regularly by their superiors.
Review of Results. Results take multiple forms, and these are reflected in the design of the
review process. Certain projects are of such importance individually that the upper managers want
to follow them closely. The remaining projects receive their reviews at lower levels. However, for the
purpose of upper management review, they are summarized to be reviewed collectively by upper
There is also a need for regular review of the quality improvement process. This is done through
audits that may extend to all aspects of managing for quality. (Refer to Section 11, ISO 9000 Family
of Standards.)
Inputs to Progress Review. Much of the database for progress review comes from the
reports issued by the project teams. However, it takes added work to analyze these reports and to prepare
the summaries needed by upper managers. Usually this added work is done by the secretary of
the quality council with the aid of the facilitators, the team leaders, and other sources such as finance.
As companies gain experience, they design standardized reporting formats to make it easy to
summarize reports by groups of projects, by product lines, by business units, by divisions, and for
the corporation. One such format, used by a large European company, determines for each project
The original estimated amount of chronic waste
The original estimated reduction in cost if the project were to be successful
The actual cost reduction achieved
The capital investment
The net cost reduction
The summaries are reviewed at various levels. The corporate summary is reviewed quarterly at
the chairman’s staff meeting (personal communication to the author).
Evaluation of Performance. One of the objectives of progress review is evaluation of performance.
This evaluation extends to individuals as well as to projects. Evaluation of individual
performance on improvement projects runs into the complication that the results are achieved by
teams. The problem then becomes one of evaluating individual contribution to team efforts. This
new problem has as yet no scientific solution. Thus each supervisor is left to judge subordinates’
contributions based on inputs from all available sources.
At higher levels of organization, the evaluations extend to judging the performance of supervisors
and managers. Such evaluations necessarily must consider results achieved on multiple projects.
This has led to evolution of measurement (metrics) to evaluate managers’ performance on projects
collectively. These metrics include
Numbers of improvement projects: initiated, in progress, completed, and aborted
Value of completed projects in terms of improvement in product performance, reduction in costs,
and return on investment
Percentage of subordinates active on project teams
Superiors then judge their subordinates based on these and other inputs.
“Recognition” as used here means “public acknowledgment of superior performance.” (Superior performance
deserves public acknowledgment.) Recognition tells recipients that their efforts are appreciated.
It adds to their self-respect and to the respect received from others.
Most companies are quite effective at providing recognition. They enlist the ingenuity of those
with special skills in communication—Human Relations, Marketing, Advertising—as well as the
line managers. The numerous forms of recognition reflect this ingenuity:
Certificates, plaques, and such are awarded for serving on project teams, serving as facilitator,
and completing training courses.
Project teams present their final report in the office of the ranking local manager.
Project summaries are published in the company news media, along with team pictures. Some
companies create news supplements or special newsletters devoted to quality improvement.
Published accounts of successful projects not only provide recognition, they also serve as case
materials for training purposes and as powerful stimulators to all.
Dinners are held to honor project teams.
Medals or prizes may be awarded to teams judged to have completed the best projects during
some designated time period. The measure of success always includes the extent of results
achieved and sometimes includes the methods used to achieve the results. [For an account of the
annual competition sponsored by Motorola, see Feder (1993); see also Motorola’s Team
Competition (1992).]
As used here, “rewards” refers to salaries, salary increases, bonuses, promotions, and so on resulting
from the annual review of employee performance. This review has in the past focused on meeting
goals for traditional parameters: costs, productivity, schedule, and quality. Now a new parameter—
quality improvement—must be added to recognize that quality improvement is to become a part of
the job description.
Note that reward differs sharply from recognition. The crucial difference lies in whether the work
is voluntary or mandatory.
Recognition is given for superior performance, which is voluntary. (People can hold their jobs by
giving adequate performance.)
Rewards are given for mandated performance—doing the work defined in the job description.
Willful failure to do this work is a violation of the employment contract and is a form of
The new parameter—quality improvement—is time-consuming. It adds a new function. It
invades the cultural pattern. Yet it is critical to the company’s ability to remain competitive. This is
why the parameter of quality improvement must enter the job descriptions and the reward system.
Failing this, employees will continue to be judged on their performance against traditional goals, and
quality improvement will suffer due to lack of priority.
One well-known company had added the parameter “performance on quality improvement” to its
annual review system. All personnel in the managerial hierarchy are then rated into one of three
classes: more than satisfactory, satisfactory, or less than satisfactory. Those who fall into the lowest
class are barred from advancement for the following 12 months (personal communication to
the author).
(For additional discussion, see Section 15, Human Resources and Quality.)
Throughout this section there have been numerous observations on the needs for training. These
needs are extensive because quality improvement is a new function in the company that assigns new
responsibility to all. To carry out these new responsibilities requires extensive training. Some details
of this training have been discussed here and there in this section. (For additional discussion, see
Section 16, Training for Quality.)
Quality improvement requires action at all levels of the organization, as summarized in Figure 5.21.
(For more on the planning and coordination of these activities on multiple levels, see Section 13,
Strategic Deployment, and Section 14, Total Quality Management.)
This section of the handbook has drawn extensively from various training materials published by the
Juran Institute, Inc., and Juran, J. M., ed. (1955), A History of Managing for Quality, sponsored by
the Juran Foundation, Inc. Quality Press, Milwaukee. I am also grateful to the Juran Foundation, Inc.,
and the Juran Institute, Inc., for permission to extract from those works.
Ackoff, Russell L. (1978). The Art of Problem Solving. John Wiley & Sons, New York, pp. 45–47.
AT&T Quality Library. For Information, call AT&T Customer Information Center, 1-800-432-6600.
Anderson, Dave (1982). “When 3 Meant 6.” New York Times, December 27, pp. C1, C3.
Berwick, Donald M., Godfrey, A. Blanton, and Roessner, Jane (1990). Curing Health Care. Jossey-Bass, San
Betker, Harry A. (1983). “Quality Improvement Program: Reducing Solder Defects on Printed Circuit Board
Assembly.” The Juran Report, No. 2, November, pp. 53–58.
Betker, Harry A. (1985). “Storyboarding: It’s No Mickey Mouse Technique.” The Juran Report, No. 5, Summer,
pp. 25–30.
Bigelow, James S., and Floyd, Raymond C. (1990). “Quality Improvement through Root Cause Analysis.”
IMPRO 1990 Conference Proceedings, Juran Institute, Inc., Wilton, CT.
Black, Dennis A., and Stump, Jerry R. (1987). “Developing AQI Team ‘Attitudes for Progress.’ ” IMPRO 1987
Conference Proceedings, Juran Institute, Inc., Wilton, CT.
Boley, Daniel C., and Petska, Joseph P. (1990). “The Genesis of a Partnership: A Tale of Two Plants.” IMPRO
1990 Conference Proceedings, Juran Institute, Inc., Wilton, CT.
Chase, Richard B., and Stewart, Douglas M. (1994). “Make Your Service Fail-Safe.” Sloan Management Review,
Spring, pp. 35–44.
Upper management Quality improvement teams Operating departments
Organize quality councils Receive and review mission Implement remedies
Secure and screen project Conduct diagnostic journey Implement controls
Select projects Conduct remedial journey
Prepare mission Deal with cultural resistance
Assign teams; establish Establish controls to hold
training gains
Review progress Report results
FIGURE 5.21 Responsibility for quality improvement.
Cohen, Leonard A. (1987). “Diet and Cancer.” Scientific American, November, pp. 42–48.
Coonley, Howard, and Agnew, P. G. (1941). The Role of Standards in the System of Free Enterprise. American
National Standards Institute, New York.
Daughton, William J. (1987). “Achieving Customer Satisfaction with a Major Automobile Manufacturer.”
IMPRO 1987 Conference Proceedings, Juran Institute, Inc., Wilton, CT.
DeWolff, Thomas P., Anderson, Ralph J., and Stout, Gordon L. (1987). “Quality Improvement Process and
Cultural Change.” IMPRO 1987 Conference Proceedings, Juran Institute, Inc., Wilton, CT.
Egan, John (1985). “The Jaguar Obsession.” Quality (EOQC), January, pp. 3–4.
Engle, David, and Ball, David (1985). “Improving Customer Service for Special Orders.” IMPRO 1985
Conference Proceedings, Juran Institute, Inc., Wilton, CT.
Feder, Barnaby J. (1993). “At Motorola, Quality Is a Team Sport.” New York Times, January 21.
Florida Power and Light Company (1984). American Productivity Center, Case Study 39, September.
Greenridge, R. M. C. (1953). “The Case of Reliability vs. Defective Components et al.” Electronic Applications
Reliability Review, No. 1, p. 12.
Gross, Neil (1989). “A Wave of Ideas, Drop by Drop.” Business Week, Innovation, pp. 22, 28, and 30.
Gulliver, Frank R. (1987). “Post-Project Appraisals Pay.” Harvard Business Review, March-April, pp. 128–132.
Hartman, Bob (1983). “Implementing Quality Improvement.” The Juran Report, No. 2, November, pp. 124–131.
IMPRO Proceedings, Annual Conference on Quality Management, Juran Institute, Inc., Wilton, CT.
Juran, J. M. (1964). Managerial Breakthrough. McGraw-Hill, New York. Revised edition, 1995.
Juran, J. M. (1975). “The Non-Pareto Principle; Mea Culpa.” Quality Progress, May, pp. 8–9.
Juran, J. M. (1978). “Japanese and Western Quality—A Contrast.” From a paper originally presented at the
International Conference on Quality Control, Tokyo, Oct. 17–20.
Juran, J. M. (1981). Juran on Quality Improvement, a series of 16 videocassettes on the subject. Juran Institute
Inc., Wilton, CT.
Juran, J. M. (1985). “A Prescription for the West—Four Years Later.” European Organization for Quality, 29th
Annual Conference. Reprinted in The Juran Report, No. 5, Summer 1985.
Juran, J. M. (1992). Juran on Quality by Design. Free Press, New York, pp. 407–425.
Juran, J. M. (1993), “Made in USA, a Renaissance in Quality.” Harvard Business Review, July-August, pp. 42–50.
Juran, J. M., ed. (1995). A History of Managing for Quality. Sponsored by Juran Foundation, Inc., Quality Press,
ASQC, Milwaukee.
Juran Report, The. Issued periodically by Juran Institute, Inc., Wilton, CT.
Kearns, David T., and Nadler, David A. (1995). How Xerox Reinvented Itself and Beat Back the Japanese. Harper
Business, New York.
Kelly, Bob, Howard, Tom, and Gambel, Anthony (1990). “Quality Forces Alcoa/Phirer Partnership.” IMPRO
1990 Conference Proceedings, Juran Institute, Inc., Wilton, CT.
Kegarise, Ronald J., and Miller, George D. (1985). “An Alcoa-Kodak Joint Team.” IMPRO 1985 Conference
Proceedings, Juran Institute, Inc., Wilton, CT.
Kinosz, Donald L., and Ice, James W. (1991). “Facilitation of Problem Solving Teams.” IMPRO 1991 Conference
Proceedings, Juran institute, Inc., Wilton, CT.
Krantz, K. Theodor (1989). “How Velcro Got Hooked on Quality.” Harvard Business Review, September-
October, pp. 34–40.
Mead, Margaret, ed. (1951). Cultural Patterns and Technical Change. UNESCO, Paris; also published by Mentor
Books, New American Library of World Literature, New York, 1955.
Metcalf, Henry C., and Urwick, L., eds. (1941). Dynamic Administration. Harper and Row, New York.
Mizell, Michael, and Strattner, Lawrence (1981). “Diagnostic Measuring in Manufacturing.” Quality, September,
pp. 29–32.
Motorola’s Team Competition (1992). Productivity Views, January/February, pp. 3–4.
NIST, National Institute of Standards and Technology, Gaithersburg, MD 20899. (It manages the U.S. Malcolm
Baldrige National Quality Award.)
Parvey, Dale E. (1990). “The Juran Improvement Methodology Applied to Information Security Systems.”
IMPRO 1990 Conference Proceedings, Juran Institute, Inc., Wilton, CT.
Pelletier, Don (1990). “Quality Improvement and the Executive Employee.” IMPRO 1990 Conference
Proceedings, Juran Institute, Inc., Wilton, CT.
Rassi, Alan J. (1991). “A Three-Phase Approach to Exceptional Customer Service.” IMPRO 1991 Conference
Proceedings, Juran Institute, Inc., Wilton, CT.
Sakai, Shinji (1994). “Rediscovering Quality—The Toyota Way.” IMPRO 1994 Conference Proceedings, Juran
Institute, Inc., Wilton, CT.
Schein, Edgar H. (1993). “How Can Organizations Learn Faster? The Challenge of Entering the Green Room.”
Sloan Management Review, Winter, pp. 85–92.
Seder, L. A., and Cowan, D. (1956). “The Span Plan of Process Capability Analysis. American Society for
Quality Control,” General Publication 3, September. (It was originally called “S.P.A.N. capability method,” in
which the letters stood for “systematic procedure for attaining necessary capability.”)
Seder, L. A.(1950). “Diagnosis with Diagrams.” Industrial Quality Control, Vol. 6, Nos. 4 and 5, January and
Smidt, Harm, and Doesema, Tjeerd (1991). “Lead Time Reduction of Project Related Purchased Components.”
IMPRO 1991 Conference Proceedings, Juran Institute, Inc., Wilton, CT.
Sterett, Kent W. (1987). “Quality at Florida Power and Light. (The Role of Facilitators and Performance
Reviews).” IMPRO 1987 Conference Proceedings, Juran Institute, Inc., Wilton, CT.
Stewart, Thomas A. (1994). “Rate Your Readiness to Change.” Fortune, February 7, pp. 106–110.
Stratton, Donald A. (1987). “Force Field Analysis: A Powerful Tool for Facilitators.” The Juran Report, No. 8,
pp. 105–111.
Traver, Robert W. (1983). “Locating the Key Variables.” ASQC Quality Congress Transactions, Milwaukee, pp.
Whipple, A. B. C. (1984). “Stranded Navy Man Who Charted the World’s Seas.” Smithsonian, March, pp.
Some Accounts of Mobilizing for Quality Improvement
Hodgson, Alan (1991). “Starting the Amersham Quality Revolution” (at Amersham International). IMPRO 1991
Conference Proceedings, Juran Institute, Inc., Wilton, CT.
Lewicki, Thomas A., and Romo, Robert A. (1987). “Quality Through Collaboration” (at General Motors Central
Foundry). IMPRO 1987 Conference Proceedings, Juran Institute, Inc., Wilton, CT.
McBride, P. S. (1993). “Short’s Total Quality Management Programme: Quality Management in Practice” (at
Short Brothers, Belfast, Northern Ireland). Quality Forum, June 1993.
Payne, Bernard J. (1987). “Overall QI/JQI: From 1% AQL to 10 ppm” (at Mullard Blackburn). IMPRO 1987
Conference Proceedings, Juran Institute, Inc., Wilton, CT.
Scanlon, Robert J. (1991). “Installing a Quality Process in Record Time” (at Southern Pacific Transportation
Company). IMPRO 1991 Conference Proceedings, Juran Institute, Inc., Wilton, CT.
Tenner, Arthur R., and De Toro, Irving J. (1992). Total Quality Management: Three Steps to Continuous
Improvement (at Banc One, based on the work of Charles A. Aubrey II). Addison-Wesley Publishing Company,
Reading, MA.
Young, John (1990). “Enabling Quality Improvement in British Telecom.” IMPRO 1991 Conference
Proceedings, Juran Institute, Inc., Wilton, CT.
James F. Riley, Jr.
Why Process Management? 6.1
The Origins of PQM 6.2
Process Quality Management (PQM)
Defined 6.3
Overview 6.6
Initiating PQM Activity 6.6
The Planning Phase: Planning the New
Process 6.8
The Transfer Phase: Transferring the New
Process Plan to Operations 6.16
Operational Management Phase:
Managing the New Process 6.18
Key Points 6.19
Critical Success Factors for PQM
implementation 6.20
Why Process Quality Management? The dynamic environment in which business is conducted
today is characterized by what has been referred to as “the six c’s:” change, complexity, customer
demands, competitive pressure, cost impacts, and constraints. All have a great impact on an
organization’s ability to meet its stated business goals and objectives. Traditionally, organizations
have responded to these factors with new products and services. Rarely have they made changes in
the processes that support the new goods and services.
Experience shows that success in achieving business goals and objectives depends heavily on
large, complex, cross-functional business processes, such as product planning, product development,
invoicing, patient care, purchasing, materials procurement, parts distribution, and the like. In the
absence of management attention over time, many of these processes become obsolete, overextended,
redundant, excessively costly, ill-defined, and not adaptable to the demands of a constantly changing
environment. For processes that have suffered this neglect (and this includes a very large number
of processes for reasons that will be discussed later in this section) quality of output falls far short
of the quality required for competitive performance.
A business process is the logical organization of people, materials, energy, equipment, and information
into work activities designed to produce a required end result (product or service) (Pall 1986).
There are three principal dimensions for measuring process quality: effectiveness, efficiency, and
adaptability. The process is effective if the output meets customer needs. It is efficient when it is
effective at the least cost. The process is adaptable when it remains effective and efficient in the face
of the many changes that occur over time. A process orientation is vital if management is to meet
customer needs and ensure organizational health.
On the face of it, the need to maintain high quality of processes would seem obvious. To understand
why good process quality is the exception, not the rule, requires a close look at how processes
are designed and what happens to them over time.
First, the design. The western business organization model, for reasons of history, has evolved
into a hierarchy of functionally specialized departments. Management direction, goals, and measurements
are deployed from the top downward through this vertical hierarchy. However, the
processes which yield the products of work, in particular those products which customers buy (and
which justify the existence of the organization), flow horizontally across the organization through
functional departments (Figure 6.1). Traditionally, each functional piece of a process is the responsibility
of a department, whose manager is held accountable for the performance of that piece.
However, no one is accountable for the entire process. Many problems arise from the conflict
between the demands of the departments and the demands of the overall major processes.
In a competition with functional goals, functional resources, and functional careers, the crossfunctional
processes are starved for attention. As a result, the processes as operated are often neither
effective nor efficient, and they are certainly not adaptable.
A second source of poor process performance is the natural deterioration to which all processes
are subject in the course of their evolution. For example, at one railroad company, the company telephone
directory revealed that there were more employees with the title “rework clerk” than with the
title “clerk.” Each of the rework clerks had been put in place to guard against the recurrence of some
serious problem that arose. Over time, the imbalance in titles was the outward evidence of processes
which had established rework as the organization’s norm.
The rapidity of technological evolution, in combination with rising customer expectations, has
created global competitive pressures on costs and quality. These pressures have stimulated an exploration
of cross-functional processes—to identify and understand them and to improve their performance.
There is now much evidence that within the total product cycle a major problem of poor
process performance lies with process management technologies. Functional objectives frequently
conflict with customer needs, served as they must be by cross-functional processes. Further, the
processes generate a variety of waste (missed deadlines, factory scrap, etc.). It is not difficult to identify
products, such as invoice generation, preparation of an insurance policy, or paying a claim, that
take over 20 days to accomplish less than 20 min of actual work. They are also not easily changed
in response to the continuously changing environment. To better serve customer needs there is a need
to restore these processes to effectiveness, efficiency, and adaptability.
The Origins of PQM. IBM Corporation was among the first American companies to see the
benefits of identifying and managing business processes. The spirit of IBM’s first efforts in manag-
FIGURE 6.1 Workflow in a functional organization. (Source: Juran Institute, Wilton, CT.)
ing business processes in the early 1980s was expressed in the words of one executive: “Focus for
improvement must be on the job process” (Kane 1986). Process Management has long been practiced
in manufacturing. In product manufacturing, the plant manager “owns” a large part of the manufacturing
process. This manager has complete responsibility for operating this part of the
manufacturing process and is accountable for the results. As owner, the manager is expected to control,
improve, and optimize the manufacturing process to meet customer needs and business needs
(cost, cycle time, waste elimination, value creation, etc.). In pursuit of these targets, managers of the
manufacturing process have developed some indispensable concepts and tools, including definition
of process requirements, step-by-step process documentation, establishment of process measurements,
removal of process defects, and assurance of process optimization. In fact, much of the science
of industrial engineering is concerned with these tasks. Recognizing the value of these tools in
manufacturing and their applicability to business processes, the IBM senior management committee
directed that process management methodology be applied to all major business processes (such as
product development, business planning, distribution, billing, market planning, etc.), and not just to
the manufacturing process.
Around the same time, a number of other North American companies, including AT&T, Ford
Motor Company, Motorola, Corning, and Hewlett-Packard, also began applying process management
concepts to their business processes. In all of these companies, the emphasis was placed on
cross-functional and cross-organizational processes. Application of process management methodology
resulted in breaking down the functional barriers within the processes. In each case, a new, permanent
managerial structure was established for the targeted process.
By mid-1985, many organizations and industries were managing selected major business
processes with the same attention commonly devoted to functions, departments, and other organizational
entities. Early efforts bore such names as Business Process Management, Continuous Process
Improvement, and Business Process Quality Improvement.
Business Process Reengineering (BPR) should be mentioned as part of this family of methodologies.
Like the methodologies mentioned previously in this section, BPR accomplishes a shift of managerial
orientation from function to process. According to the consultants who first described BPR and
gave it its name, BPR departs from the other methodologies in its emphasis on radical change of
processes rather than on incremental change. Furthermore, BPR frequently seeks to change more than
one process at the same time. Because of the economic climate of the early 1990s, and the outstanding
payback that some writers attribute to BPR, its popularity grew rapidly for a time.
However, there is evidence, including the testimony of Michael Hammer, one of the most widely
read writers on BPR, that in many early applications, the lure of rapid improvement caused some
managers (and their consultants), who ignored human limitations, to impose too much change in too
short a time, with a devastating effect on long-term organization performance. Furthermore, in many
early applications, users became so fascinated by the promise of radical change that they changed
everything, overlooking elements of the existing process design that worked perfectly well and
would have been better carried over as part of the new design. Such a carryover would have saved
time, reduced demand on the designers, and produced a better result.
Much has been published on process management. AT&T (1988), Black (1985), Gibson
(1991–92), Hammer and Champy (1993), Kane (1986 and 1992), Pall (1987), Riley (1989),
Rummler (1992), Schlesiona (1988), and Zachman (1990) have all proposed similar methodological
approaches that differ from one another in minor details. The specific details of the methodology presented
in this section were developed by consultants at the Juran Institute, Inc. [Gibson et al. (1990);
Riley et al. (1994)], based on years of collective experience in a variety of industries.
Process Quality Management (PQM) Defined. The methodology described in this section
is one which has been introduced with increasing success by a number of prominent corporations,
including the ones already mentioned. While it may vary in name and details from company to company,
the methodology possesses a core of common features which distinguishes it from other
approaches to managing quality. That core of features includes: a conscious orientation toward customers
and their needs; a specific focus on managing a few key cross-functional processes which most
affect satisfaction of customer needs; a pattern of clear ownership—accountability for each key
process; a cross-functional team responsible for operating the process; application at the process level
of quality-management processes—quality control, quality improvement, and quality planning. In
this section, the methodology will be referred to as process quality management, or PQM.
Before discussing the details of PQM, an example will illustrate how a process, operating in a traditional
functional hierarchy, may respond poorly to a seemingly minor change in an operating environment,
and how the effects of that change can stimulate dramatic improvements, made possible by
applying the process management approach. It also illustrates how the potential for dramatic
improvement offered by a new technology (information technology, in this case) is more easily recognized
when there is accountability for making those improvements.
In the early 1980s, a major multinational manufacturer of information processing systems decided
to change its traditional rent-or-lease product pricing policy to include outright purchase. This strategic
change led to a complete revision of the company’s contracting policies, including terms and conditions.
Instead of firm list prices, published discounts were now available; for especially large procurements,
the company offered unpublished discounts with a number of financing options. A new contract management
process evolved out of the new policy, an incremental modification of the existing process. The
new process had to accommodate special contracts with a variety of nonstandard terms and conditions.
Within 2 years, more than 10 percent of the company’s revenue was generated by “special contracts.”
However, as the percentage of this revenue increased, the ratio of sales closed to proposals
made plummeted to fewer than 1 out of 5—a process yield of 20 percent. Both customers and field
marketing representatives complained about the long turnaround time (the time elapsed from receipt
of a request for a proposal until delivery of the proposal to the customer), which averaged 14 weeks.
The process was simply unresponsive to customer business needs.
Facing lost business opportunities and a barrage of complaints from field marketing representatives,
the executive quality council targeted this process for the application of process quality management.
The director of contract management was designated as the process owner, and formed a
process management team comprising representatives from field marketing, field administration,
business systems, product development, finance, marketing practices, and legal services.
Originally, the contract management process was a serial set of steps (Figure 6.2). The process
started with the field marketing representative, who received a request for a special contract proposal
from the customer. A draft of the contract proposal was then prepared in the branch office
with the help of branch administration and reviewed by the branch manager. Subsequently, it was
submitted for regional management review (usually in another geographic location) and finally for
a comprehensive evaluation at headquarters by large-account marketing, marketing practices,
finance, and business systems. If the proposal was in order, a contract was prepared. The contract
management department then arranged for up to 28 sequential approvals of the contract at the executive
level, involving various functions, such as product development, finance, legal services, and
the like.
Having successfully passed all these hurdles, the contract was then approved and returned to marketing
division headquarters for further refinement and processing. Eventually, some 3 to 4 months
later, the proposed contract returned to the branch office for presentation to the customer. In many
instances, it arrived too late. The customer had taken the business to a competitor.
The process management team flow-charted the process, and validated a number of hypotheses.
These included: manual processing was slow; there were postal service delays; the serial approval
process, consisting of up to 28 high-level executive signoffs, took too long; the memo-generated
guidelines for the format and content of the contract proposal were vague, conflicting, and difficult
to access; and there was substantial resistance to changing to the new, purchase-only strategy, especially
by the finance, business practices, and legal functions.
After several months of process redesign and test, the team launched the new contract management
process, shown in Figure 6.3. The new process addressed all causes of delay that the team had
discovered. It made especially good use of new information technology support, which was unavailable
years before when the contract management process began operation.
In designing the new process, the team incorporated a number of important new features:
 The team wrote new guidelines for contract proposal preparation and installed them on-line, using
a national electronic mail system. They established a system to keep the guidelines continuously
updated. This measure accelerated both preparation and transmission of contract proposals.
 They arranged for approval authority for simple contracts—those of lower dollar volume or having
no special engineering requirements—to be delegated to the regional marketing manager.
 They established two review boards for concurrent review and approval of special contracts. The
concurrent processes replaced the serial process.
 They found that the teamwork required by the cross-functional arrangement had the added effect
of reducing interfunctional rivalry and resistance to the new marketing policy.
 They established turnaround time requirements for each activity in the process, then measured and
tracked actual performance against the standards. Whenever they experienced delays beyond the
specified time targets, they initiated corrective action. For example, each review board had 5 business
days to review proposals, approve them, and pass them on. With the target established, it was
a relatively simple matter for the board to monitor its own performance against the standard.
This new management approach resulted in an 83 percent improvement in average turnaround time
(from 14 weeks to 17 days), and an increase in process yield of 180 percent (from 20 to 56 percent). Still,
the team was not satisfied. They implemented two more process redesigns in the next 3 years. After 5
years of PQM focus, the special-contract management process was performing at a 60 percent yield. For
simple contracts, which account for 92 percent of the process volume, the turnaround time is 24 hours.
Before it was redesigned, this process consumed the equivalent of 117 full-time people; as of
1995, after the several redesigns, it required fewer than 60. Special-contract revenue now exceeds 30
FIGURE 6.2 Special-contract management process (before application of process-management principles). (Source: Juran Institute,
Wilton, CT.)
Branch Office Regional Office HQ Large-Account HQ Special-Contract Assorted High-Level
Marketing Management Managers
Prepare special- Review, approve, Log in, review, and Review, provide price, terms, 28 sequential approvals
contract proposal? forward? assign for approval? and conditions; schedule
Plan presentation Consolidate changes and
approvals; finalize special
? ?
Branch Office Regional HQ Large-Account HQ Special-Contract Assorted High-Level Review Review
Office Marketing Management Managers Board #1 Board #2
Prepare special- Review, Screen, log in, and [Sequential Evaluate and Evaluate
contract approve? follow review board approvals approve? and
proposal? actions? eliminated]? approve
Plan Finalize special
presentation contract
FIGURE 6.3 Special-contract management process (after application of process-management principles). (Source: Juran Institute,
Wilton, CT.)
percent of total U.S. revenue—an all-time high. Customers and company management agree that the
present process performance may be judged effective and efficient. As additional redesigns are
required to respond to the inevitable changes in environment, the managers believe that the process
will also prove to be adaptable.
Overview. A PQM effort is initiated when executive management selects key processes, identifies
owners and teams, and provides them with process mission statements and goals. After the owners
and team are trained in process methodology, they work through the three phases of PQM methodology:
planning, transfer, and operational management.
The planning phase, in which the process design (or redesign) takes place, involves five steps:
1. Defining the present process.
2. Determining customer needs and process flow.
3. Establishing process measurements.
4. Conducting analyses of measurement and other data.
5. Designing the new process. The output is the new process plan.
Planning is the most time-consuming of the three phases.
The transfer phase is the second phase, in which the plans developed in the first phase are handed
off from the process team to the operating forces and put into operation.
Operational management is the third phase of PQM. Here, the working owner and team first
monitor new process performance, focusing on process effectiveness and efficiency measurements.
They apply quality control techniques, as appropriate, to maintain process performance. They use
quality improvement techniques to rid the process of chronic deficiencies. Finally, they conduct a
periodic executive management review and assessment to ensure that the process continues to meet
customer needs and business needs, and remains competitive.
Replanning, the cycling back to the first phase, is invoked when indicated. PQM is not a one-time
event; it is itself a continuous process.
Initiating PQM Activity
Selecting the Key Process(es). Organizations operate dozens of major cross-functional business
processes. From these a few key processes are selected as the PQM focus. The organization’s Strategic
Plan provides guidance in the selection of the key processes. (See Section 13, Strategic Deployment.)
There are several approaches to selecting key business processes:
 The Critical Success Factor approach holds that for any organization relatively few (no more than
eight) factors can be identified as “necessary and sufficient” for the attainment of its mission and
vision. Once identified, these factors are used to select the key business processes and rank them
by priority (Hardaker and Ward 1987).
 The Balanced Business Scorecard (Kaplan and Norton 1992) measures business performance in four
dimensions: financial performance, performance in the eyes of the customer, internal process performance,
and performance in organization learning and innovation. For each dimension, performance
measures are created and performance targets are set. Using these measures to track performance
provides a “balanced” assessment of business performance. The processes which create imbalances
in the scorecard are identified as the processes that most need attention—the key processes.
 Another approach is to invite upper management to identify a few (four to six) organizationspecific
critical selection criteria to use in evaluating the processes. Examples of such criteria are:
effect on business success, effect on customer satisfaction, significance of problems associated
with the process, amount of resources currently committed to the process, potential for improvement,
affordability of adopting process management, and effect of process on schedule. Using the
criteria and some simple scoring system (such as “low, medium, or high”), the managers evaluate
the many processes from the long list of the organization’s major business processes (10 to 25 of
them) and, by comparing the evaluations, identify the key processes. (The long list may be prepared
in advance in a process identification study conducted separately, often by the chief quality
officer, and often with the support of a consultant.
Whatever approach is used to identify key processes, the process map can be used to display the
results. The “process map” is a graphic tool for describing an organization in terms of its business
processes and their relationships to the organization’s principal stakeholders. The traditional organization
chart answers the question: “Who reports to whom?” The process map answers the question:
“How does the organization’s work get done?”
Figure 6.4 describes the work of the Educational Testing Service (ETS), the organization that prepares
and administers educational entrance examinations in the United States. In this process map,
organizations and departments are represented by shaded blocks labeled in bold type. The key operational
units of ETS, including external units designated “partners” by ETS, are located within a
boundary line labeled “ETS.” The important business processes of ETS are listed within that boundary
and marked by an asterisk (*). These are the processes eligible for the PQM focus, shown in their
relationship to various parts of the organization. Picturing the organization from a process perspective
provides upper management with a useful tool in thinking about and discussing the organization
in terms of its work and the processes it employs to get the work done.
Organizing: Assigning Ownership, Selecting the Team, and PQM Infrastructure. Because certain
major cross-functional business processes, the key processes, are critical to business success, the
FIGURE 6.4 Process map of major business processes at Educational Testing Services. (Source: Juran
Institute, Wilton, CT.)
quality council sees to it that those processes are organized in a special way. After selecting key
processes, the quality council appoints a process owner, who is responsible for making the process
effective, efficient, and adaptable, and is accountable for its performance (Riley, 1989 and 1994).
For large complex processes, especially in large companies, a two-tier ownership arrangement is
most often used. An appointed executive owner operates as a sponsor, champion, and supporter at
the upper management level, and is accountable for process results. At the operating level, a working
owner, usually a first- or second-level manager, leads the process-management team responsible
for day-to-day operation. The owner assignments—executive owner and working owner—are
ongoing. The major advantages of this structure are that there is at the same time “hands on” involvement
and support of upper management and adequate management of the process details.
The process-management team is a peer-level group which includes a manager or supervisor
from each major function within the process. Each member is an expert in a segment of the process.
Ideally, process management teams have no more than eight members, and the individuals chosen
should be proven leaders. The team is responsible for the management and continuous improvement
of the process. The team shares with the owner the responsibilities for effectiveness and efficiency.
Most commonly, the team assignments are ongoing.
From time to time a process owner creates an ad hoc team to address some special issue (human
resources, information technology, activity-based costing, etc.). The mission of such a projectoriented
team is limited, and the team disbands when the mission is complete. The ad hoc team is
different from the process-management team.
Figure 6.5 is a simplified diagram of a multifunctional organization and one of its major processes.
The shaded portions include: the executive owner, the working owner, the process management team,
and the stakeholders—functional heads at the executive level who have work activities of the business
process operating within their function. Customarily, the stakeholders are members of the quality council,
along with the executive owner. Taken together, this shaded portion is referred to as the PQM
Establishing the Team’s Mission and Goals. The preliminary process mission and improvement
goals for the process are communicated to the owners (executive and working levels) and team by
the quality council. To do their jobs most effectively, the owners and team must make the mission
and goals their own. They do this in the first step of the planning phase: defining the process.
The Planning Phase: Planning the New Process. The first phase of PQM is Planning,
which consists of five steps: (1) defining the process, (2) discovering customer needs and flow-
FIGURE 6.5 Organization infrastructure for process management in multifunctional organizations.
(Source: Juran Institute, Wilton, CT.)
charting the process, (3) establishing measurements of the process, (4) analyzing process measurements
and other data, and (5) designing (or redesigning) the process. The output of the Planning
Phase is the new process plan.
Defining the Current Process. The owner(s) and team collaborate to define the process precisely.
In accomplishing this, their starting point and principal reference is the process documentation
developed by the quality council during the selection of key processes and identification of owners
and teams. This documentation includes preliminary statements of mission and goals.
Effective mission and goal statements explicitly declare:
 The purpose and scope of the process
 “Stretch” targets for customer needs and business needs
(The purpose of the stretch target is to motivate aggressive process improvement activity.)
As an example, a mission statement for the Special-Contract Management Process is: Provide
competitive special pricing and supportive terms and conditions for large information systems procurements
that meet customer needs for value, contractual support, and timeliness at affordable cost.
The goals for the same process are:
1. Deliver approved price and contract support document within 30 days of date of customer’s letter
of intent.
2. Achieve a yield of special-contract proposals (percent of proposals closed as sales) of not less
than 50 percent.
The team must reach consensus on the suitability of these statements, propose modifications for
the quality council’s approval, if necessary, and also document the scope, objectives, and content.
Based on available data and collective team experience, the team will document process flow, the
process strengths and weaknesses, performance history, measures, costs, complaints, environment,
resources, and so on. This will probably involve narrative documentation and will certainly require
the use of flow diagrams.
Bounding the business process starts with an inventory of the major subprocesses—six to eight
of them is typical—that the business process comprises. The inventory must include the “startswith”
subprocess (the first subprocess executed), the “ends-with” subprocess (the last executed), and
the major subprocesses in between. If they have significant effect on the quality of the process output,
activities upstream of the process are included within the process boundary. To provide focus
and avoid ambiguity, it is also helpful to list subprocesses which are explicitly excluded from the
business process. The accumulating information on the process components is represented in diagram
form, which evolves, as the steps of the planning phase are completed, from a collection of subprocesses
to a flow diagram.
Figure 6.6 shows the high-level diagram of the special-contract process that resulted from process
analysis but before the process was redesigned. At the end of the process definition step such a diagram
is not yet a flow diagram, as there is no indication of the sequence in which the subprocesses
occur. Establishing those relationships as they presently exist is the work of Step 2.
Discovering Customer Needs and Flowcharting the Process. For the process to do its work well,
the team must identify all of the customers, determine their needs, and prioritize the information.
Priorities enable the team to focus its attention and spend its energies where they will be most effective.
(The subject of identifying customers and their needs is covered in detail in Section 3, The
Quality Planning Process.)
Determining customer needs and expectations requires ongoing, disciplined activity. Process
owners must ensure that this activity is incorporated in the day-to-day conduct of the business
process as the customer requirements subprocess and assign accountability for its performance. The
output of this vital activity is a continually updated customer requirement statement.
On the process flow chart it is usual to indicate the key suppliers and customers and their roles
in the process, as providers or receivers of materials, product, information, and the like. Although the
diagram can serve a number of specialized purposes, the most important here is to create a common,
high-level understanding among the owner and team members of how the process works—how the
subprocesses relate to each other and to the customers and suppliers and how information and product
move around and through the process. In creating the process flow chart, the team will also verify
the list of customers and may, as understanding of the process deepens, add to the list of
The process flow chart is the team’s primary tool for analyzing the process to determine whether
it can satisfy customer needs. By walking through the chart together, step by step, sharing questions
and collective experience, the team determines whether the process is correctly represented, making
adjustments to the diagram as necessary to reflect the process as it presently operates.
When the step is complete, the team has a starting point for analysis and improvement of the
process. In Figure 6.8, product flow is shown by solid lines and information flow by dotted lines.
Establishing Process Measurements. What gets measured, gets done. Establishing, collecting,
and using the correct measures is critical in managing business process quality. “Process capability,”
“process performance,” and other process measures have no practical significance if the process they
purport to describe is not managed. To be managed, the process must fulfill certain minimum conditions:
a. It has an owner.
b. It is defined.
c. Its management infrastructure is in place.
d. Its requirements are established.
e. Its measurements and control points are established.
f. It demonstrates stable, predictable, and repeatable performance.
A process which fulfills these minimum conditions is said to be manageable. Manageability is
the precondition for all further work in PQM.
FIGURE 6.6 High-level diagram of the special-contract process, an output of process analysis. (Source:
Juran Institute, Wilton, CT.)
Of these criteria, (a) through (d) have already been addressed in this section. Criteria (e) and (f)
are addressed in the following.
Process Measurements (See also Section 9). In deciding what aspects of the process to measure,
we look for guidance to the process mission and to our list of customer needs. Process measures
based on customer needs provide a way of measuring process effectiveness. For example, if
the customer requires delivery of an order within 24 hours of order placement, we incorporate into
our order-fulfillment process a measure such as “time elapsed between receipt of order and delivery
of order,” and a system for collecting, processing, summarizing, and reporting information
from the data generated. The statistic reported to the executive owner will be one such as “percent
of orders delivered within 24 hours,” a statistic which summarizes on-time performance. The team
will also need data on which to base analysis and correction of problems and continuous improvement
of the process. For this purpose, the team needs data from which they can compute such
descriptive statistics as distribution of delivery times by product type, and so on. The uses to which
the data will be put must be thought through carefully at the time of process design to minimize
the redesign of the measures and measurement systems.
Process measures based on cost, cycle time, labor productivity, process yield, and the like are
measures of process efficiency. Suppose that a goal for our order-fulfillment process is to reduce
order-picking errors to one error per thousand order lines. Managing to that goal requires identification
of order-picking errors in relation to the number of order lines picked. For order-picking errors
that are inadvertent—that is, when they happen, the picker is unaware of them—measuring them
requires a separate inspection to identify errors. In a random audit on a sample of picked orders, an
inspector identifies errors and records them. As with delivery-time measurement, the team must
think through all the uses it will make of these measurements. For a report of estimated error rate,
the data needed are: number of errors and number of order lines inspected. To improve process performance
in this category, the data must help the team identify error sources and determine root
cause. For that to occur, each error must be associated with time of day, shift, product type, size of
package, etc., so that the data can be stratified to test various theories of root cause.
While not a measurement category, process adaptability is an important consideration for process
owners and teams. Adaptability will be discussed later in this section.
Process measurements must be linked to business performance. If certain key processes must run
exceptionally well to ensure organization success, it follows that collective success of the key processes
is good for the organization’s performance. Process owners must take care to select process measures
that are strongly correlated with traditional business indicators, such as revenue, profit, ROI, earnings
per share, productivity per employee, and so on. In high-level business plan reviews, managers are
motivated and rewarded for maintaining this linkage between process and organization performance
measures because of the two values which PQM supports: organization success is good, and process
management is the way we will achieve organization success.
Figure 6.7 shows some typical process measurements and the traditional business indicators with
which they are linked. To illustrate, “percent of sales quota achieved” is a traditional business indicator
relating to the business objective of improving revenue. The special-contract management
process has a major impact on the indicator, since more than 30 percent of U.S. revenue comes from
that process. Therefore, the contract close rate (ratio of the value of firm contracts to the total value
of proposals submitted) of the special-contract management process is linked to percent of sales
quota and other traditional revenue measures, and is therefore a measure of great importance to management.
Measurement points appear on the process flow diagram.
Control Points. Process measurement is also a part of the control mechanisms established to maintain
planned performance in the new process. To control the process requires that each of a few
selected process variables be the control subjects of a feedback control loop. Typically, there will be
five to six control points at the macroprocess level for variables associated with: external output,
external input, key intermediate products, and other high-leverage process points.
The control points in the special-contract management process are represented graphically in
Figure 6.8. Feedback loop design and other issues surrounding process control are covered in detail
in Section 4, The Quality Control Process.
Process Variability, Stability, and Capability. As in all processes, business processes exhibit variability.
The tools of statistical process control such as Shewhart charts (see Section 45, Statistical
Process Control) help the team to minimize process variation and assess process stability.
Evaluation of process capability is an important step in process quality improvement. Process
capability is a measure of variation in a process operating under stable conditions. “Under stable conditions”
means that all variation in the process is attributable to random causes. The usual criterion for
stability is that the process, as plotted and interpreted on a Shewhart control chart, is “in control.”
The Traditional Business View The Process View
Business Objective Business Indicator Key Process Process Measure
Higher revenue Percent of sales quota Contract management Contract close rate
Percent of revenue plan Product development Development cycle time
Value of orders canceled Account management Backlog management
after shipment and system assurance
Receivable days Billing quality index
Reduce costs Inventory turns Manufacturing Manufacturing cycle
FIGURE 6.7 Linkages among business objectives, traditional business indicators, and process measures generated
by the process-management approach—a few examples. (Source: Juran Institute, Wilton, CT.)
FIGURE 6.8 Flowchart of the special-contract management process, including process control points.
(Source: Juran Institute, Wilton, CT.)
Statistical process control, process capability, and associated tools are useful components of the
process team’s tool kit. They are covered in detail in Section 44, Basic Statistical Methods.
The output of the measurement step is a measurement plan, a list of process measurements to be
made and the details of making each one—who will make it, how it will be made, on what schedule,
and so on.
Analyzing the Process. Process Analysis is performed for the following purposes:
 Assess the current process for its effectiveness and efficiency.
 Identify the underlying causes of any performance inadequacy.
 Identify opportunities for improvement.
 Make the improvements.
First, referring to the process flowchart, the team breaks the process into its component activities
using a procedure called “process decomposition,” which consists of progressively breaking apart
the process, level by level, starting with the macrolevel. As decomposition proceeds, the process is
described in ever finer detail.
As the strengths and weaknesses of the process are understood at one level, the process management
team’s interim theories and conclusions will help decide where to go next with the analysis.
The team will discover that certain subprocesses have more influence on the performance of the
overall business process than others (an example of the Pareto principle). These more significant subprocesses
become the target for the next level of analysis.
Decomposition is complete when the process parts are small enough to judge as to their effectiveness
and efficiency. Figure 6.9 gives examples from three levels of decomposition (subprocess,
activity, and task) of three typical business processes (procurement, development engineering, and
office administration).
Measurement data are collected according to the measurement plan to determine process effectiveness
and efficiency. The data are analyzed for effectiveness (conformance to customer needs) and
long-term capability to meet current and future customer requirements.
The goal for process efficiency is that all key business processes operate at minimum total
process cost and cycle time, while still meeting customer requirements.
Process effectiveness and efficiency are analyzed concurrently. Maximizing effectiveness and efficiency
together means that the process produces high quality at low cost; in other words, it can provide
the most value to the customer.
“Business process adaptability” is the ability of a process to readily accommodate changes both in
the requirements and the environment, while maintaining its effectiveness and efficiency over time.
To analyze the business process, the flow diagram is examined in four steps and modified as necessary.
The steps are:
 Examine each decision symbol
Is this a checking activity?
Business Process Subprocess Activity Task
Procurement Vendor selection Vendor survey Documentation of outside
Development Hardware design Engineering change Convening the Change
engineering Board
Office administration Providing secretarial Calendar management Making a change to existing
services calendar
FIGURE 6.9 Process decomposition—examples of process elements disclosed within typical business processes.
(Source: Juran Institute, Wilton, CT.)
If so, is this a complete check, or do some types of errors go undetected?
Is this a redundant check?
 Examine each rework loop
Would we need to perform these activities if we had no failures?
How “long” is this rework loop (as measured in number of steps, time lost, resources consumed,
Does this rework loop prevent the problem from recurring?
 Examine each activity symbol
Is this a redundant activity?
What is the value of this activity relative to its cost?
How have we prevented errors in this activity?
 Examine each document and database symbol
Is this necessary?
How is this kept up to date?
Is there a single source for this information?
How can we use this information to monitor and improve the process?
The “Process Analysis Summary Report” is the culmination and key output of this process analysis
step. It includes the findings from the analysis, that is, the reasons for inadequate process performance
and potential solutions that have been proposed and recorded by owner and team as analysis progressed.
The completion of this report is an opportune time for an executive owner/stakeholder review.
The owner/stakeholder reviews can be highly motivational to owners, teams, stakeholders, and
the Quality Council. Of particular interest is the presentation of potential solutions for improved
process operation. These have been collected throughout the planning phase and stored in an idea
bin. These design suggestions are now documented and organized for executive review as part of the
process analysis summary report presentation.
In reviewing the potential solutions, the executive owner and quality council provide the selection
criteria for acceptable process design alternatives. Knowing upper management’s criteria for
proposed solutions helps to focus the process-management team’s design efforts and makes a favorable
reception for the reengineered new process plan more likely.
Designing (or Redesigning) the Process. In Process Design, the team defines the specific operational
means for meeting stated product goals. The result is a newly developed Process Plan. Design
changes fall into five broad categories: workflow, technology, people and organization, physical
infrastructure, and policy and regulations.
In the design step, the owner and team must decide whether to create a new process design or to
redesign the existing process. Creating a new design might mean radical change; redesign generally
means incremental change with some carryover of existing design features.
The team will generate many design alternatives, with input from both internal and external
sources. One approach to generating these design alternatives from internal sources is to train tasklevel
performers to apply creative thinking to the redesign of their process.
Ideas generated in these sessions are documented and added to the idea bin. Benchmarking can
provide a rich source of ideas from external sources, including ideas for radical change.
Benchmarking is discussed in detail in Section 12.
In designing for process effectiveness, the variable of most interest is usually process cycle time.
In service-oriented competition, lowest process cycle time is often the decisive feature. Furthermore,
cycle-time reduction usually translates to efficiency gains as well. For many processes, the most
promising source of cycle-time reduction is the introduction of new technology, especially information
Designing for speed creates surprising competitive benefits: growth of market share and reduction
of inventory requirements. Hewlett-Packard, Brunswick Corp., GE’s Electrical Distribution and
Control Division, AT&T, and Benetton are among the companies who have reported stunning
achievements in cycle-time reduction for both product development and manufacturing (Dumaine,
1989). In each of the companies, the gains resulted from efforts based on a focus on major processes.
Other common features of these efforts included:
 Stretch objectives proposed by top management
 Absolute adherence to schedule, once agreed to
 Application of state-of-the art information technology
 Reduction of management levels in favor of empowered employees and self-directed work teams
 Putting speed in the culture
In designing for speed, successful redesigns frequently originate from a few relatively simple
guidelines: eliminate handoffs in the process, eliminate problems caused upstream of activity,
remove delays or errors during handoffs between functional areas, and combine steps that span businesses
or functions. A few illustrations are provided:
 Eliminate handoffs in the process: A “handoff” is a transfer of material or information from one
person to another, especially across departmental boundaries. In any process involving more than
a single person, handoffs are inevitable. It must be recognized, however, that the handoff is timeconsuming
and full of peril for process integrity—the missed instruction, the confused part identification,
the obsolete specification, the miscommunicated customer request.
In the special-contract management process, discussed previously in this section, the use of concurrent
review boards eliminated the 28 sequential executive approvals and associated handoffs.
 Eliminate problems caused upstream of activity. Errors in order entry at a U.S. computer company
were caused when sales representatives incorrectly configured systems. As a result, the cost
of the sales-and-order process was 30 percent higher than that of competitors, and the error rates
for some products were as high as 100 percent. The cross-functional redesign fixed both the configurations
problem and sales-force skills so that on-time delivery improved at significant cost savings
(Hall, Rosenthal, and Wade 1993).
 Remove delays or errors during handoffs between functional areas: The processing of a new policy
at a U.K. insurance company involved 10 handoffs and took at least 40 days to complete. The
company implemented a case-manager approach by which only one handoff occurred and the policy
was processed in less than 7 days (Hall, Rosenthal, and Wade 1993).
 Combine steps that span businesses or functions: At a U.S. electronics equipment manufacturer,
as many as seven job titles in three different functions were involved in the nine steps required to
design, produce, install, and maintain hardware. The company eliminated all but two job titles,
leaving one job in sales and one job in manufacturing (Hall, Rosenthal, and Wade 1993).
The Ford accounts payable process provides a classic example of process redesign. Details are
given by Hammer and Champy (1993). Process Quality Management is successful when the design
step involved radical change. Hammer and Champy propose the following principles for such radical
change of a process:
 Organize the process around outcomes, not tasks.
 Have those who use the output of the process perform the process.
 Incorporate information-processing work into the real work that produces the information.
 Treat geographically dispersed resources as though they were centralized.
 Coordinate parallel functions within the process, not in subsequent steps.
 Put the decision point where the work is performed and build control into the process.
 Capture information only once and at the source.
Before the new design is put into place, a design review is in order. Its purpose is to temper the
enthusiasm of the team with the objectivity of experienced outsiders. Typically, the process owner
assembles a panel of experts from within the organization (but outside the process) to provide the
evaluation of design alternatives.
Process design testing is performed to determine whether the process design alternative will work
under operating conditions. Design testing may include trials, pilots, dry runs, simulations, etc. The
results are used to predict new process performance and cost/benefit feasibility.
Successful process design requires employee participation and involvement. To overlook such
participation creates a lost opportunity and a barrier to significant improvement. The creativity of the
first-line work force in generating new designs can be significant.
Byrne (1993) reports that many companies are adopting “horizontal organizational design,”which features
the use of self-directed work teams organized around the process. Eastman Chemical has over 1000
teams; increasing reliance on self-directed teams has enabled the company to eliminate senior VP positions
for administration, manufacturing, and R&D. (See also Section 15, Human Resources and Quality.)
Lexmark International, a former IBM division, abolished 60 percent of the management jobs
in manufacturing and support services. Instead, they organized around cross-functional teams
Creating the New Process Plan. After we have redefined a key process, we must document the
new process and carefully explain the new steps. The new process plan now includes the new
process design and its control plan for maintaining the new level of process performance. The
new process plan for the special-contract management process, shown as a high-level process
schematic, is shown in Figure 6.10.
The Transfer Phase: Transferring the New Process Plan to Operations. The
transfer phase consists of three steps: (1) planning for implementation problems, (2) planning for
implementation action, and (3) deploying the new process plan.
FIGURE 6.10 Completed process plan diagram for the special-contract management process. (Source: Juran
Institute, Wilton, CT.)
Planning for Implementation Problems. A major PQM effort may involve huge expenditures and
precipitate fundamental change in an organization, affecting thousands of jobs. All of this poses
major management challenges. All of the many changes must be planned, scheduled, and completed
so that the new process may be deployed to operational management. Figure 6.11 identifies specific
categories of problems to be addressed and the key elements that are included.
Of the five categories listed in Figure 6.11, People and Organization is usually the source of the
most challenging change issues in any PQM effort. Implementation issues in the people and organizational
design category include: new jobs, which are usually bigger; new job descriptions; training
people in the new jobs; new performance plans and objectives; new compensation systems (incentive
pay, gainsharing, and the like); new recognition and reward mechanisms; new labor contracts with
unions; introduction of teamwork and team-building concepts essential to a process orientation; formation
of self-directed work teams; team education; reduction of management layers; new reporting
relationships; development and management of severance plans for those whose jobs are eliminated;
temporary continuation of benefits; out-placement programs; and new career paths based on knowledge
and contribution, rather than on promotion within a hierarchy. The list goes on. Additionally,
there are changes in technology, policy, physical infrastructure, etc., to be dealt with.
The importance of change management skills becomes clear. Deploying a new process can be a
threat to those affected. The owner and team must be skilled in overcoming resistance to change.
Creating Readiness for Change: A Model for Change. Change happens when four conditions are
combined. First, the current state must be seen as unsatisfactory, even painful; it must constitute a
tension for change. Second, there must be a satisfactory alternative, a vision of how things can be
better. Third, some practical steps must be available to reach the satisfactory state, including instruction
in how to take the steps, and support during the journey. Fourth, to maintain the change, the
organization and individuals must acquire skills and reach a state of self-efficacy.
These four conditions reinforce the intent to change. Progress toward that change must be monitored
continuously so as to make the change a permanent one. In the operational management phase,
operational controls, continuous improvement activity, and ongoing review and assessment all contribute
to ensuring that the new process plan will continue to perform as planned. (See also
Resistance to Change and how to deal with it in Section 5, The Quality Improvement Process.)
Planning for Implementation Action. The output of this step is a complex work plan, to be carried
out by the Owner and Process Management Team. They will benefit from skills in the techniques
of Project Management. (See Section 17, Project Management and Product Development.)
Deploying the New Process Plan. Before actually implementing the new process, the team tests
the process plan. They test selected components of the process and may carry out computer simulations.
The purpose is to predict the performance of the new process and determine feasibility. Also,
the tests help the team refine the “roll out” of the process and decide whether to conduct parallel
Category Key Elements Included
Workflow Process anatomy (macro/micro, cross-functional, intrafunctional, interdepartmental,
and intradepartmental)
Technology Information technology and automation
People and organization Jobs, job description, training and development, performance management,
compensation (incentive-based or not), recognition/reward,
union involvement, teams, self-directed work teams, reporting relationships
and delayering
Infrastructure (physical) Location, space, layout, equipment, tools, and furnishings
Policy/regulations Government, community, industry, company, standards, and culture
New-process design issues
FIGURE 6.11 Design categories. (Source: Juran Institute, Wilton, CT.)
operation (old process and new process running concurrently). The team must decide how to deploy
the new process. There are several options:
 Horizontal deployment, function by function.
 Vertical deployment, top down, all functions at once.
 Modularized deployment, activity by activity, until all are deployed.
 Priority deployment, subprocesses and activities in priority sequence, those having the highest
potential for improvement going first.
 Trial deployment, a small-scale pilot of the entire process, then expansion for complete implementation.
This technique was used in the first redesign of the Special-Contract Management
process, that is, a regional trial preceded national expansion. The insurance company USAA conducts
all pilot tests of new process designs in their Great Lakes region. In addition to “working the
bugs out of the new design before going national,” USAA uses this approach as a “career-broadening
experience for promising managers,” and to “roll out the new design to the rest of the organization
with much less resistance” (Garvin 1995).
Full deployment of the new process includes the development and deployment of an updated control
plan. Figure 6.12 lists the contents of a new process plan.
Operational Management Phase: Managing the New Process. The Operational
Management Phase begins when the process is put into operation. The major activities in operational
management are: (1) process quality control, (2) process quality improvement, and (3) periodic
process review and assessment.
Process Quality Control. “Process control” is an ongoing managerial process, in which the actual
performance of the operating process is evaluated by measurements taken at the control points, comparing
the measurements to the quality targets, and taking action on the difference. The goal of
process control is to maintain performance of the business process at its planned level. (See Section
4, The Quality Control Process).
Process Quality Improvement. By monitoring process performance with respect to customer
requirements, the process owner can identify gaps between what the process is delivering and what
is required for full customer satisfaction. These gaps are targets for process quality improvement
 Process mission
 Process goals
 Process management infrastructure (that is, owner/team/stakeholders)
 Process contract
 Process description/model
 Customer requirements (that is, customer list, customer needs, and requirements statement)
 Process flow
 Measurement plan
 Process analysis summary report
 Control plan
 Implementation action plan
 Resource plan
FIGURE 6.12 Contents of complete process plan (Source: Juran Institute, Wilton, CT.)
efforts. They are signaled by defects, complaints, high costs of poor quality, and other deficiencies.
(See Section 5, The Quality Improvement Process.)
Periodic Process Review and Assessment. The owner conducts reviews and assessments of current
process performance to ensure that the process is performing according to plan. The review
should include review and assessment of the process design itself to protect against changes in the
design assumptions and anticipated future changes such as changes in customer needs, new technology
or competitive process designs. It is worthwhile for the process owner to establish a schedule
for reviewing the needs of customers and evaluating and benchmarking the present process.
As customer needs change, process measures must be refined to reflect these changes. This continuous
refinement is the subject of a measurement management subprocess, which is established by the
owners and team and complements the customer needs subprocess. The two processes go hand in hand.
The process management category in the Malcolm Baldrige National Quality Award criteria
(1998) provides a basis for management review and assessment of process performance.
Other external award criteria from worldwide sources, as well as many national and international
standards, serve as inspiration and guidance for owners and teams contemplating process reviews.
(See Section 14, Total Quality Management, and Section 11, The ISO 9000 Family of International
The criteria of the Malcolm Baldrige National Quality Award have come to be regarded as the de
facto definition of TQM. (See Section 14.) Process quality management is an important concept
within the TQM framework.
Organizations have learned not to limit managerial attention to the financial dimension. They
have gained experience in defining, identifying, and managing the quality dimension. They are
accustomed to thinking strategically—setting a vision, mission, and goals, all in alignment. And they
will have experience reviewing progress against those goals.
The quality improvement process, which began in Japan in the 1950s and was widely deployed
in the United States in the early 1980s, was an important step beyond functional management.
Organizations found that quality improvement required two new pieces of organization machinery—
the quality council and the cross-functional project team. The Quality Council usually consists of the
senior management team; to its traditional responsibility for management of finance the responsibility
for the management of quality is added. The project team recognizes that, in a functional
organization, responsibility for reduction of chronic deficiencies has to be assigned to a crossfunctional
PQM is a natural extension of many of the lessons learned in early quality improvement activities.
It requires a conceptual change—from reliance on functional specialization to an understanding
of the advantages of focusing on major business processes. It also requires an additional piece of
organization machinery: an infrastructure for each of the major processes.
Key Points. PQM is distinguished by the following:
 A strategic orientation, that is
A clear mission, values, and vision for the organization
Strategic goals tied to the organization vision, which are shared by executive leadership and
deployed throughout the organization in the form of key business objectives
Alignment and linkage of the organization’s processes to its vision, strategic goals, and objectives
 A cross-functional orientation in place of the hierarchical organization.
 Cross-functional process teams, supported by the management system (education, communication,
performance management, recognition and reward, compensation, new career path structures,
etc.). The mission of each team is to dramatically improve the effectiveness, efficiency, and adaptability
of each major business process to which it is assigned.
 Prime organizational focus on the needs of customers, external and internal, and business needs
such as cost, cycle time, waste elimination.
 The driving of all work processes by quality of products and services and overall value creation.
Critical Success Factors for PQM Implementation. The following factors are important
to the success of a PQM initiative:
 Leadership from the top of the organization
 Management which communicates the vision, strategic goals, and key business objectives throughout
the organization
 Vision shared by all in the organization
 Employees empowered and accountable to act in support of these key business objectives
 Expertise in change management available throughout the organization to facilitate dramatic
 Continuous improvement
 Widespread skills in project management to enable the many PQM teams to manage schedules,
costs, and work plans being coordinated and implemented throughout the organization
 Executive management promotion of the importance, impact, progress, and success of the PQM
effort throughout the organization, and to external stakeholders
 Upper management’s obligation is to enable and promote three principal objectives: customer
focus, process orientation, and empowered employees at all levels
Leaders of those organizations who have adopted PQM as a management tool know that Process
Quality Management is a continuous managerial focus, not a single event or a quick fix. They also
know that a constant focus on business processes is essential to the long-term success of their organization.
AT&T Quality Steering Committee (1988). Process Quality Management & Improvement Guidelines, AT&T
Information Center, Indianapolis, IN.
Black, John (1985). Business Process Analysis—Guide to Total Quality, Boeing Commercial Airplane Company,
Seattle, WA, revised 1987.Byrne, John A. (1993). “The Horizontal Corporation,” Business Week, Dec. 20, pp.
Dumaine, Brian (1989). “How Managers Can Succeed Through Speed,” Fortune, Feb. 13, pp. 54–60.
Garvin, David A. (1995). “Leveraging Processes for Strategic Advantage,” Harvard Business Review, Sept./Oct.,
vol. 73, no. 5, pp. 77–90.
Gibson, Michael J. W. (1991–92). “The Quality Process: Business Process Quality Management,” International
Manufacturing Strategy Resource Book, pp. 167–179.
Gibson, Michael J. W., Gabriel, A., and Riley, James F., Jr. (1990). Managing Business Process Quality (MBPQ),
1st ed., Juran Institute, Inc., Wilton, CT.
Hall, Gene, Rosenthal, Jim, and Wade, Judy (1993). “How to Make Reengineering Really Work,” Harvard
Business Review, Nov./Dec., vol. 71, no. 6, pp. 199–130.
Hammer, Michael, and Champy, James (1993). Reengineering the Corporation, Harper Collins, New York.
Hardaker, Maurice, and Ward, Bryan K. (1987). “Getting Things Done: How to Make a Team Work,” Harvard
Business Review, Nov./Dec., vol. 65, pp. 112–119.
Kane, Edward J. (1986). “IBM’s Focus on the Business Process,” Quality Progress, April, p. 26.
Kane, Edward J. (1992). “Process Management Methodology Brings Uniformity to DBS,” Quality Progress,
June, vol. 25, no. 6, pp. 41–46.
Kaplan, Robert S., and Norton, David P. (1992). “The Balanced Scorecard—Measures that Drive Performance,”
Harvard Business Review, Jan./Feb., vol. 7, no. 1, pp. 71–79, reprint #92105.
Pall, Gabriel A. (1987). Quality Process Management, Prentice-Hall, Inc., Englewood Cliffs, NJ.
Riley, James F., Jr. (1989). Executive Quality Focus: Discussion Leader’s Guide, Science Research Associates,
Inc., Chicago.
Riley, James F., Jr., Pall, Gabriel A., and Harshbarger, Richard W. (1994). Reengineering Processes for
Competitive Advantage: Business Process Quality Management (BPQM), 2nd ed., Juran Institute, Inc., Wilton,
Rummler, Geary (1992). “Managing the White Space: The Work of Geary Rummler,” Training and Development,
Special Report, August, pp. 26–30.
Schlesiona, Peter (1988). Business Process Management, Science Research Associates, Inc., Chicago.
Zachman, James W. (1990). “Developing and Executing Business Strategies Using Process Quality
Management,” IMPRO Conference Proceedings, Juran Institute, Wilton, CT., pp. 2a-9–21.
J. M. Juran
National Affluence and Organization 7.2
Life Behind the Quality Dikes 7.3
Voluntary Obsolescence 7.3
Involuntary Obsolescence 7.3
The Spectrum of Affluence 7.4
Fitness for Use and Conformance to
Specification 7.4
Stated Needs and Real Needs 7.6
Psychological Needs 7.6
“User-Friendly” Needs 7.7
The Need to Be Kept Informed 7.7
Cultural Needs 7.8
Needs Traceable to Unintended Use 7.8
Inadequate Available Products 7.11
Relief from Onerous Chores 7.11
Reduction of Time for Service 7.11
Changes in Customer Habits 7.11
Role of the Entrepreneur 7.11
Consumer Products 7.12
Price Differences 7.12
Efforts to Quantify Value 7.13
Industrial Products 7.14
Commodity versus Specialty or System
Specialties; the “Bundled” Price 7.14
Effect of Quality Superiority 7.15
Consumer Preference and Share of
Market 7.16
Industrial Products and Share of Market
Quality and Competitive Bidding 7.18
Building Quality Leadership 7.20
The “Market Leader” Concept 7.20
Carryover of Failure-Prone Features 7.21
Concept of the Optimum 7.22
Steps in Life Cycle Cost Analysis 7.22
Breadth of Application 7.23
Application to Consumer Products 7.23
Application to Industrial Products 7.25
Application to Defense Industries 7.26
Cultural Resistance 7.26
Contracts Based on Amount of Use 7.26
Perfectionism in Quality of Design 7.27
Perfectionism in Quality of Conformance
The Perfectionists 7.28
Quality affects company economics in two principal ways:
The effect of quality on costs: In this case “quality” means freedom from troubles traceable to
office errors, factory defects, field failures, and so on. Higher “quality” means fewer errors, fewer
defects, and fewer field failures. It takes effort to reduce the numbers of such deficiencies, but in
the great majority of cases, the end result is cost reduction. This type of effect of quality on company
economics is discussed in Section 8, Quality and Costs.
The effect of quality on income: In this case “quality” means those features of the product which
respond to customer needs. Such features make the product salable and provide “product satisfaction”
to customers. Higher quality means better and/or more features which provide greater
satisfaction to customers.
This section focuses on the relationship between product features and company income.
(“Company” includes any operating institution—an industrial company, a government agency, a
school, and so on. “Income” means gross receipts, whether from sales, appropriations, tuitions, and
so on.) The section discusses the forces through which quality affects income and the methods in use
for studying the cause-effect relationships. Closely related to this subject of quality and income are
two other sections of this handbook:
Market Research and Marketing (Section 18)
Customer Service (Section 25)
The above two effects of quality—on costs and on income—interact with each other. Product deficiencies
not only add to suppliers’ and customers’ costs, they also discourage repeat sales. Customers
who are affected by field failures are, of course, less willing to buy again from the guilty supplier. In
addition, such customers do not keep this information to themselves—they publicize it so that it
becomes an input to other potential buyers, with negative effects on the sales income of the supplier.
In recent decades there has been much study of the effect of poor quality on company economics.
(See generally, Section 8, Quality and Costs.) In contrast, study of the effect of quality on income
has lagged. This imbalance is all the more surprising since most upper managers give higher priority
to increasing income than to reducing costs. This same imbalance presents an opportunity for
improving company economics through better understanding of the effect of quality on income.
The ability of an industrial company to secure income is strongly influenced by the economic climate
and by the cultural habits which the various economies have evolved. These overriding influences
affect product quality as well as other elements of commerce.
National Affluence and Organization. The form of a nation’s economy and its degree of
affluence strongly influence the approach to its quality problems.
Subsistence Economies. In such economies the numerous impoverished users have little choice
but to devote their income to basic human needs. Their protection against poor quality is derived
more from their collective political power than from their collective economic power. Most of the
world’s population remains in a state of subsistence economy.
Planned Economies. In all countries there are some socialized industries—government monopolies
for some products or services. In some countries the entire economy is so organized. These
monopolies limit the choice of the user to those qualities which result from the national planning and
its execution. For elaboration, see Section 36, Quality and the National Culture.
Shortages and Surpluses. In all economies, a shortage of goods (a “sellers’ market”) results in a
relaxing of quality standards. The demand for goods exceeds the supply, so users must take what they
can get (and bid up the price to boot). In contrast, a buyers’ market results in a tightening of quality
Life Behind the Quality Dikes. As societies industrialize, they revise their lifestyle in order
to secure the benefits of technology. Collectively, these benefits have greatly improved the quality of
life, but they have also created a new dependence. In the industrial societies, great masses of human
beings place their safety, health, and even their daily well-being behind numerous “quality dikes.” For
elaboration, see Section 35, Quality and Society, under the heading Life Behind the Quality Dikes.
Voluntary Obsolescence. As customers acquire affluence, the industrial companies increasingly
bring out new products (and new models of old products) which they urge prospective users to
buy. Many of the users who buy these new models do so while possessing older models which are
still in working order. This practice is regarded by some economists and reformers as a reprehensible
economic waste.
In their efforts to put an end to this asserted waste, the reformers have attacked the industrial companies
who bring out these new models and who promote their sale. Using the term “planned obsolescence,”
the reformers imply (and state outright) that the large companies, by their clever new
models and their powerful sales promotions, break down the resistance of the users. Under this theory,
the responsibility for the waste lies with the industrial companies who create the new models.
In the experience and judgment of the author, this theory of planned obsolescence is mostly nonsense.
The simple fact, obvious both to manufacturers and consumers, is that the consumer makes
the decision (of whether to discard the old product and buy the new). Periodically, this fact is dramatized
by some massive marketing failure.
A few decades ago E.I. DuPont de Nemours & Co., Inc. (DuPont) brought out the product
Corfam, a synthetic material invented to compete with leather for shoe uppers (and for other
applications). Corfam was a technological triumph. Though costly, it possessed excellent properties
for shoe uppers: durability, ease of care, shape retention, scuff resistance, water repellency,
and ability to “breathe.” DuPont became a major supplier of shoe uppers materials, but in 1971
it withdrew from the business because Corfam “never attained sufficient sales volume to show a
Industry observers felt that the high durability of Corfam was an irrelevant property due to rapid
style obsolescence; i.e., the life of the shoes was determined not by the inherent durability of Corfam,
but by style obsolescence. In essence, a large corporation undertook a program which was antagonistic
to obsolescence, but the users decided against it. DuPont’s investment in Corfam may have
exceeded $100 million.
In a case involving an even larger investment, the Ford Motor Company’s Edsel automobile failed
to gain consumer acceptance despite possessing numerous product innovations and being promoted
by an extensive marketing campaign.
Involuntary Obsolescence. A very different category of obsolescence consists of cases in
which long-life products contain failure-prone components which will not last for the life of the
product. The life of these components is determined by the manufacturer’s design. As a result, even
though the user decides to have the failed component replaced (to keep the product in service), the
manufacturer has made the real decision because the design determined the life of the component.
This situation is at its worst when the original manufacturer has designed the product in such a
way that the supplies, spare parts, and so on are nonstandard, so that the sole source is the original
manufacturer. In such a situation, the user is locked into a single source of supply. Collectively, such
cases have lent themselves to a good deal of abuse and have contributed to the consumerism movement.
(For elaboration, see Juran 1970.)
Industrial companies derive their income from the sale of their products. These sales are made to
“customers,” but customers vary in their functions. Customers may be merchants, processors, ultimate
users, and so on, with resulting variations in customer needs. Response to customer needs
requires a clear understanding of just what those needs are.
Human needs are complex and extend beyond technology into social, artistic, status, and other
seemingly intangible areas. Suppliers are nevertheless obliged to understand these intangibles in
order to be able to provide products which respond to such needs.
The Spectrum of Affluence. In all economies the affluence of the population varies across
a wide spectrum. Suppliers respond to this spectrum through variations in product features. These
variations are often called “grades.”
For example, all hotels provide overnight sleeping accommodations. Beyond this basic service,
hotels vary remarkably in their offerings, and the grades (deluxe, four star, and so on) reflect this
variation. In like manner, any model of automobile provides the basic service of point-to-point transportation.
However, there are multiple grades of automobiles. The higher grades supply services
beyond pure transportation—higher levels of safety, comfort, appearance, status, and so on.
Fitness for Use and Conformance to Specification. Customers and suppliers sometimes
differ in their definition of what is quality. Such differences are an invitation to trouble. To
most customers, quality means those features of the product which respond to customer needs. In
addition, quality includes freedom from failures, plus good customer service if failures do occur. One
comprehensive definition for the above is “fitness for use.”
In contrast, many suppliers had for years defined quality as conformance to specification at the
time of final product test. This definition fails to consider numerous factors which influence quality
as defined by customers: packaging, storage, transport, installation, reliability, maintainability, customer
service, and so on.
Table 7.1 tabulates some of the differences in viewpoint as applied to long-life goods.
The ongoing revolution in quality has consisted in part of revising the suppliers’ definition of
quality to conform more nearly with the customers’ definition.
Cost of Use. For consumable products, the purchase price paid by the customer is quite close to
the cost of using (consuming) the product. However, for long-lived products, the cost of use can
diverge considerably from the purchase price because of added factors such as operating costs, maintenance
costs, downtime, depreciation, and so on.
The centuries-old emphasis on purchase price has tended to obscure the subsequent costs of use.
One result has been suboptimization; i.e., suppliers optimize their costs rather than the combined
costs of suppliers and customers.
The concept of life-cycle costing offers a solution to this problem, and progress is being made in
adopting this concept. (See Life Cycle Costing, below.)
Degrees of User Knowledge. In a competitive market, customers have multiple sources of supply.
In making a choice, product quality is an obvious consideration. However, customers vary greatly in
their ability to evaluate quality, especially prior to purchase.
Table 7.2 summarizes the extent of customer knowledge and strength in the marketplace as related
to quality matters.
The broad conclusions which can be drawn from Table 7.2 are as follows:
 Original equipment manufacturers (OEMs) can protect themselves through their technological
and/or economic power as much as through contract provisions. Merchants and repair shops must
rely mainly on contract provisions supplemented by some economic power.
 Small users have very limited knowledge and protection. The situation of the small user requires
some elaboration.
With some exceptions, small users do not fully understand the technological nature of the product.
The user does have sensory recognition of some aspects of fitness for use: the bread smells
TABLE 7.1 Contrasting Views: Customer and Suppliers
Principal views
Aspects Of customers Of manufacturers
What is bought A service needed by the Goods made by the manufacturer
Definition of quality Fitness for use during the Conformance to specification on final test
life of the product
Cost Cost of use, including Cost of manufacture
Purchase price
Operating costs
Loss on resale
Responsibility for Over the entire useful life During the warranty period
keeping in service
Spare parts A necessary evil A profitable business
Source: Juran’s Quality Control Handbook, 4th ed., McGraw-Hill, New York, 1988, p. 3.7.
TABLE 7.2 Customer Influences on Quality
Original equipment Dealers and repair
Aspects of the problem manufacturers (OEMs) shops Consumers
Makeup of the market A few, very large Some large customers Very many, very small
customers plus many smaller ones customers
Economic strength of any Very large, cannot be Modest or low Negligible
one customer ignored
Technological strength Very high; has engineers Low or nil Nil (requires technical
of customer and laboratories assistance)
Political strength of Modest or low Low to nil Variable, but can be very
customer great collectively
Fitness for use is judged Qualification testing Absence of consumer Successful usage
mainly by: complaints
Quality specifications Customers Manufacturer Manufacturer
dominated by:
Use of incoming inspection Extensive test for Low or nil for dealers; In-use test
conformance to in-use tests by repair
specification shops
Collection and analysis of Good to fair Poor to nil Poor to nil
failure data
Source: Juran’s Quality Control Handbook, 4th ed., McGraw-Hill, New York, 1988, p. 3.8.
fresh-baked, the radio set has clear reception, the shoes are good-looking. Beyond such sensory
judgments, and especially concerning the long-life performance of the product, the small user must
rely mainly on prior personal experience with the supplier or merchant. Lacking such prior experience,
the small user must choose from the propaganda of competing suppliers plus other available
inputs (neighbors, merchants, independent laboratories, and so on).
To the extent that the user does understand fitness for use, the effect on the supplier’s income is
somewhat as follows:
As seen by the user, the product or service is The resulting income to the supplier is
Not fit for use None, or in immediate jeopardy
Fit for use but noticeably inferior to Low due to loss of market share or need to
competitive products lower prices
Fit for use and competitive At market prices
Noticeably superior to competitive products High due to premium prices or greater share of
In the foregoing, the terms “fitness for use,” “inferior,” “competitive,” and “superior” all relate to
the situation as seen by the user. (The foregoing table is valid as applied to both large customers and
small users.)
Stated Needs and Real Needs. Customers state their needs as they see them, and in their
language. Suppliers are faced with understanding the real needs behind the stated needs and translating
those needs into suppliers’ language.
It is quite common for customers to state their needs in the form of goods, when their real needs
are for the services provided by those goods. For example:
Stated needs Real needs
Food Nourishment, pleasant taste
Automobile Transportation, safety, comfort, etc.
Color TV Entertainment, news, etc.
Toothpaste Clean teeth, sweet breath, etc.
Preoccupation with selling goods can divert attention from the real needs of customers.
Two hair net manufacturers were in competition. They devoted much effort to improving the qualities
of the product and to strengthening their marketing techniques. But hair nets became extinct when someone
developed a hair spray which gave the user a better way of providing the basic service—holding her
hair in place. (Private communication to J. M. Juran.)
In a classic, widely read paper, “Marketing Myopia,” Levitt (1960), stressed service orientation
as distinguished from product orientation. In his view, the railroads missed an opportunity for expansion
due to focus on railroading rather than on transportation. In like manner, the motion picture
industry missed an opportunity to participate in the growing television industry due to focus on
movies rather than on entertainment. (Levitt 1960.)
To understand the real needs of customers requires answers to questions such as: Why are you
buying this product? What service do you expect from it?
Psychological Needs. For many products, customer needs extend beyond the technological
features of the product; the needs also include matters of a psychological nature. Such needs apply
to both goods and services.
A man in need of a haircut has the option of going to (1) a “shop” inhabited by “barbers” or (2)
a “salon” inhabited by “hair stylists.” Either way, he is shorn by a skilled artisan. Either way, his
resulting outward appearance is essentially the same. What differs is his remaining assets and his
sense of well-being.
What applies to services also applies to physical goods. There are factories in which chocolatecoated
candies are conveyed by a belt to the packaging department. At the end of the belt are two
teams of packers. One team packs the chocolates into modest cardboard boxes destined for budgetpriced
merchant shops. The other team packs the chocolates into satin-lined wooden boxes destined
to be sold in deluxe shops. The resulting price for a like amount of chocolate can differ by
severalfold. The respective purchasers encounter other differences as well: the shop decor, level
of courtesy, promptness of service, sense of importance, and so on. However, the goods are identical.
Any chocolate on that conveyer belt has not the faintest idea of whether it will end up in a
budget shop or in a deluxe shop.
Technologists may wonder why consumers are willing to pay such price premiums when the
goods are identical. However, to many consumers, the psychological needs are perceived as real
needs, and the consumers act on their perceptions. Most suppliers design their marketing strategies
to respond to customers’ perceived needs.
“User-Friendly” Needs. The “amateur” status of many users has given rise to the term “user
friendly” to describe a condition which enables amateurs to use technological and other complex
products with confidence. For example:
The language of published information should be simple, nonambiguous, and readily understood.
Notorious offenders have included legal documents, owners’ operating manuals, forms to be
filled out, and so on. Widely used forms (such as Federal tax returns) should be field tested on a
sample of the very people who will later be faced with filling out the forms.
Products should be broadly compatible. Much of this has been done through standardization
committees or through natural monopolies. An example of lack of such compatibility during the
1980s was the personal computer—many personal computers were able to “talk” to computers
made by the same manufacturer but not to computers made by other manufacturers.
The Need to Be Kept Informed. Customers sometimes find themselves in a state of uncertainty:
Their train is late, and they don’t know when to expect it; there is a power outage, and they
don’t know when power will be restored. In many such cases, the supplier company has not established
the policies and processes needed to keep customers informed. In actuality, the customers,
even if kept informed, usually have no choice but to wait it out. Nevertheless, being kept informed
reduces the anxiety—it provides a degree of assurance that human beings are aware of the problem
and that it is in the process of being solved.
The New York subway system rules require conductors to explain all delays lasting two minutes
or more. One survey reported that this rule was followed only about 40 percent of the time. A
City Hall report concluded that “shortage of information is a significant source of public antagonism
toward the Transit Authority” (Levine 1987).
In contrast, some airlines go to pains to keep their customers informed of the reasons for a delay
and of the progress being made in providing a remedy.
A different category of cases involves companies secretly taking actions adverse to quality but
without informing the customer. The most frequent are those in which products not conforming to
specification are shipped to unwary customers. In the great majority of such cases, the products are
fit for use despite the nonconformances. In other cases, the matter may be debatable. In still other
cases, the act of shipment is at the least unethical and at the worst illegal.
In a highly publicized case, Oldsmobile cars were being delivered containing Chevrolet engines.
Yet the Oldsmobile sales promotion had emphasized the quality of its engines. In due course the
manufacturer made restitution but not before suffering adverse publicity.
Once discovered, any secretive actions tend to arouse suspicions, even if the product is fit for customer
use. The customers wonder, “What else has been done secretly without our being informed?”
The usual reason for not informing the customer is a failure to raise the question: What shall we
tell the customers? It would help if every nonconformance document included a blank space headed
“What is to be communicated to the customers?” The decision may be to communicate nothing, but
at least the question has been faced.
Cultural Needs. The needs of customers, especially internal customers, include cultural
needs—preservation of status, continuity of habit patterns, and still other elements of what is broadly
called the cultural pattern. Some of the inability to discover customer needs is traceable to failure to
understand the nature and even the existence of the cultural pattern.
Cultural needs are seldom stated openly—mostly they are stated in disguised form. A proposed
change which may reduce the status of some employee will be resisted by that employee. The stated
reasons for the resistance will be on plausible grounds, such as the effect on costs. The real reason
will not emerge. No one will say, “I am against this because it will reduce my status.” Discovery of
the real needs behind the stated needs is an important step toward a meeting of the minds.
(For elaboration on the nature of cultural patterns and the “rules of the road,” see Section 5, The
Quality Improvement Process, under Resistance to Change; see also Juran 1964, Chapter 9.)
Needs Traceable to Unintended Use. Many quality failures arise because the customer
uses the product in a manner different from that intended by the supplier. This practice takes many
Untrained workers are assigned to processes requiring trained workers.
Equipment is overloaded or is allowed to run without adherence to maintenance schedules.
The product is used in ways never intended by the supplier.
All this influences the relationship between quality and income. The critical question is whether
the quality planning should be based on intended use or actual use. The latter often requires adding
a factor of safety during the planning. For example:
Fuses and circuit breakers are designed into electrical circuits for protection against overloads.
Software is written to detect spelling errors.
Public utility invoicing may include a check of customers’ prior usage to guard against errors in
reading the meters.
Such factors of safety may add to the cost. Yet they may well result in an optimal overall cost by
helping to avoid the higher cost arising from actual use or misuse.
When products fail, a new set of customer needs arises—how to restore service and get compensated
for the associated losses and inconvenience. These new needs are communicated through customer
complaints, which then are acted on by special departments such as Customer Service.
Inadequate company response to consumer complaints and to the terms of warranties has contributed
importantly to the rise of the “consumerism” movement. (See Section 35, Quality and
Society, under The Growth of Consumerism.)
Studies of how to respond to customer complaints have identified the key features of a response
system which meets customer needs. (For elaboration, see Section 25, Customer Service; see also,
United States Office of Consumer Affairs, 1985–86.)
Complaints also affect product salability. This has been researched in studies commissioned by
the United States Office of Consumer Affairs. The findings may be summarized as follows:
Of customers who were dissatisfied with products, nearly 70 percent did not complain. The proportions
varied with the type of product involved. The reasons for not complaining included: the
effort to complain was not worth it; the belief that complaining would do no good; lack of knowledge
of how to complain.
Over 40 percent of the complaining customers were unhappy with the responsive action taken by
the suppliers. Here again the percentage varied depending on the type of product involved.
Future salability is strongly influenced by the action taken on complaints. Figure 7.1 shows
broadly the nature of consumer behavior following product dissatisfaction. This strong influence
extends to brand loyalty. Figure 7.2 shows the extent of this influence as applied to “large ticket”
durable goods, financial services, and automobile services, respectively. A similar, strong influence
extends also to product line loyalty.
That same research concluded that an organized approach to complaint handling provides a high
return on investment. The elements of such an organized approach may include:
A response center staffed to provide 24-h access by consumers
A toll-free telephone number
A computerized database
Percent of customers intending to purchase again
Potential loss over $100
Potential loss of $1 to $5
but no
Complaint not
FIGURE 7.1 Consumer behavior after experiencing product dissatisfaction. [Planning for Quality, 2nd ed. (1990),
Juran Institute Inc., Wilton, CT, pp. 4–12.]
 Intent of repurchase
Intent of repurchase
Intent of repurchase
FIGURE 7.2 Consumer loyalty versus complaint resolution.
Large-ticket durable goods; financial services; automotive
services. (Planning for Quality, 2nd ed. (1990),Juran
Institute, Inc., Wilton, CT, pp. 4–14.)
Special training for the personnel who answer the telephones
Active solicitation of complaints to minimize loss of customers in the future
(For added detail, see the full report, United States Office of Consumer Affairs, 1985–86.)
The most simplistic assumption is that customers are completely knowledgeable as to their needs and
that market research can be used to extract this information from them. In practice, customer knowledge
can be quite incomplete. In some cases the customer may be the last person to find out. It is
unlikely that any customer ever expressed the need for a Walkman (a miniature, portable audiotape
player) before such devices came on the market. However, once they became available, many customers
discovered that they needed one.
These gaps in customer knowledge are filled in mainly by the forces of the competitive market
and by the actions of entrepreneurs.
Inadequate Available Products. When available products are perceived as inadequate, a
vacuum waiting to be filled emerges. Human ingenuity then finds ways to fill that vacuum:
The number of licensed New York taxicabs has remained frozen for years. The resulting shortage
has been filled by unlicensed cabs, limousines, and so on.
Government instructions for filling out tax forms have been confusing to many taxpayers. One
result has been the publication of some best-selling books on how to prepare tax returns.
The service provided by tradesmen has been widely regarded as expensive and untimely. One
result has been the growth of a large do-it-yourself industry.
Relief from Onerous Chores. There seems to be no end to the willingness of affluent people
to pay someone else to do onerous chores. Much former kitchen work is now being done in factories
(soluble coffee, canned foods, and more). The prices of the processed foods are often several
times the prices of the raw foods. Yet to do the processing at home involves working for a very low
hourly wage. Cleaning chores have been extensively transferred to household appliances. The end is
not in sight. The same kinds of transfer have taken place on a massive scale with respect to industrial
chores (data processing, materials handling, etc.)
Reduction of Time for Service. Some cultures exhibit an urge to “get it over with.” In such
cultures, those who can serve customers in the shortest time are rewarded by a higher share of market.
A spectacular example of this urge is the growth of the “fast food” industry. In other industries,
a major factor in choosing suppliers is the time spent to get service. An example is choice of gasoline
filling stations. [See Ackoff (1978), Fable 5.4, p. 108.] This same need for prompt service is an
essential element in the urge to go to “just-in-time” manufacture.
Changes in Customer Habits. Customer habits can be notoriously fickle. Obvious examples
are fashions in clothing and concerns over health that have reduced the consumption of beef and
increased that of poultry. Such shifts are not limited to consumers. Industrial companies often launch
“drives,” most of which briefly take center stage and then fade away. The associated “buzz words”
similarly come and go.
Role of the Entrepreneur. The entrepreneur plays a vital role in providing customers with
new versions of existing products. In addition, the entrepreneur identifies new products, some of them
unheard of, which might create customer needs where none have existed previously. Those new products
have a shocking rate of mortality, but the rewards can be remarkably high, and that is what attracts
the independent entrepreneur. Moreover, the entrepreneurs can make use of the power of advertising
and promotion, which some do very effectively. The legendary Charles Revson, founder of Revlon,
stated it somewhat as follows: “In our factory we make lipstick; in our advertising we sell hope.”
Discovery of customer needs is critical to generating income; it is also one of the major steps on the
“quality planning road map.” That road map includes other steps which, in varying degrees, influence
the relationship between quality and income. (For elaboration, see Section 3, The Quality
Planning Process.)
There is general awareness that product price bears some rational relationship to product quality.
However, researchers on the subject have often reported confused relationships, some of which
appear to run contrary to logical reasoning. To interpret this research, it is useful to separate the subject
into consumer products and industrial products.
Consumer Products. Numerous researchers have tried to quantify the correlation between
product quality and product price. (See, for example, Riesz 1979; also Morris and Bronson 1969.) A
major database for this research has been the journal Consumer Reports, a publication of Consumers
Union, a nonprofit supplier of information and advice to consumers. The specific information used
in the research consisted of Consumer Reports’ published quality ratings of products, along with the
associated prevailing market prices.
The research generally concluded that there is little positive correlation between quality ratings and
market prices. For a significant minority of products, the correlation was negative. Such conclusions
were reached as to foods, both convenience and nonconvenience (Riesz 1979). Similar conclusions were
reached for other consumable products, household appliances, tools, and other long-life products
(Morris and Bronson 1969).
Researchers offer various theories to explain why so many consumers seem to be acting contrary
to their own best interests:
 The quality ratings are based solely on evaluations of the functional features of the products—the
inherent quality of design. The ratings do not evaluate various factors which are known to influence
consumer behavior. These factors include: service in such forms as attention, courtesy,
promptness; also decor in such forms as pleasant surroundings, attractive packaging.
 Consumers generally possess only limited technological literacy, and most are unaware of the
quality ratings.
 Lacking objective quality information, consumers give weight to the image projected by manufacturers
and merchants through their promotion and advertising.
 The price itself is perceived by many consumers as a quality rating. There appears to be a widespread
belief that a higher-priced product is also a higher-quality product. Some companies have exploited
this belief as a part of their marketing and pricing strategy (“Pricing of Products Is Still an Art” 1981).
Price Differences. Premium-priced products usually run about 10 to 20 percent higher than
other products. For example, branded products often are priced in this range relative to generic products.
However, there are many instances of much greater price differences.
Haircuts given in some “salons” sell at several times the price prevailing in “barber shops.”
Chocolates packaged in elegant boxes and sold in deluxe shops may sell for several times the
price of the identical chocolates packaged in simple boxes and sold in budget shops.
The spectrum of restaurant meal prices exceeds an order of magnitude.
Branded pharmaceuticals may sell for several times the price of generic drugs which are asserted
to be therapeutically equivalent.
What emerges is that for many consumers, perception of the quality-price relationship is derived
from unique interpretations of the terms used:
Quality is interpreted as including factors which go beyond the functional features of the product.
Price is interpreted as relating to “value” and is paid for those added factors, along with the inherent
functional features.
Price premiums are conspicuous, and are often resisted fiercely by buyers, even when there
are clear quality differences. In contrast, buyers are usually willing to reward superior quality
with higher share of market in lieu of price differences. In many cases, the supplier can gain more
from higher share of market than from price premiums, because of the arithmetic of the breakeven
chart. (See Figure 7.3, below.) For an interesting case example involving the risks of price
premiums based on superior quality, see Smith (1995).
Efforts to Quantify Value. Efforts to quantify value have largely been limited to functional
properties of products. For example, the U.S. Department of Transportation evaluated the perfor-
FIGURE 7.3 Break-even chart. (Juran’s Quality Control Handbook, 4th ed., McGraw-
Hill, New York, pp. 3.13.)
mance of automobile tires for several qualities, notably tread wear. That made it possible to estimate
the cost per unit of distance traveled (under standard test conditions). (See “Consumer Guide on
Tires” 1980.) Consumers Union sometimes makes such evaluations for consumer products. (See, for
example, “Dishwashing Liquids” 1984.)
Such evaluations can be useful to consumers. However, the impact of such information is limited
because of consumer unawareness and because consumer perceptions are based on broader concepts
of quality and value.
Industrial Products. Industrial products also employ the concepts of quality, price, and value.
However, industrial buyers are generally much better informed of the significance of these concepts.
In addition, industrial buyers are better provided with the technological and economic information
needed to make rational decisions.
The principle of “standard product, market price” can be difficult to apply due to product quality
A company making standard power tools improved the reliability of the tools, but the marketing
manager resisted increasing the price on the ground that since they were standard tools he would
lose share of market if he raised prices. A field study then disclosed that the high-reliability tools
greatly reduced the costs of the (industrial) users in maintenance and especially in downtime.
This information then became the means of convincing users to accept a price increase (of
$500,000 per year). (Consulting experience of J. M. Juran.)
Commodity versus Specialty or System. An important question in much industrial buying
is whether the product being bought is a commodity or something broader. The greater breadth
may involve a specialty or a system of which the commodity is a part, but which includes other
attributes of special value to the buyer.
Commodities are typically bought at market prices, and the price strongly influences the purchasing
decisions. However, a perceived quality superiority is nevertheless an asset which may be
translated into higher share of market or into a price premium. Many companies have opted for price
premiums despite the fact that customers resist accepting price premiums more strongly than awarding
higher market share.
The report entitled Pricing High Quality Products (PIMS 1978) raises questions concerning this
strategy. According to the report, the market is willing to pay premium prices for high-quality products.
However, if the premium price is not demanded, the market responds by awarding so high an
increase in market share that the supplier ends up with a return on investment greater than that resulting
solely from premium pricing.
Perceived quality superiority takes many forms: predictable uniformity of product, promptness of
delivery, technological advice and service, assistance in training customer personnel, prompt assistance
in troubleshooting, product innovation, sharing of information, joint quality planning, and joint
projects for quality improvement. For a case example of a joint quality improvement project involving
Aluminum Company of America and Eastman Kodak, see Kegarise and Miller (1986). (See also
Kegarise et al. 1987.)
Specialties; the “Bundled” Price. Specialties are standard products which are tailored
specifically for use by specific customers. The product is “special” because of added nonstandard
features and services which become the basis for “bundled” prices. Bundled prices provide no breakdown
of price between the goods (commodities) and the associated additional features and services.
Bundled prices are an advantage to the supplier as long as the product remains a specialty and
requires the added features and services. However, if wide use of the specialty results in standardization,
the need for the added services diminishes. In such cases it is common for competitors to
offer the standard product at lower prices but without the technical services. This is a form of
“unbundling” the price. (For an interesting research on pricing in the chemicals industry, along with
an approach to evaluation of the “additional attributes,” see Gross 1978.)
Growth in share of market is among the highest goals of upper managers. Greater market share
means higher sales volume. In turn, higher sales volume accelerates return on investment disproportionally
due to the workings of the break-even chart (Figure 7.3).
In Figure 7.3, to the right of the break-even line, an increase of 20 percent in sales creates an
increase of 50 percent in profit, since the “constant” costs do not increase. (Actually, constant costs
do vary with volume, but not at all in proportion.) The risks involved in increasing market share are
modest, since the technology, production facilities, market, and so on are already in existence and of
proved effectiveness.
Effect of Quality Superiority. Quality superiority can often be translated into higher share
of market, but it may require special effort to do so. Much depends on the nature and degree of superiority
and especially on the ability of the buyer to perceive the difference and its significance.
Quality Superiority Obvious to the Buyer. In such cases, the obvious superiority can be translated
into higher share of market. This concept is fully understood by marketers, and they have from time
immemorial urged product developers to come up with product features which can then be propagandized
to secure higher share of market. Examples of such cases are legion.
Quality Superiority Translatable into Users’ Economics. Some products are outwardly “alike”
but have unlike performances. An obvious example is the electric power consumption of an appliance.
In this and similar examples, it is feasible to translate the technological difference into the language
of money. Such translation makes it easier for amateurs in technology to understand the
significance of the quality superiority.
The power tool case (above) realized the same effect. The superior reliability was translated into
the language of money to secure a price premium. It could instead have been used to secure higher
share of market. In the tire wear case (above) there was a translation into cost per unit of distance
The initiative to translate may also be taken by the buyer. Some users of grinding wheels keep
records on wheel life. This is then translated into money—grinding wheel costs per 1000 pieces
processed. Such a unit of measure makes it unnecessary for the buyer to become expert in the technology
of abrasives.
Collectively, cases such as the above can be generalized as follows:
There is in fact a quality difference among competing products.
This difference is technological in nature so that its significance is not understood by many users.
It is often possible to translate the difference into the language of money or into other forms within
the users’ systems of values.
Quality Superiority Minor but Demonstrable. In some cases, quality superiority can secure added
share of market even though the “inferior” product is fit for use.
A manufacturer of antifriction bearings refined his processes to such an extent that his products
were clearly more precise than those of his competitors. However, competitors’ products were fit
for use, so no price differential was feasible. Nevertheless, the fact of greater precision impressed
the clients’ engineers and secured increased share of market. (Consulting experience of J. M.
In consumer products, even a seemingly small product difference may be translated into
increased market share if the consumers are adequately sensitized.
A manufacturer of candy-coated chocolates seized on the fact that his product did not create
chocolate smudge marks on consumers’ hands. He dramatized this in television advertisements
by contrasting the appearance of children’s hands after eating his and competitors’ (uncoated)
chocolate. His share of market rose dramatically.
Quality Superiority Accepted on Faith. Consumers can be persuaded to accept, on faith, assertions
of product superiority which they themselves are unable to verify. An example was an ingenious
market research on electric razors. The sponsoring company (Schick) employed an
independent laboratory to conduct the tests. During the research, panelists shaved themselves
twice, using two electric razors one after the other. On one day the Schick razor was used first and
a competing razor immediately after. On the next day the sequence was reversed. In all tests the
contents of the second razor were weighed precisely. The data assertedly showed that when the Schick
was the second razor, its contents weighed more than those of competitors. The implication was
that Schick razors gave a cleaner shave. Within a few months the Schick share of market rose as
September 8.3 percent
December 16.4 percent
In this case, the consumers had no way to verify the accuracy of the asserted superiority. They had
the choice of accepting it on faith, or not. Many accepted it on faith.
No Quality Superiority. If there is no demonstrable quality superiority, then share of market is
determined by marketing skills. These take such forms as persuasive propaganda, attractive packaging,
and so on. Price reductions in various forms can provide increases in share of market, but this
is usually temporary. Competitors move promptly to take similar action.
Consumer Preference and Share of Market. Consumers rely heavily on their own senses
to aid them in judging quality. This fact has stimulated research to design means for measuring quality
by using human senses as measuring instruments. This research has led to development of objective
methods for measuring consumer preference and other forms of consumer response. A large body of literature
is now available, setting out the types of sensory tests and the methods for conducting them. (For
elaboration, see Section 23, Inspection and Test, under Sensory Tests.)
At first these methods were applied to making process control and product acceptance decisions.
More recently the applications have been extended into areas such as consumer preference testing,
new product development, advertising, and marketing.
For some products it is easy to secure a measure of consumer preference through “forced choice”
testing. For example, a table is set up in a department store and passers-by are invited to taste two
cups of coffee, A and B, and to express their preference. Pairs of swatches of carpet may be shown
to panels of potential buyers with the request that they indicate their preferences. For comparatively
simple consumer products, such tests can secure good data on consumer preference.
The value of consumer preference data is greatly multiplied through correlation with data on
share of market. Figure 7.4 shows such a correlation for 41 different packaged consumer food products.
This was an uncommonly useful analysis and deserves careful study.
Each dot on Figure 7.4 represents a food product sold on supermarket shelves. Each product has
competitors for the available shelf space. The competing products sell for identical prices and are
packaged in identically sized boxes containing identical amounts of product. What may influence the
consumer are
 The contents of the package, as judged by senses and usage, which may cause the consumer to prefer
product A over product B.
 The marketing features such as attractiveness of the package, appeal of prior advertising, and reputation
of the manufacturer.
On Figure 7.4 the horizontal scale shows consumer preference over the leading competitor as
determined by statistically sound preference testing. The vertical scale shows the share of market
versus the leading competitor, considering the two as constituting 100 percent.
In Figure 7.4 no product showed a consumer preference below 25 percent or above 75 percent.
Such preference levels would mean that the product is so superior (or inferior) that three users out
of four can detect the difference. Since all other factors are essentially equal, a product which is so
overwhelmingly preferred takes over the entire market, and its competition disappears.
In contrast to the vacant areas on the horizontal scale of consumer preference, the vertical scale
of share of market has data along the entire spectrum. One product (marked A on Figure 7.4) lies
squarely on the 50 percent consumer preference line, which probably means (under forced-choice
testing) that the users are guessing as to whether they prefer that product or its competitor. Yet product
A has only 10 percent share of market, and its competitor 90 percent. Not only that, this inequality
in share of market has persisted for years. The reason is that the 90 percent company was the first to
bring that product to market. As a result it acquired a “prior franchise” and has retained its position
through good promotion.
The conclusion is that when competing products are quite similar in consumer preference, any effect
of such small quality differentials is obscured by the effect of the marketing skills. In consequence, it is
logical to conclude that when quality preferences are clearly evident to the user, such quality differences
are decisive in share of market, all other things being equal. When quality differences are slight, the decisive
factor in share of market is the marketing skills.
As a corollary, it appears that companies are well advised to undertake quality improvements
which will result in either (1) bringing them from a clearly weak to an acceptable preference or (2)
bringing them from an acceptable preference to a clearly dominant preference. However, companies
are not well advised to undertake quality improvements which will merely move them from one
acceptable level to another, since the dominant role in share of market in such cases is played by the
marketing skills. [For elaboration, see Juran (1959)].
It is easy for technologists to conclude that what they regard as important in the product is also
of prime concern to the user. In the carpet industry, the engineers devote much effort to improving
FIGURE 7.4 Consumer preference versus share of market.
(Juran’s Quality Control Handbook, 4th ed., McGraw-Hill, New
York, p. 3.15.)
wear qualities and other technological aspects of fitness for use. However, a market survey established
that consumers’ reasons for selecting carpets were primarily sensory:
Color 56 percent
Pattern 20 percent
Other sensory qualities 6 percent
Nonsensory qualities 18 percent
For more complex consumer products it is feasible, in theory, to study the relation of quality to
market share by securing quantitative data on (1) actual changes in buying patterns of consumers and
(2) actions of suppliers which may have created these changes. In practice, such information is difficult
to acquire. It is also difficult to conclude, in any one instance, why the purchase was of model
A rather than B. What does emerge are “demographic” patterns, i.e., age of buyers, size of family,
and so on, which favor model A rather than B. (For elaboration, see Section 18, Market Research and
Marketing.) For products sold through merchants, broad consumer dissatisfaction with quality can
translate into “merchant preference,” with extensive damage to share of market.
A maker of household appliances was competitive with respect to product features, price, and
promptness of delivery. However, it was not competitive with respect to field failure, and this
became a major source of complaints from consumers to the merchants. Within several years the
maker (B) lost all of its leadership in share of market, as shown in the table below. This table stimulated
the upper managers of company B to take action to improve product reliability.
Leaders in market share during:
Model price Base year Base year plus 1 Base year plus 2 Base year plus 3
High A C C C
Medium B B C C
Low C C C C
Special B B B C
Industrial Products and Share of Market. Industrial products are sold more on technological
performance than on sensory qualities. However, the principle of customer preference
applies, as does the need to relate quality differences to customer preference and to share of market.
The methodology is discussed in Section 18, Market Research and Marketing.
Quality and Competitive Bidding. Many industrial products and, especially, large systems,
are bought through competitive bidding. Most government agencies are required by law to
secure competitive bids before awarding large contracts. Industrial companies require their purchasing
managers to do the same. The invitations to bid usually include the parameter of quality, which
may be specified in detail or though performance specifications.
To prospective suppliers the ratio of awards received to bids made is of great significance. The
volume of sales and profit depends importantly on this ratio. In addition, the cost of preparing
bids is substantial; for large systems, the cost of bid preparation is itself large. Finally, the ratio
affects the morale of the people involved. (Members of a winning team fight with their competitors;
members of a losing team fight with each other.) It is feasible to analyze the record of prior
bids in order to improve the percent of successful bids. Table 7.3 shows such an analysis involving
20 unsuccessful bids.
To create Table 7.3, a multifunctional team analyzed 20 unsuccessful bids. It identified the main
and contributing reasons for failure to win the contract. The team’s conclusions show that the installation
price was the most influential factor—it was a contributing cause in 10 of the 14 cases which
included bids for installation. This finding resulted in a revision of the process for estimating the
installation price and to an improvement in the bidding/success ratio.
Among marketers there has always been a school of thought which gives quality the greatest weight
among the factors which determine marketability. A survey by Hopkins and Bailey (1971) of 125
senior marketing executives as to their preference for their own product superiority showed the following:
Percent of marketing executives giving first
Form of product superiority preference to this form
Superior quality 40
Lower price (or better value) 17
More features, options, or uses 12
All others 31
Such opinions are supported by the “Profit Impact of Market Strategy” (PIMS) study, (Schoeffler,
Buzzell, and Heany 1974). The PIMS study, involving 521 businesses, undertook (among other things)
to relate (1) quality competitiveness to (2) share of market. The findings can be expressed as follows:
TABLE 7.3 Analysis of Unsuccessful Bids*
Bid not accepted due to
Contract Quality of Product price Installation Reciprocal
proposal design price† buying Other
Totals 7 8 10 (of 14) 4 1
*Contributing reason; main reason
†Only 14 bids were made for installation.
Source: Juran’s Quality Control Handbook, 4th ed., McGraw-Hill, New York, 1988, p. 3.7.
Number of businesses in these zones of share of market
Quality versus competitors Under 12% 12–26% 27%
Inferior 79 58 35
Average 51 63 53
Superior 39 55 88
Total 169 176 176
Building Quality Leadership. Quality leadership is often the result of an original quality
superiority which gains what marketers call a “prior franchise.” Once gained, this franchise can be
maintained through continuing product improvement and effective promotion.
Companies which have attained quality leadership have usually done so on the basis of one of
two principal strategies:
 Let nature take its course. In this approach, companies apply their best efforts, hoping that in time
these efforts will be recognized.
 Help nature out by adopting a positive policy—establish leadership as a formal goal and then set
out to reach that goal.
Those who decide to make quality leadership a formal goal soon find that they must also answer
the question: Leadership in what? Quality leadership can exist in any of the multiple aspects of fitness
for use, but the focus of the company will differ depending on which aspect is chosen.
If quality leadership is to consist of: The company must focus on:
Superior quality of design Product development, systems development
Superior quality of conformance Manufacturing quality controls
Availability Reliability and maintainability programs
Guarantees, field services Customer service capability
Once attained, quality leadership endures until there is clear cumulative evidence that some competitor
has overtaken the leader. Lacking such evidence, the leadership can endure for decades and
even centuries. However, quality leadership can also be lost through some catastrophic change.
A brewery reportedly changed its formulation in an effort to reduce costs. Within several years,
its share of market declined sharply. The original formula was then restored but market share did
not recover. (See “The Perils of Cutting Quality” 1982.)
In some cases, the quality reputation is built not around a specific company but around an association
of companies. In that event, this association adopts and publicizes some mark or symbol. The
quality reputation becomes identified with this mark, and the association goes to great lengths to protect
its quality reputation.
The medieval guilds imposed strict specifications and quality controls on their members. Many
medieval cities imposed “export controls” on selected finished goods in order to protect the quality
reputation of the city (Juran 1995, Chapter 7).
The growth of competition in quality has stimulated the expansion of strategic business planning
to include planning for quality and quality leadership. (For elaboration, see Section 13, Strategic
The “Market Leader” Concept. One approach to quality leadership is through product
development in collaboration with the leading user of such products—a user who is influential in the
market and hence is likely to be followed. For example, in the medical field, an individual is “internationally
renowned; a chairman of several scientific societies; is invited to congresses as speaker or
chairman; writes numerous scientific papers” (Ollson 1986).
Determining who is the leading user requires some analysis. (In some respects the situation is
similar to the marketer’s problem of discovering who within the client company is the most influential
in the decision to buy.) Ollson lists 10 leader types, each playing a different role.
Carryover of Failure-Prone Features. Quality leadership can be lost by perpetuating
failure-prone features of predecessor models. The guilty features are well known, since the resulting
field failures keep the field service force busy restoring service. Nevertheless, there has been
much carryover of failure-prone features into new models. At the least, such carryover perpetuates
a sales detriment and a cost burden. At its worst, it is a cancer which can destroy seemingly healthy
product lines.
A notorious example was the original xerographic copier. In that case the “top 10” list of field
failure modes remained essentially identical, model after model. A similar phenomenon existed
for years in the automobile industry.
The reasons behind this carryover have much in common with the chronic internal wastes which
abound in so many companies:
 The alarm signals are disconnected. When wastes go on, year after year, the accountants incorporate
them into the budgets. That disconnects the alarm signals—no alarms ring as long as actual
waste does not exceed budgeted waste.
 There is no clear responsibility to get rid of the wastes. There are other reasons as well. The technologists
have the capability to eliminate much of the carryover. However, those technologists are
usually under intense pressure from the marketers to develop new product and process features in
order to increase sales. In addition, they share a distaste for spending their time cleaning up old
problems. In their culture, the greatest prestige comes from developing the new.
The surprising result can be that each department is carrying out its assigned responsibilities, and
yet the product line is dying. Seemingly nothing short of upper management intervention—setting
goals for getting rid of the carryover—can break up the impasse.
At the highest levels of management, and among boards of directors, there is keen interest in financial
measures such as net income and share prices on the stock markets. It is known that quality influences
these measures, but so do other variables. Separating out the effect of quality has as yet not
been feasible other than through broad correlation studies.
During the early 1990s, some of the financial press published articles questioning the merits
of the Malcolm Baldrige National Quality Award, Total Quality Management (TQM), and other
quality initiatives. These articles were challenged, and one result was analysis of the stock price
performance of Baldrige Award winners compared with that of industrial companies generally.
The results were striking. From the dates of receiving the Award, the stock price of the Baldrige
winners had advanced 89 percent, as compared with 33 percent for the broad Standard & Poor’s
index of 500 stocks (Business Week 1993, p. 8.)
In 1991 the General Accounting Office (GAO) published the results of a study of 20 “finalist”
applicants for the Baldrige Award (companies which were site-visited). The report concluded that “In
nearly all cases, companies that used total quality management practices achieved better employee
relations, higher productivity, greater customer satisfaction, increased market share, and improved
profitability” (General Accounting Office 1991).
In its simplest form, a sales contract sets out an agreed price for a specific product (goods or services),
e.g., X cents for a glass of milk; Y dollars for a bus ticket. For such consumable products, the
purchase price is also the cost of using the product. Drinking the milk or riding the bus normally
involves no added cost for the user, beyond the original purchase price. Expressed in equation form,
Purchase price  cost of use
For long-life products, this simple equation is no longer valid. Purchase price expands to include
such factors as cost of capital invested, installation cost, and deductions for resale value. Cost of use
expands to include costs of operation and maintenance. It is true even for “simple” consumer products.
For some articles of clothing, the cumulative costs of cleaning and maintenance can exceed the
original purchase price.
The famous comedian Ed Wynn is said to have worn the same $3.50 shoes throughout his long
career. They cost him $3000 in repairs.
Concept of the Optimum. The basic concept of life-cycle costing is one of finding the
optimum—finding that set of conditions which (1) meets the needs of both supplier and customer
and (2) minimizes their combined costs. (Life cycle cost is only one of several names given to this
concept of an optimum. Other names include: cost of ownership, cost of use or usage, mission
cost, lifetime cost.)
The life cycle cost concept is widely applicable, but application has lagged. The concept can be
defined in models which identify the factors to be considered, the data to be acquired, and the equations
to be used in arriving at the optimum. The lag in application is not due to difficulty in setting
up the models. Instead, the lag is due to inadequacies in acquiring the needed data, and especially to
cultural resistance. (See below.)
Steps in Life Cycle Cost Analysis. The literature has gone far to organize life cycle cost
analyses. The steps set out below represent a typical organized approach. For elaboration on various
organized approaches, see: Brook and Barasia (1977); Ebenfelt and Ogren (1974); Stokes and Stehle
(1968); Toohey and Calvo (1980); Wynholds and Skratt (1977).
Identify the Life Cycle Phases. Optimizing requires striking a balance among numerous costs,
some of which are antagonistic to others. The starting point is to identify the phases or activities
through which the product goes during its life cycle. These phases are mapped out in a flow diagram
as an aid to the team doing the analysis. Typical phases include: product research; product development;
product design; manufacturing planning; production; installation; provision of spares; operation;
maintenance; support services; modifications; disposal.
Identify the Cost Elements. The next step is to identify the cost elements for each phase. For
example, operating costs for civilian aircraft include: maintenance labor and material, spares holding,
delay/flight interruptions, administrative, insurance, training, flight operation, crew, aircraft and
traffic servicing, fuel and oil (Rose and Phelps 1979). For an example from the Tennessee Valley
Authority, see Duhan and Catlin 1973.
Acquire the Cost Data. This step can be a formidable obstacle. The prevailing accounting systems
provide only part of the essential cost information. The rest must be acquired by special study—by
estimate or by enlarging the accounting system. The work involved can be reduced by concentrating
on the vital few cost categories—those which involve most of the money. Attention must also be
given to those categories which are highly sensitive, i.e., they are leveraged to respond to small
changes in other factors—the “cost drivers.”
Analyze the Relationships. This step quantifies the interrelationship among the cost factors. For
example, a comparatively simple analysis establishes that for automotive vehicles, tire wear correlates
mainly with distance traveled and speed of travel. For aircraft, tire wear correlates mainly with
number of landings and takeoffs.
However, many analyses are far more complex. A common example is the relationship of (1)
designed mean time between failures (MTBF) and mean time to repair (MTTR) to (2) the subsequent
costs of operation and maintenance Repair and Maintenance (R&M). For some products (such as
certain military categories) repair and maintenance costs over the life of the product run to multiples
of the original purchase price. These R&M costs are highly sensitive to the designed MTBF and
MTTR. Efforts to quantify the interrelationship run into complex estimates bounded by a wide range
of error. For a case example involving military avionics, see Toohey and Calvo (1980).
Formulate Aids to Decision Making. The purpose of these analyses is to aid decision making.
Typically the decision maker first establishes which categories of cost are to be included in the decisionmaking
process. Then, on the basis of the analysis, equations are set up to arrive at the life cycle cost
in terms of those same established cost categories. For example, the state of Virginia arrived at the
following equation for estimating cost per hour for a certain class of highway machinery:
Cost per hour equals initial price, plus repair parts, plus foregone interest, less resale value, all
divided by operating hours (Doom 1969).
Breadth of Application. Ideally the life cycle cost analysis should provide aid to making
strategic decisions on optimizing costs. In practice this is feasible only for simple products or for
problems of limited scope: state government purchase of room air conditioners (Doom 1969); optimum
inventory levels (Dushman 1970); repair level strategy, i.e., discard, base repair, or depot repair
(Henderson 1979); effect of test system requirements of operation and support costs (Gleason 1981);
optimization of number of thermal cycles (Shumaker and DuBuisson 1976).
Probably the widest application has been in the area of industrial products. (See below, under
Application to Industrial Products.)
Irrespective of the area of application, most investigators have concluded that the decisions which
determine life cycle cost are concentrated in the early stages of the product life cycle. Figure 7.5 is
a typical model (Bj?rklund 1981).
Figure 7.5 shows that life cycle cost is determined mainly by decisions made during the very
early phases of the life cycle. Such a concentration makes clear the need for providing the product
researchers, developers, and designers with a good database on the subsequent costs of production,
installation, operation, and maintenance.
Application to Consumer Products. In a classic study, Gryna (1970) found that for various
household appliances and television sets, the ratio of life cycle costs to original price ranged
from 1.9 to 4.8. (See Table 7.4.)
A study, Consumer Appliances: The Real Cost (M.I.T. 1974), found the following proportions of
life cycle costs to prevail during the year 1972 for color TV sets and household refrigerators:
Elements of life cycle costs Color TV sets Refrigerators
Purchase price 53 36
Power 12 58
Service 35 6
Total 100 100
Lund (1978) provides some supplemental information based on a follow-up study.
Fody (1977) reported on how a U.S. government agency made its first application of the life cycle
cost concept to procurement of room air conditioners. The suppliers made their bids on the basis of
original price. However, the agency considered in addition the expected electric power cost based on
certified energy efficiency ratings. The basis for awarding the contracts then became the lowest life
cycle costs rather than the lowest bid price.
Life cycle costs for automobiles have been studied in depth. Table 7.5 (Federal Highway
Administration 1984) shows life cycle costs for intermediate size cars driven 120,000 mi (192,000
km) in 12 years.
Although data on life cycle costs of consumer products have become increasingly available, consumer
use of such data has lagged. The major reasons include:
Percent of life cycle costs
Concept &
Design & development
Percentage of LCC
tied up by decisions
Use &
FIGURE 7.5 Phases affecting life cycle cost adapted from Bj?rklund, 1981, p. 3. (Juran’s Quality Control
Handbook, 4th ed., McGraw-Hill, New York, pp. 3.23.)
TABLE 7.4 Life cycle Costs: Consumer Products
Cost of Ratio to (life
operation plus cycle cost to
Product Original price, $ maintenance, $ Total cost, $ original price)
Room air conditioner 200 465 665 3.3
Dishwasher 245 372 617 2.5
Freezer 165 628 793 4.8
Range, electric 175 591 766 4.4
Range, gas 180 150 330 1.9
Refrigerator 230 561 791 3.5
TV (black and white) 200 305 505 2.5
TV (color) 560 526 1086 1.9
Washing machine 235 617 852 3.6
Source: Juran’s Quality Control Handbook, 4th ed., McGraw-Hill, New York, p. 3.23.
1. Cultural resistance (see below)
2. The economics of administering numerous small long-life contracts
3. The complexities created by multiple ownership
The most notable example of multiple ownership is passenger automobiles. In the United States,
these often go through multiple ownership before being scrapped. Even under short-term warrantees,
transfer of ownership creates problems of administering warrantee contracts. Existing practice usually
imposes a charge for such transfer between successive owners. For contracts over the useful life
of the product, this problem becomes considerably more complicated.
Application to Industrial Products. Application to industrial products has probably been
the area of greatest progress. A major example is seen in the airlines’ evolution of life cycle costing
strategy for aircraft maintenance. A critical element was the creation of an adequate database relative
to field operation and maintenance. Data analysis then resulted in a change in grand strategy for
maintenance, from the overhaul concept to the concept of on-condition maintenance. In addition, the
data analysis resulted in a superior feedback to product designers and manufacturers. For an uncommonly
well-documented explanation, see Nowlan and Heap (1978).
Part of the competition to sell industrial equipment consists of convincing prospective buyers that
their operating and maintenance costs will be low. In some cases this conviction is created by guaranteeing
the operating costs or by offering low-cost maintenance contracts. Some manufacturers provide
record-keeping aids to enable users to accumulate data on competitive products as an aid to
future purchasing decisions. Some industrial users build up data banks on cost of downtime for various
types of industrial equipment as an input to future decision making.
The approach to making decisions to acquire capital equipment follows generally the steps set
out above under the heading Steps in Life-Cycle Cost Analysis. Kaufman (1969) gives an explanation
of methodology along with case examples of application.
TABLE 7.5 Life-Cycle Costs, Automobiles
Original price $10,320
Additional “ownership” costs
Accessories 198
Registration 240
Titling 516
Insurance 6,691
Scheduled maintenance 1,169
Nonoperating taxes 33
Subtotal $8,847
Operation and maintenance costs
Gasoline $6,651
Unscheduled maintenance 4,254
Tires 638
Oil 161
Gasoline tax, federal 514
Gasoline tax, other 771
Sales taxes 130
Parking, tolls 1,129
Subtotal $14,248
Grand total $33,415
Source: Juran’s Quality Control Handbook, 4th ed.,
McGraw-Hill, New York, p. 3.24.
Application to Defense Industries. During the twentieth century many governments greatly
expanded their acquisition of military weaponry, both in volume and in complexity. Mostly the governments
acquired these weapons by purchase rather than by expansion of government arsenals and
shipyards. It was most desirable that the life cycle cost concept be applied to such weapons. However,
a major obstacle was the deeply rooted practice of buying on the basis of the lowest bid price.
Starting in about the 1960s the U.S. Department of Defense organizations stepped up their efforts
to make the life cycle cost concept effective in procurement contracts. Directives were issued to
define the new emphasis and to clear away old obstacles. However, as events unfolded, it became
evident that to apply the concept to government procurement was more difficult than for comparable
situations in civilian procurement. The differences have their origin in such factors as the nature
of the respective missions, the system of priorities, the organization for decision making and the
extent of public scrutiny. [For a more detailed discussion of these differences, see Gansler (1974),
Pedrick (1968), and Bryan (1981).]
The urge for applying the life cycle concept to military products has stimulated an extensive literature.
Most of the published papers relate to division of the subject and to the structure of models.
[See, for example, Barasia and Kiang (1978), Peratino (1968), and Ryan (1968).]
There are also numerous papers on application. These are mainly directed at subsystems, e.g.,
optimizing inventory levels. Alternatively, the applications are directed at lower-level components. A
published example relates to standardization of electronic modules (Laskin and Smithhisler 1979).
Another example deals with standardization of test equipment (Rosenberg and Witt 1976). [See also
Eustis (1977) and Gallagher and Knobloch (1971).] Application to subsystems or lower-level components
obviously runs the risk of suboptimizing unless care is taken to examine the impact of any
proposed change on related subsystems or components.
Cultural Resistance. Cultural resistance is a major force holding back the application of the
life cycle cost concept. Purchase based on original price has dominated commercial practice for
thousands of years. The skills, habit patterns, and status of many persons—product designers, purchasing
managers, marketers—have long been built around the original purchase price concept.
Changing to life cycle costing demands a change in habit patterns, with associated risks of damage
to long-standing skills and status.
The most deeply rooted habits are probably those of consumers—small buyers for personal use.
They keep few records on costs of operation and maintenance and tend to underestimate the
amounts. For less-than-affluent consumers, the purchase of a costly product is obscured by the fact
that they may lack the capital needed even for the original price and hence must borrow part of it. In
addition, the laws of sales are well worked out as applied to original price contracts but are still
in evolution as applied to life cycle cost contracts.
Obviously, makers of consumer goods cannot abandon marketing on original price when such is
the cultural pattern. What they can do is to experiment by offering some optional models designed
for lower cost of usage as a means of gaining experience and time for the day when life cycle costing
comes into wider use.
Makers of industrial products also face cultural resistance in trying to use life cycle costing as a
business opportunity. However, with good data they can make a persuasive case and strike responsive
chords in buyers who see in these data a way to further the interests of their companies and
Contracts Based on Amount of Use. An alternative approach to life cycle costing is
through sales contracts which are based on the amount of use. Such contracts shift all the life cycle
costs to the supplier, who then tends to redesign the system in a way which optimizes the cost of providing
The public utilities—e.g., telephone, power—are long-standing examples. These utilities neither
sell a product nor do they often even lease a product; they sell only the service (e.g., watt-hours of
electricity, message units of telephone service). In such cases, the ownership of the equipment
remains with the utility, which also has the responsibility for keeping the equipment maintained and
repaired. As a result, the income of the utility is directly bound up with keeping the equipment in
service. There are numerous other instances; e.g., the rental car is often rented based on the actual
mileage driven; laundromat machines are rented on the basis of hours of use.
Sale of goods can sometimes be converted into a sale of use. It is common practice for vehicle
fleets to “buy” tires based on mileage. Airlines buy engines based on hours of use. There is much
opportunity for innovation in the use of this concept.
For consumer products, the metering of actual use adds many complications. Common practice
is therefore to use elapsed time as an approximation of amount of use.
The human being exhibits an instinctive drive for precision, beauty, and perfection. When unrestrained
by economics, this drive has created the art treasures of the ages. In the arts and in esthetics,
this timeless human instinct still prevails.
In the industrial society, there are many situations in which this urge for perfection coincides with
human needs. In food and drug preparation, certain organisms must be completely eliminated or they
will multiply and create health hazards. Nuclear reactors, underground mines, aircraft, and other
structures susceptible to catastrophic destruction of life require a determined pursuit of perfection to
minimize dangers to human safety. So does the mass production of hazardous products.
However, there are numerous other situations in which the pursuit of perfection is antagonistic to
society, since it consumes materials and energy without adding to fitness for use, either technologically
or esthetically. This wasteful activity is termed “perfectionism” because it adds cost without
adding value.
Perfectionism in Quality of Design. This is often called “overdesign.” Common examples
Long-life designs for products which will become obsolete before they wear out.
Costly finishes on nonvisible surfaces.
Tolerances or features added beyond the needs of fitness for use. (The military budget reviewers
call this “gold-plating.”)
Some cases of overdesign are not simple matters of yes or no. For example, in television reception
there are “fringe areas” which give poor reception with conventional circuit design. For such areas, supplemental
circuitry is needed to attain good quality of image. However, this extra circuitry is for many
areas an overdesign and a waste. The alternative of designing an attachment to be used only in fringe
areas creates other problems, since these attachments must be installed under nonfactory conditions.
Overdesign can also take place in the areas of reliability and maintainability. Examples include:
“Worst case” designs that guard against failures resulting from a highly unlikely combination of
adverse conditions. Such designs can be justified in critical situations but seldom otherwise.
Use of unduly large factors of safety.
Use of heavy-duty or high-precision components for products which will be subjected to conventional
Defense against overdesign is best done during design review, when the design is still fluid. The
design review team commonly includes members from the functions of production, marketing, use,
and customer service. Such a team can estimate the economic effects of the design. It can then challenge
those design features which do not contribute to fitness for use and which therefore will add
costs without adding value. Some forms of design review classify the characteristics, e.g., essential,
desirable, unessential. The unessential then become prime candidates for removal.
Perfectionism in Quality of Conformance. Typical examples include:
Insistence on conformance to specification despite long-standing successful use of nonconforming
Setting appearance standards at levels beyond those sensed by users
One defense against this type of perfectionism is to separate two decisions which are often confused:
(1) the decision on whether product conforms to specification and (2) the decision on whether
nonconforming product is fit for use. Decision 1 may be relegated to the bottom of the hierarchy.
Decision 2 should be made only by people who have knowledge of the conditions of use.
A further defense is to quantify the costs and then shift the burden of proof.
A marketing head insisted on overfilling the product packages beyond the label weight. The production
head computed the cost of the overfill and then challenged the marketer to prove that this
cost would be recovered through added sales.
The Perfectionists. Those who advocate perfectionism often do so with the best intentions
and always for reasons which seem logical to them. The resulting proposals are nevertheless of no
benefit to users for one of several common reasons:
The added perfection has no value to the user. (The advocate is not aware of this.)
The added perfection has value to the user but not enough to make up for the added cost. (The
advocate is unaware of the extent of the costs involved.)
The added perfection is proposed not to benefit the user but to protect the personal position of the
advocate who has some functional interest in perfection but no responsibility for the associated
The weaknesses of such proposals all relate back to costs: ignorance of the costs; no responsibility
for the costs; indifference to costs due to preoccupation with something else. Those who do have
responsibility for the costs should quantify them and then dramatize the results in order to provide
the best challenge.
This section has drawn extensively from the following:
Juran, J. M. (1991). Juran on Quality by Design. The Free Press, a division of Macmillan, New York.
Juran, J. M., ed. (1995). A History of Managing for Quality. Sponsored by Juran Foundation, Inc. Quality Press,
Leadership for the Quality Century: A Reference Guide, 6th ed. (1996). Juran Institute, Inc., Wilton, CT.
Planning for Quality, 2nd ed. (1990). Juran Institute, Inc., Wilton, CT.
The author is grateful to the copyright holders for permission to quote from these works.
Ackoff, Russell L. (1978). The Art of Problem Solving. John Wiley & Sons, New York.
Barasia, R. K. and Kiang, T. D. (1978). “Development of a Life Cycle Management Cost Model.” Proceedings,
Annual Reliability and Maintainability Symposium. IEEE, New York, pp. 254–260.
Bj?rklund, O. (1981). “Life Cycle Costs; An Analysis Tool for Design, Marketing, and Service.” Q-Bulletin of
Brook, Cyril, and Barasia, Ramesh (1977). “A Support System Life Cycle Cost Model.” Proceedings, Annual
Reliability and Maintainability Symposium. IEEE, New York, pp. 297–302.
Bryan, N. S. (1981). “Contracting for Life Cycle Cost to Improve System Affordability.” Proceedings, Annual
Reliability and Maintainability Symposium. IEEE, New York, pp. 342–345.
Business Week (1993). “Betting to Win on the Baldie Winners.” October 18.
Consumer Appliances: The Real Cost (1974). Prepared jointly by the Massachusetts Institute of Technology
Center for Policy Alternatives with the Charles Stark Draper Laboratory, Inc. Sponsored by the National Science
Foundation. Available from RANN Document Center, National Science Foundation, Washington, D.C. 20550.
“Consumer Guide on Tires Is Issued by U.S. Agency” (1980). New York Times, December 31, p. 6.
“Dishwashing Liquids” (1984). Consumer Reports, July, pp. 412–414.
Doom, I. F. (1969). “Total Cost Purchasing Applied to Heavy Equipment Procurement.” Virginia Highway
Research Council, Charlottesville.
Duhan, Stanley, and Catlin, John C., Sr. (1973). “Total Life Cost and the ‘Ilities’.” Proceedings, Annual
Reliability and Maintainability Symposium. IEEE, New York, pp. 491–495.
Dushman, Allan (1970). “Effect of Reliability on Life Cycle Inventory Cost.” Proceedings, Annual Reliability
and Maintainability Symposium. IEEE, New York, pp. 549–561.
Ebenfelt, Hans, and Ogren, Stig (1974). “Some Experiences from the Use of an LCC Approach.” Proceedings,
Annual Reliability and Maintainability Symposium. IEEE, New York, pp. 142–146.
Eustis, G. E. (1977). “Reduced Support Costs for Shipboard Electronic Systems.” Proceedings, Annual
Reliability and Maintainability Symposium. IEEE, New York, pp. 316–319.
Federal Highway Administration, Office of Highway Planning, Highway Statistics Division (1984). “Cost of
Owning and Operating Automobiles and Vans.” Washington, D.C.
Fody, Theodore, J. (1977). “The Procurement of Window Air Conditioners Using Life-Cycle Costing.”
Proceedings, Annual Reliability and Maintainability Symposium. IEEE, New York, pp. 81–88.
Gallagher, B. M., and Knobloch, W. H. (1971). “Helicopter Auxiliary Power Unit Cost of Ownership.”
Proceedings, Annual Reliability and Maintainability Symposium. IEEE, New York, pp. 285–291.
Gansler, J. S. (1974). “Application of Life-Cycle Costing to the DOD System Acquisition Decision Process.”
Proceedings, Annual Reliability and Maintainability Symposium. IEEE, New York, pp. 147–148.
General Accounting Office (1991). “Management Practices: U.S. Companies Improve Performance Through
Quality Efforts.” Washington, DC.
Gleason, Daniel (1981). “The Cost of Test System Requirements.” Proceedings, Annual Reliability and
Maintainability Symposium. IEEE, New York, pp. 108–113.
Gross, Irwin (1978). “Insights from Pricing Research.” In Pricing Practices and Strategies. E. L. Bailey (ed.).
The Conference Board, New York, pp. 34–39.
Gryna, F. M., Jr. (1970). “User Costs of Poor Product Quality.” (Doctoral dissertation, University of Iowa, Iowa
City.) For additional data and discussion, see Gryna, F. M., Jr. (1977). “Quality Costs: User vs. Manufacturer.”
Quality Progress, June, pp. 10–13.
Henderson, J. T., P. E. (1979). “A Computerized LCC/ORLA Methodology.” Proceedings Annual Reliability and
Maintainability Symposium. IEEE, New York, pp. 51–55.
Hopkins, David S., and Bailey, Earl (1971). “New Product Pressures.” The Conference Record, June, pp. 16–24.
Juran, J. M. (1959). “A Note on Economics of Quality.” Industrial Quality Control, February, pp. 20–23.
Juran, J. M. (1964, 1995). Managerial Breakthrough (first published 1964). McGraw-Hill, New York.
Juran, J. M. (1970). “Consumerism and Product Quality.” Quality Progress, July, pp. 18–27.
Juran, J. M. (1992). Juran on Quality by Design. The Free Press, a division of Macmillan, New York.
Juran, J. M. (ed.) (1995). A History of Managing for Quality. Quality Press, Milwaukee.
Kaufman, R. J. (1969). “Life-Cycle Costing: Decision-Making Tool for Capital Equipment Acquisitions.”
Journal of Purchasing, August, pp. 16–31.
Kegarise, R. J., and Miller, G. D. (1986). “An Alcoa-Kodak Joint Team.” The Juran Report, no. 6, pp. 29–34.
Kegarise, R. J., Heil, M., Miller, G. D., and Miller, G. (1987). “A Supplier/Purchaser Project: From Fear to Trust.”
The Juran Report, no. 8, pp. 248–252.
Laskin, R., and Smithhisler, W. L. (1979). “The Economics of Standard Electronic Packaging.” Proceedings,
Annual Reliability and Maintainability Symposium. IEEE, New York, pp. 67–72.
Levine, Richard (1987). “Breaking Routine: Voice of the Subway.” New York Times, January 15.
Levitt, Theodore (1960). “Marketing Myopia.” Harvard Business Review, July-August, pp. 26–28ff.
Lund, Robert T. (1978). “Life Cycle Costing: A Business and Societal Instrument.” Management Review, April,
pp. 17–23.
M.I.T. (1974). “The Real Cost.” Massachusetts Institute of Technology Study. Cambridge, MA.
Morris, Ruby Turner, and Bronson, Claire Sekulski (1969). “The Chaos of Competition Indicated by Consumer
Reports.” Journal of Marketing, July, pp. 26–34.
Nowlan, F. S., and Heap, H. F. (1978). Reliability Centered Maintenance. Document ADA066579. Defense
Documentation Center, Alexandria, VA.
Ollson, John Ryding (1986). “The Market-Leader Method; User-Oriented Development.” Proceedings 30th
EOQC Annual Conference, pp. 59–68.
Pedrick, P. C. (1968). “Survey of Life-Cycle Costing Practices of Non-Defense Industry.” Proceedings, Annual
Reliability and Maintainability Symposium. IEEE, New York, pp. 188–192.
Peratino, G. S. (1968). “Air Force Approach to Life Cycle Costing.” Proceedings, Annual Reliability and
Maintainability Symposium. IEEE, New York, pp. 184–187.
“The Perils of Cutting Quality.” (1982). New York Times, August 22.
PIMS (1978). Pricing High-Quality Products. The Strategic Planning Institute, Cambridge, MA.
“Pricing of Products Is Still an Art.” (1981). Wall Street Journal, November 25, pp. 25, 33.
Riesz, Peter C. (1979). “Price-Quality Correlations for Packaged Food Products.” Journal of Consumer Affairs.
Rose, John, and Phelps, E. L. (1979). “Cost of Ownership Application to Airplane Design.” Proceedings, Annual
Reliability and Maintainability Symposium, IEEE, New York, pp. 47–50.
Rosenberg, H., and Witt, J. H. (1976). “Effects on LCC of Test Equipment Standardization.” Proceedings, Annual
Reliability and Maintainability Symposium. IEEE, New York, pp. 287–292.
Ryan, W. J. (1968). “Procurement Views of Life-Cycle Costing.” Proceedings, Annual Reliability and
Maintainability Symposium. IEEE, New York, pp. 164–168.
Schoeffler, Sidney, Buzzell, Robert D., and Heany, Donald F. (1974). “Impact of Strategic Planning on Profit
Performance.” Harvard Business Review, March-April, pp. 137–145.
Shumaker, M. J., and DuBuisson, J. C. (1976). “Tradeoff of Thermal Cycling vs. Life Cycle Costs.” Proceedings,
Annual Reliability and Maintainability Symposium. IEEE, New York, pp. 300–305.
Smith, Lee (1995). “Rubbermaid Goes Thump.” Fortune, October 2, pp. 90–104.
Stokes, R. G., and Stehle, F. N. (1968). “Some Life-Cycle Cost Estimates for Electronic Equipment: Methods
and Results.” Proceedings, Annual Reliability and Maintainability Symposium, IEEE, New York, pp. 169–183.
Toohey, Edward F., and Calvo, Alberto B. (1980). “Cost Analyses for Avionics Acquisition.” Proceedings, Annual
Reliability and Maintainability Symposium. IEEE, New York, pp. 85–90.
United States Office of Consumer Affairs (1985–86). “Consumer Complaint Handling in America; An Update
Study.” Washington, DC, 20201.
Wynholds, Hans W., and Skratt, John P. (1977). “Weapon System Parametric Life-Cycle Cost Analysis.”
Proceedings, Annual Reliability and Maintainability Symposium. IEEE, New York, pp. 303–309.
Frank M. Gryna
Lessons Learned 8.2
Objectives of Evaluation 8.3
Internal Failure Costs 8.4
External Failure Costs 8.6
Appraisal Costs 8.7
Prevention Costs 8.7
Example from Manufacturing Sector 8.8
Example from Service Sector 8.9
Finalizing the Definitions 8.10
International Standards and Quality
Costs 8.11
Sequence of Events 8.12
Data Collection 8.13
Roles to Support Improvement 8.20
Optimum Cost of Poor Quality 8.20
Summarizing the Data 8.23
Bases for Comparison 8.23
Reporting the Results 8.24
This section discusses how quality has an impact on the costs of goods and services in an organization.
Section 7, Quality and Income, addresses the issue of quality and sales revenue. Thus,
the two sections provide a framework of how quality is related to the total financial picture of an
We identify and measure the costs associated with poor quality for three reasons: to quantify the
size of the quality problem to help justify an improvement effort, to guide the development of that
effort, and to track progress in improvement activities. Among the concepts and methodologies covered
are traditional categories of quality costs, a broadened concept of categories including lost revenue
and process capability costs, activity-based costing, data collection methods, return on quality,
presentation of findings, gaining approval for an improvement effort, using cost data to support continuous
improvement, optimum quality level, and reporting cost data. The underlying theme in the
section is the use of quality-related costs to support a quality improvement effort rather than as a system
of reporting quality costs.
We will follow the convention of using the term “product” to denote goods or services.
During the 1950s there evolved numerous quality-oriented staff departments. The heads of these new
departments were faced with “selling” their activities to the company managers. Because the main
language of those managers was money, the concept of studying quality-related costs provided the
vocabulary to communicate between the quality staff departments and the company managers.
Over the decades, as the staff quality specialists extended their studies, some surprises emerged:
1. The quality-related costs were much larger than had been shown in the accounting reports. For
most companies, these costs ran in the range of 10 to 30 percent of sales or 25 to 40 percent of
operating expenses. Some of these costs were visible, some of them were hidden.
2. The quality costs were not simply the result of factory operation, the support operations were also
major contributors.
3. The bulk of the costs were the result of poor quality. Such costs had been buried in the standards,
but they were in fact avoidable.
4. While these quality costs were avoidable, there was no clear responsibility for action to reduce
them, neither was there any structured approach for doing so.
Quality specialists used the data to help justify quality improvement proposals and to track the
cost data over time.
Those early decades of experience led to some useful lessons learned.
Lessons Learned. These lessons, discussed below, can help us to formulate objectives for
tracking and analyzing the impact of quality on costs.
The Language of Money Is Essential. Money is the basic language of upper management. Despite
the prevalence of estimates, the figures provide upper managers with information showing the overall
size of the quality costs, their prevalence in areas beyond manufacture, and the major areas for
potential improvement.
Without the quality cost figures, the communication of such information to upper managers is
slower and less effective.
The Meaning of “Quality Costs.” The term “quality costs” has different meanings to different people.
Some equate “quality costs” with the costs of poor quality (mainly the costs of finding and correcting
defective work); others equate the term with the costs to attain quality; still others use the
term to mean the costs of running the Quality department. In this handbook, the term “quality costs”
means the cost of poor quality.
Quality Cost Measurement and Publication Does Not Solve Quality Problems. Some organizations
evaluate the cost of poor quality and publish it in the form of a scoreboard in the belief that
publication alone will stimulate the responsible managers to take action to reduce the costs. These
efforts have failed. The realities are that publication alone is not enough. It makes no provision to
identify projects, establish clear responsibilities, provide resources to diagnose and remove causes of
problems, or take other essential steps. New organization machinery is needed to attack and reduce
the high costs of poor quality (see, generally, Section 5, The Quality Improvement Process).
Scoreboards, if properly designed, can be a healthy stimulus to competition among departments,
plants, and divisions. To work effectively, the scoreboard must be supplemented by a structured
improvement program. In addition, scoreboards must be designed to take into account inherent differences
in operations among various organizational units. Otherwise, comparisons made will
become a source of friction.
Scope of Quality Costs Is Too Limited. Traditionally, the measurement of quality cost focuses
on the cost of nonconformities, i.e., defects in the goods or services delivered to external and
internal customers. These are often called external and internal failure costs. An important cost
that is not measured is lost sales due to poor quality (this is called a “hidden cost” because it is
not easily measured). Another omitted cost is the extra cost in processes that were producing conforming
output but which are inefficient. These inefficiencies are due to excess product or
process variability (even though within specification limits) or inefficiencies due to redundant or
non-value-added process steps.
Traditional Categories of Quality Costs Have Had a Remarkable Longevity. About 1945, a pioneering
effort proposed that quality-related costs be assigned to one of three categories: failure costs,
appraisal costs, and prevention costs. The pioneers emphasized that these categories were not the
only way to organize quality costs; the important point was to obtain a credible estimate of the total
quality cost. But many practitioners found the categories useful and even found ingenious ways to
adapt the categories to special applications such as engineering design.
The experience that led to these lessons learned also included some changes in the quality movement:
1. An explosion in the acceptance of the concept of continuous improvement in all sectors—profit,
nonprofit, and public.
2. Progress in understanding and in quantifying the impact of quality on sales revenue.
3. Emphasis on examining cross-functional processes to reduce errors and cycle time and
improve process capability to increase customer satisfaction. These analyses confirm the benefits
of (a) diagnosis of causes to reduce errors and (b) process analysis to identify redundant
work steps and other forms of non-value-added activities.
From the lessons learned and the changes in the quality movement, we can identify some objectives
for evaluating quality and costs.
Objectives of Evaluation. The primary objectives are
1. Quantify the size of the quality problem in language that will have an impact on upper management.
The language of money improves communication between middle managers and upper
managers. Some managers say: “We don’t need to spend time to translate the defects into dollars.
We realize that quality is important, and we already know what the major problems are.” Typically
when the study is made, these managers are surprised by two results. First, the quality costs turn out
to be much higher than had been realized. In many industries they are in excess of 20 percent of
sales. Second, while the distribution of the quality costs confirms some of the known problem areas,
it also reveals other problem areas that had not previously been recognized.
2. Identify major opportunities for reduction in cost of poor quality throughout all activities
in an organization. Costs of poor quality do not exist as a homogeneous mass. Instead, they occur
in specific segments, each traceable to some specific cause. These segments are unequal in size,
and a relative few of the segments account for the bulk of the costs. A major byproduct of evaluation
of costs of poor quality is identification of these vital few segments. This results in setting
priorities to assure the effective use of resources. We need to collect data on the cost of poor quality,
analyze the data, and plan an improvement strategy that attacks chunks of the glacier rather
than ice chips.
3. Identify opportunities for reducing customer dissatisfaction and associated threats to sales revenues.
Some costs of poor quality are the result of customer dissatisfaction with the goods or service
provided. This dissatisfaction results in a loss of current customers—“customer defections”—and an
inability to attract new customers (for elaboration see Section 18 under Linking Customer
Satisfaction Results to Customer Loyalty and to Processes). Addressing the areas of dissatisfaction
helps to improve retention of current customers and create new customers.
4. Provide a means of measuring the result of quality improvement activities instituted to achieve
the opportunities in 2 and 3 above. Measuring progress helps to keep a focus on improvement and
also spotlights conditions that require removal of obstacles to improvements.
5. Align quality goals with organization goals. Measuring the cost of poor quality is one of four
key inputs for assessing the current status of quality (the others are market standing on quality relative
to competition, the organization quality culture, and the activities composing the quality system).
Knowing the cost of poor quality (and the other elements) leads to the development of a
strategic quality plan consistent with overall organization goals.
Collectively, these objectives strive to increase the value of product and process output and
enhance customer satisfaction. This section uses the framework shown in Figure 8.1. Note that this
framework extends the traditional concept of quality costs to reflect not only the costs of nonconformities
but also process inefficiencies and the impact of quality on sales revenue. Sometimes, the
term “economics of quality” is employed to describe the broader concept and differentiate it from
the traditional concept of “quality cost.”
We must emphasize the main objective in collecting this data, i.e., to energize and support quality
improvement activities. This is summarized in Figure 8.2. The term “cost of quality” used in this
figure includes the prevention, appraisal, and failure categories which are discussed below.
Many companies summarize these costs into four categories. Some practitioners also call these categories
the “cost of quality.” These categories and examples of typical subcategories are discussed
Internal Failure Costs. These are costs of deficiencies discovered before delivery which are
associated with the failure (nonconformities) to meet explicit requirements or implicit needs of external
or internal customers. Also included are avoidable process losses and inefficiencies that occur
even when requirements and needs are met. These are costs that would disappear if no deficiencies
Failure to Meet Customer Requirements and Needs. Examples of subcategories are costs associated
Scrap: The labor, material, and (usually) overhead on defective product that cannot economically
be repaired. The titles are numerous—scrap, spoilage, defectives, etc.
Rework: Correcting defectives in physical products or errors in service products.
Lost or missing information: Retrieving information that should have been supplied.
Failure analysis: Analyzing nonconforming goods or services to determine causes.
Scrap and rework—supplier: Scrap and rework due to nonconforming product received from
suppliers. This also includes the costs to the buyer of resolving supplier quality problems.
FIGURE 8.1 Components of the cost of poor quality.
Cost of
Cost of
Cost of poor quality
Cost of
lost opportunities
for sales revenue
One hundred percent sorting inspection: Finding defective units in product lots which contain
unacceptably high levels of defectives.
Reinspection, retest: Reinspection and retest of products that have undergone rework or other
Changing processes: Modifying manufacturing or service processes to correct deficiencies.
Redesign of hardware: Changing designs of hardware to correct deficiencies.
Redesign of software: Changing designs of software to correct deficiencies.
For organizations For processes
Identification Process
1. Identify Cost-of-Quality work
2. Collect Cost-of-Quality data
Use Cost-of-Quality information
as input to develop/enhance quality
improvement plans, establish Costof-
Quality reduction objectives, and
implement those plans
Establish the need
for and make the
selection of
processes for
Identify and prioritize
opportunities for
the selected
3. Analyze Cost-of-Quality information
to motivate and prioritize process
improvement projects/activities
FIGURE 8.2 Cost of quality and quality improvement. (AT&T 1990, p. 16.)
Scrapping of obsolete product: Disposing of products that have been superseded.
Scrap in support operations: Defective items in indirect operations.
Rework in internal support operations: Correcting defective items in indirect operations.
Downgrading: The difference between the normal selling price and the reduced price due to
quality reasons.
Cost of Inefficient Processes. Examples of subcategories are
Variability of product characteristics: Losses that occur even with conforming product (e.g.,
overfill of packages due to variability of filling and measuring equipment).
Unplanned downtime of equipment: Loss of capacity of equipment due to failures.
Inventory shrinkage: Loss due to the difference between actual and recorded inventory
Variation of process characteristics from “best practice”: Losses due to cycle time and costs
of processes as compared to best practices in providing the same output. The best-practice
process may be internal or external to the organization.
Non-value-added activities: Redundant operations, sorting inspections, and other non-valueadded
activities. A value-added activity increases the usefulness of a product to the customer; a
non-value-added activity does not. (The concept is similar to the 1950s concept of value engineering
and value analysis.)
For elaboration on variation and the cost of quality, see Reeve (1991). For a discussion of waste
in “white collar” processes, see Quevedo (1991).
External Failure Costs. These are costs associated with deficiencies that are found after product
is received by the customer. Also included are lost opportunities for sales revenue. These costs
also would disappear if there were no deficiencies.
Failure to Meet Customer Requirements and Needs. Examples of subcategories are
Warranty charges: The costs involved in replacing or making repairs to products that are still
within the warranty period.
Complaint adjustment: The costs of investigation and adjustment of justified complaints attributable
to defective product or installation.
Returned material: The costs associated with receipt and replacement of defective product
received from the field.
Allowances: The costs of concessions made to customers due to substandard products accepted
by the customer as is or to conforming product that does not meet customer needs.
Penalties due to poor quality: This applies to goods or services delivered or to internal processes
such as late payment of an invoice resulting in a lost discount for paying on time.
Rework on support operations: Correcting errors on billing and other external processes.
Revenue losses in support operations: An example is the failure to collect on receivables from
some customers.
Lost Opportunities for Sales Revenue. Examples are
Customer defections: Profit margin on current revenue lost due to customers who switch for
reasons of quality. An important example of this category is current contracts that are canceled
due to quality.
New customers lost because of quality: Profit on potential customers lost because of poor
New customers lost because of lack of capability to meet customer needs: Profit on potential
revenue lost because of inadequate processes to meet customer needs.
Appraisal Costs. These are the costs incurred to determine the degree of conformance to quality
requirements. Examples are
Incoming inspection and test: Determining the quality of purchased product, whether by
inspection on receipt, by inspection at the source, or by surveillance.
In-process inspection and test: In-process evaluation of conformance to requirements.
Final inspection and test: Evaluation of conformance to requirements for product acceptance.
Document review: Examination of paperwork to be sent to customer.
Balancing: Examination of various accounts to assure internal consistency.
Product quality audits: Performing quality audits on in-process or finished products.
Maintaining accuracy of test equipment: Keeping measuring instruments and equipment in calibration.
Inspection and test materials and services: Materials and supplies in inspection and test work
(e.g., x-ray film) and services (e.g., electric power) where significant.
Evaluation of stocks: Testing products in field storage or in stock to evaluate degradation.
In collecting appraisal costs, what is decisive is the kind of work done and not the department
name (the work may be done by chemists in the laboratory, by sorters in Operations, by testers in
Inspection, or by an external firm engaged for the purpose of testing). Also note that industries use
a variety of terms for “appraisal,” e.g., checking, balancing, reconciliation, review.
Prevention Costs. These are costs incurred to keep failure and appraisal costs to a minimum.
Examples are
Quality planning: This includes the broad array of activities which collectively create the overall
quality plan and the numerous specialized plans. It includes also the preparation of procedures
needed to communicate these plans to all concerned.
New-products review: Reliability engineering and other quality-related activities associated
with the launching of new design.
Process planning: Process capability studies, inspection planning, and other activities associated
with the manufacturing and service processes.
Process control: In-process inspection and test to determine the status of the process (rather
than for product acceptance).
Quality audits: Evaluating the execution of activities in the overall quality plan.
Supplier quality evaluation: Evaluating supplier quality activities prior to supplier selection,
auditing the activities during the contract, and associated effort with suppliers.
Training: Preparing and conducting quality-related training programs. As in the case of
appraisal costs, some of this work may be done by personnel who are not on the payroll of the
Quality department. The decisive criterion is again the type of work, not the name of the department
performing the work.
Note that prevention costs are costs of special planning, review, and analysis activities for quality.
Prevention costs do not include basic activities such as product design, process design, process
maintenance, and customer service.
The compilation of prevention costs is initially important because it highlights the small investment
made in prevention activities and suggests the potential for an increase in prevention costs
with the aim of reducing failure costs. The author has often observed that upper management
immediately grasps this point and takes action to initiate an improvement effort. Experience also
suggests, however, that continuing measurement of prevention costs can usually be excluded in
order to (1) focus on the major opportunity, i.e., failure costs, and (2) avoid the time spent discussing
what should be counted as prevention costs.
This part of the section focuses on the question “How much is it costing our organization by not
doing a good job on quality?” Thus we will use the term “cost of poor quality.” Most (but not all) of
the total of the four categories is the cost of poor quality (clearly, prevention costs are not a cost of
poor quality.) Strictly defined, the cost of poor quality is the sum of internal and external failure costs
categories. But this assumes that those elements of appraisal costs—e.g., 100 percent sorting inspection
or review—necessitated by inadequate processes are classified under internal failures. This
emphasis on the cost of poor quality is related to a later focus in the section, i.e., quality improvement,
rather than just quality cost measurement.
A useful reference on definitions, categories, and other aspects is Campanella (1999). For an
exhaustive listing of elements within the four categories see Atkinson, Hamburg, and Ittner (1994).
Winchell (1991) presents a method for defining quality cost terms directly in the language used by
an organization.
Example from Manufacturing Sector. An example of a study for a tire manufacturer is
shown in Table 8.1.
Some conclusions are typical for these studies:
1. The total of almost $900,000 per year is large.
2. Most (79.1 percent) of the total is concentrated in failure costs, specifically in “waste-scrap” and
consumer adjustments.
3. Failure costs are about 5 times the appraisal costs. Failure costs must be attacked first.
4. A small amount (4.3 percent) is spent on prevention.
TABLE 8.1 Annual Quality Cost—Tire manufacturer
1. Cost of quality failures—losses
a. Defective stock $ 3,276 0.37
b. Repairs to product 73,229 8.31
c. Collect scrap 2,288 0.26
d. Waste-scrap 187,428 21.26
e. Consumer adjustments 408,200 46.31
f. Downgrading products 22,838 2.59
g. Customer ill will Not counted
h. Customer policy adjustment Not counted
Total $697,259 79.10%
2. Cost of appraisal
a. Incoming inspection $ 23,655 2.68
b. Inspection 1 32,582 3.70
c. Inspection 2 25,200 2.86
d. Spot-check inspection 65,910 7.37
Total $147,347 16.61%
3. Cost of prevention
a. Local plant quality control engineering $ 7,848 0.89
b. Corporate quality control engineering 30,000 3.40
Total $ 37,848 4.29%
Grand total $882,454 100.00%
5. Some consequences of poor quality could not be quantified, e.g., “customer ill will” and “customer
policy adjustment.” Here, the factors were listed as a reminder of their existence.
As a result of this study, management decided to increase the budget for prevention activities.
Three engineers were assigned to identify and pursue specific quality improvement projects.
Example from Service Sector. Table 8.2 shows a monthly quality cost report for an installment
loan process at one bank. Only those activities that fall into the four categories of quality cost
are shown. The monthly cost of quality of about $13,000 for this one process is equivalent to about
$160,000 per year for this bank. Table 8.2 quantifies loan costs throughout a typical consumer loan
life cycle (including loan payoff). Note that the internal and external failure costs account for about
half of total quality costs. These failure costs—which are preventable—are now tracked and managed
for reduction or elimination so that they do not become unintentionally built into the operating
structure. From the moment the customer contacts the bank with a problem, all costs related to
resolving the problem are external failure costs. (Also note the significant amount of appraisal cost.)
These two examples illustrate studies at the plant level and the process level, but studies can be
conducted at other levels, e.g., corporate, division, plant, department, process, product, component, or
on a specific problem. Studies made at higher levels are typically infrequent, perhaps annual.
Increasingly, studies are conducted as part of quality improvement activities on one process or one
problem and then the frequency is guided by the needs of the improvement effort. For further discussion,
see below under Using Cost of Poor Quality Concept to Support Quality Improvement.
The concept of cost of poor quality applies to a gamut of activities. For examples from manufacturing
see Finnegan and Schottmiller (1990) and O’Neill (1988). In the service sector, useful discussions
are provided as applied to hotels (Bohan and Horney 1991) and for educational testing (Wild and Kovacs
TABLE 8.2 Quality Cost Report—Installment Loans
Operation Prevention Appraisal Internal failure External failure
Making a loan:
Run credit check 0 0 26 0
Process GL tickets and I/L input sheets 0 0 248 0
Review documents 0 3014 8 0
Make document corrections 0 0 1014 0
Follow up on titles, etc. 0 157 0 0
Review all output 0 2244 0 0
Correct rejects and incorrect output 0 0 426 0
Correct incomplete collateral report 0 0 0 78
Work with dealer on problems 0 0 0 2482
I/L system downtime 0 0 520 0
Time spent training on I/L 1366 0 0 0
Loan payment:
Receive and process payments 0 261 784 0
Respond to inquiries when no coupon
is presented with payments 0 0 784 0
Loan payoff:
Process payoff and release document 0 0 13 0
Research payoff problems 0 0 13 0
Total cost of quality (COQ) 1366 5676 3836 2560
COQ as % of total quality cost 10.2 42.2 28.5 19.1
COQ as % of reported salary expense 2.6 10.8 7.3 4.9
Source: Adapted from Aubrey (1988).
1994). Applications have also been made to functional areas. For marketing, see Carr (1992) and Nickell
(1985); for engineering see Schrader (1986); for “white collar” see Hou (1992) and Keaton et al. (1988).
The Conference Board (1989) presents the results of a survey of 111 companies (manufacturing and service)
on current practices in measuring quality costs.
Finalizing the Definitions. Although many organizations have found it useful to divide the
overall cost into the categories of internal failure, external failure, appraisal, and prevention, the
structure may not apply in all cases. Clearly, the practitioner should choose a structure that suits company
need. In defining the cost of poor quality for a given organization, the following points should
be kept in mind.
1. The definitions should be tailor-made for each organization. The usual approach is to review
the literature and select those detailed categories which apply to the organization. The titles used
should meet the needs of the organization, not the literature. This selected list is then discussed with
the various functions to identify additional categories, refine the wording, and decide on broad
groupings, if any, for the costs. The resulting definitions are “right” for the organization.
2. The key categories are the failure cost elements because these provide the major opportunity
for reduction in costs and for removal of the causes of customer dissatisfaction. These costs should
be attacked first. Appraisal costs are also an area for reduction, especially if the causes of the failures
are identified and removed so as to reduce the need for appraisal.
3. Agreement should be reached on the categories of cost to include before any data are collected.
Upper management should be a party to this agreement. Initially, summarized data on scrap and rework
can gain management’s attention and stimulate the need for a full study. Such summaries can be an
impetus for management to become personally involved, for example, by calling and chairing the meetings
to finalize the definition of the cost of poor quality. The quality specialist and the accountant both
have key roles.
4. Certain costs routinely incurred may have been accepted as inevitable but are really part of
the cost of poor quality. Examples are the costs of redesigning the product made necessary by
deficiencies in fitness for use and the costs of changing the manufacturing process because of an
inability to meet product specifications. If the original design and original manufacturing plans had
been adequate, these costs would not have occurred. Typically, these costs have been accepted as
normal operating costs, but should be viewed as opportunities for improvement and subsequent cost
5. As the detailed categories of the cost of poor quality are identified, some categories will be
controversial. Much of the controversy centers around the point: “These are not quality-related costs
but costs that are part of normal operating expenses and therefore should not be included.” Examples
are inclusion of full overhead in calculating scrap costs, preventive maintenance, and loss in morale.
In most companies, the cost of poor quality is a large sum, frequently larger than the company’s
profits. This is true even when the controversial categories are not included, so it is prudent to omit
these categories and avoid the controversy in order to focus attention on the major areas of potential
cost reduction. Some efforts to quantify quality costs have failed because of tenacious insistence by
some specialists that certain controversial categories be included. A useful guide is to ask: “Suppose
all defects disappeared. Would the cost in question also disappear?” A “yes” answer means that the
cost is associated with quality problems and therefore should be included. A “no” answer means that
the category should not be included in the cost of poor quality.
At the minimum, controversial categories should be separated out of the totals so that attention
will be directed to the main issues, i.e., the failure costs.
Hidden Costs. The cost of poor quality may be understated because of costs which are difficult to
estimate. The “hidden” costs occur in both manufacturing and service industries and include:
1. Potential lost sales (see above under External Failure Costs).
2. Costs of redesign of products due to poor quality.
3. Costs of changing processes due to inability to meet quality requirements for products.
4. Costs of software changes due to quality reasons.
5. Costs of downtime of equipment and systems including computer information systems.
6. Costs included in standards because history shows that a certain level of defects is inevitable and
allowances should be included in standards:
a. Extra material purchased: The purchasing buyer orders 6 percent more than the production
quantity needed.
b. Allowances for scrap and rework during production: History shows that 3 percent is “normal”
and accountants have built this into the cost standards. One accountant said, “Our scrap
cost is zero. The production departments are able to stay within the 3 percent that we have
added in the standard cost and therefore the scrap cost is zero.” Ah, for the make-believe
“numbers game.”
c. Allowances in time standards for scrap and rework: One manufacturer allows 9.6 percent
in the time standard for certain operations to cover scrap and rework.
d. Extra process equipment capacity: One manufacturer plans for 5 percent unscheduled
downtime of equipment and provides extra equipment to cover the downtime. In such cases,
the alarm signals ring only when the standard value is exceeded. Even when operating within
those standards, however, the costs should be a part of the cost of poor quality. They represent
opportunities for improvement.
7. Extra indirect costs due to defects and errors. Examples are space charges and inventory charges.
8. Scrap and errors not reported. One example is scrap that is never reported because of fear of
reprisals, or scrap that is charged to a general ledger account without an identification as scrap.
9. Extra process costs due to excessive product variability (even though within specification limits):
For example, a process for filling packages with a dry soap mix meets requirements for
label weight on the contents. The process aim, however, is set above label weight to account
for variability in the filling process. See Cost of Inefficient Processes above under Internal
Failure Costs.
10. Cost of errors made in support operations, e.g., order filling, shipping, customer service, billing.
11. Cost of poor quality within a supplier’s company. Such costs are included in the purchase price.
These hidden costs can accumulate to a large amount—sometimes three or four times the
reported failure cost. Where agreement can be reached to include some of these costs, and where
credible data or estimates are available, then they should be included in the study. Otherwise, they
should be left for future exploration.
Progress has been made in quantifying certain hidden costs, and therefore some of them have been
included in the four categories discussed above. Obvious costs of poor quality are the tip of the iceberg.
Atkinson et al. (1991) trace the evolution of the cost of quality, present research results from four
organizations (manufacturing and service), and explain how cost of quality data is applied in continuous
improvement programs.
International Standards and Quality Costs. The issue of quality costs is addressed in
ISO 9004-1 (1994), Quality Management and Quality System Elements—Guidelines, Section 6,
“Financial Considerations of Quality Systems.” This standard is advisory rather than mandatory.
Three approaches to data collection and reporting are identified (but others are not excluded):
1. Quality costing approach: This is the failure, appraisal, and prevention approach described
2. Process cost approach. This approach collects data for a process rather than a product. All
process costs are divided into cost of conformity and cost of nonconformity. The cost of conformity
includes all costs incurred to meet stated and implied need of customers. Note that this is
the cost incurred when a process is running without failure, i.e., material, labor, and overhead
including prevention and process control activities. This cost includes process inefficiencies. The
cost of nonconformity is the traditional cost of internal and external failures. The focus is to
reduce both the cost of conformity and the cost of nonconformity.
3. Quality loss approach: This approach includes, but goes beyond, internal and external failure
costs. Conceptually it tries to collect data on many of the “hidden” costs such as loss of sales revenue
due to poor quality, process inefficiencies, and losses when a quality characteristic deviates
from a target value even though it is within specification limits. Under this approach the costs can
be estimated by using the Taguchi quality loss function.
For a comparison of these three approaches, see Schottmiller (1996). To provide further guidance,
Technical Committee 176 of the International Organization for Standardization is developing a document,
ISO/CD 10014, Guideline for Managing the Economics of Quality. This document will
address both costs and customer satisfaction and will apply to “for profit” and “not for profit” organizations.
Shepherd (1998) reviews the experiences with quality costs of over 50 organizations that
successfully implemented ISO 9000.
A study of the cost of poor quality is logically made by the accountant, but the usual approach follows
a different scenario. A quality manager learns about the quality cost concept and speaks with
the accountant about making a study. The accountant responds that “the books are not kept that way.”
The accountant does provide numbers on scrap, rework, or certain other categories, but is not persuaded
to define a complete list of categories and collect the data. The quality manager then follows
one of two routes: (1) unilaterally prepares a definition of the categories and collects data or (2) presents
to upper management the limited data provided by the accountant, and recommends that a full
study be made using the resources of Accounting, Quality, and other functions. The second approach
is more likely to achieve acceptance of the results of the study.
Sequence of Events. The following sequence applies to most organizations.
1. Review the literature on quality costs. Consult others in similar industries who have had experience
with applying quality cost concepts.
2. Select one organizational unit of the company to serve as a pilot site. This unit may be one plant,
one large department, one product line, etc.
3. Discuss the objectives of the study with the key people in the organization, particularly those in
the accounting function. Two objectives are paramount: determine the size of the quality problem
and identify specific projects for improvement.
4. Collect whatever cost data are conveniently available from the accounting system and use this
information to gain management support to make a full cost study.
5. Make a proposal to management for a full study. The proposal should provide for a task force of
all concerned parties to identify the work activities that contribute to the cost of poor quality.
Work records, job descriptions, flowcharts, interviews, and brainstorming can be used to identify
the activities.
6. Publish a draft of the categories defining the cost of poor quality. Secure comments and revise.
7. Finalize the definitions and secure management approval.
8. Secure agreement on responsibility for data collection and report preparation.
9. Collect and summarize the data. Ideally, this should be done by Accounting.
10. Present the cost results to management along with the results of a demonstration quality
improvement project (if available). Request authorization to proceed with a broader companywide
program of measuring the costs and pursuing projects. See below under Gaining Approval
for the Quality Improvement Program.
Clearly, the sequence must be tailored for each organization.
The costs associated with poor quality typically span a variety of departments (see Figure 8.3),
and thus it is important to plan for this in data collection.
Data Collection. The initial study collects cost data by several approaches:
1. Established accounts: Examples are appraisal activities conducted by an Inspection department
and warranty expenses to respond to customer problems.
2. Analysis of ingredients of established accounts: For example, suppose an account called “customer
returns” reports the cost of all goods returned. Some of the goods are returned because they
are defective. Costs associated with these are properly categorized as “cost of poor quality.” Other
goods may be returned because the customer is reducing inventory. To distinguish the quality
costs from the others requires a study of the basic return documents.
3. Basic accounting documents: For example, some product inspection is done by Production
department employees. By securing their names and the associated payroll data, we can quantify
these quality costs.
4. Estimates: Input from knowledgeable personnel is clearly important. In addition, several
approaches may be needed.
a. Temporary records: For example, some production workers spend part of their time repairing
defective product. It may be feasible to arrange with their supervisor to create a temporary
record to determine the repair time and thereby the repair cost. This cost can then be projected
for the time period to be covered by the study.
b. Work sampling: Here, random observations of activities are taken and the percent of time
spent in each of a number of predefined categories can then be estimated (see Esterby 1984).
In one approach, employees are asked to record the observation as prevention, appraisal, failure,
or first time work (AT&T 1990, p. 35).
c. Allocation of total resources: For example, in one of the engineering departments, some of
the engineers are engaged part time in making product failure analyses. The department, however,
makes no provision for charging engineering time to multiple accounts. Ask each engineer
to make an estimate of time spent on product failure analysis by keeping a temporary
activity log for several representative weeks. As the time spent is due to a product failure, the
cost is categorized as a failure cost.
d. Unit cost data: Here, the cost of correcting one error is estimated and multiplied by the number
of errors per year. Examples include billing errors and scrap. Note that the unit cost per
error may consist of costs from several departments.
e. Market research data: Lost sales revenue due to poor quality is part of the cost of poor quality.
Although this revenue is difficult to estimate, market research studies on customer satisfaction
and loyalty can provide input data on dissatisfied customers and customer defections.
Cost-of-poor-quality activity
Location in organization
Dept. A Dept. B Dept. C Dept. D Etc.
Discover status of late order  
Correct erroneous bills  
Expedite installation of late shipment  
Troubleshoot failures on installation   
Perform warranty repairs   
Dispose of scrap 
Replace unacceptable installation   
FIGURE 8.3 Costs of poor quality across departments. (Romagnole and Williams 1995.)
Calculations can then be made to estimate the lost revenue. Table 8.3 shows a sample case
from the banking industry. Note that this approach starts with annual revenue and does not
take into account the loss over the duration of years that the customer would have been loyal
to the company. For a more comprehensive calculation, see Section 18 under Linking
Customer Satisfaction Results to Customer Loyalty Analysis and to Processes. Also note that
the calculations in Table 8.3 do not consider the portion of satisfied customers who will be
wooed away by competition.
A special problem is whether all of the defective product has been reported. An approach to estimate
the total amount of scrap is the input-output analysis. A manufacturer of molded plastic parts
followed this approach:
1. Determine (from inventory records) the pounds of raw material placed in the manufacturing
2. Determine the pounds of finished goods shipped. If necessary, convert shipments from units to
3. Calculate the overall loss as step 1 minus step 2.
4. Make subtractions for work in process and finished goods inventory.
The result is the amount of raw material unaccounted for and presumably due to defective product
(or other unknown reasons). A comparison of this result and the recorded amount of defectives
provides a practical check on the recorded number of defectives.
Another special problem is the rare but large cost, e.g., a product liability cost. Such costs can be
handled in two ways: (1) report the cost in a special category separating it from the total for the other
categories or (2) calculate an expected cost by multiplying the probability of occurrence of the
unlikely event by the cost if the event does occur.
A common mistake in data collection is the pursuit of great precision. This is not necessary—it
is a waste of resources and time. We determine the cost of poor quality in order to justify and support
quality improvement activities and identify key problem areas. For that purpose, a precision of
±20 percent is adequate in determining the cost of poor quality.
Briscoe and Gryna (1996) discuss the categories and data collection as applied to small business.
One of the issues in calculating the costs of poor quality is how to handle overhead costs. Three
approaches are used in practice: include total overhead using direct labor or some other base, include
TABLE 8.3 Revenue Lost through Poor Quality
$10,000,000 Annual customer service revenue
1,000 Number of customers
Ч 25% Percent dissatisfied
250 Number dissatisfied
Ч 75% Percent of switchers
(60–90% of dissatisfied)
188 Number of switchers
Ч $10,000 Average revenue per customer
$1,880,000 Revenue lost through poor quality
Source: The University of Tampa (1990).
variable overhead only (the usual approach), or do not include overhead at all. The allocation of overhead
can, of course, have a significant impact on calculating the total cost of poor quality and also
on determining the distribution of the total over the various departments. Activity-based costing
(ABC) can help by providing a realistic allocation of overhead costs.
Traditionally, manufacturing overhead costs are allocated to functional departments and to products
based on direct labor hours, direct labor dollars, or machine hours. This method works fine but
only if the single base used (e.g., direct labor hours) accounts for most of the total operational costs,
as has been true in the past. Times have changed.
During the past 20 years, many manufacturing and service firms have experienced significant
changes in their cost structure and the way their products are made. Direct labor accounted for
about 50 to 60 percent of the total cost of the product or service, with overhead constituting about
20 percent of the cost. As companies became more automated through robotics and other means,
direct labor costs have declined to about 5 to 25 percent and overhead costs have increased to about
75 percent in the total cost mix of the product.
Activity-based costing (ABC) is an accounting method that aims to improve cost-effectiveness
through a focus on key cost elements. In doing this, ABC allocates overhead based on the factors (activities)
which cause overhead cost elements to be incurred. These causal factors—called cost drivers—
are measurable activities that increase overhead costs. ABC refines the way product costs are
determined by using computer technology to economically track overhead costs in smaller categories
with many cost drivers. These cost drivers are analogous to an allocation base such as direct labor hours
in traditional allocation of overhead. But instead of just one cost driver (e.g., direct labor hours) there
are many cost drivers such as machine setups, purchase orders, shipments, maintenance requests, etc.
For each cost driver, an overhead rate is determined by dividing total costs for the driver (e.g., total cost
for all machine setups) by the number of driver events (e.g., number of setups). The results might be
for example, $90 per machine setup, $40 per purchase order, $120 per shipment, $80 per maintenance
request. These overhead rates can then be applied to specific products, thus recognizing that every product
or service does not utilize every component of overhead at exactly the same intensity, or in some
cases, may not even use a given component at all. This more precise allocation of overhead can change
the total cost of poor quality and the distribution over departments, thus influencing the priorities for
improvement efforts.
Activity-based costing is more than an exercise in allocating overhead. This broader viewpoint of
“activity-based cost management” emphasizes improvement in terms of cost reduction to be applied
where it is most needed.
A basic accounting text that explains both activity-based costing and quality costs is Garrison and
Noreen (l994). Other references relating activity-based costing and quality costs are Dawes and Siff
(1993), Hester (1993), and Krause and Gryna (1995).
Improvement requires an investment of resources, and the investment must be justified by the blossoming
benefits of improvement. The long-term effect of applying the cost of poor quality concept
is shown in Figure 8.4. We will call the comparison of benefits to investment the “return on quality”
(ROQ). Thus ROQ is really a return of investment (ROI) in the same sense as other investments such
as equipment or an advertising program.
Using the expanded scope of cost of poor quality (see above under Categories of Quality Costs),
the benefits of an improvement effort involve both reductions in cost and increases in sales revenue.
Some of the issues involved in estimating the benefits are
Reduced cost of errors: Expected savings, of course, must be based on specific plans for
improvement. Often, such plans have a goal of cutting these costs by 50 percent within 5 years,
but such a potential benefit should not be assumed unless the problem areas for improvement
have been explicitly identified and an action plan with resources has been developed. In estimating
present costs, don’t inflate the present costs by including debatable or borderline items.
Decisive review meetings will get bogged down in debating the validity of the figures instead of
discussing the merits of the proposals for improvement.
Improved process capability: Expected savings can come from a reduction in variability (of
product characteristics or process characteristics) and other process losses such as redundant
operations, sorting inspections, retrieving missing information, and other non-value-added activities.
As with other benefits, these expected savings must be based on improvement plans.
Reduced customer defections: One early indicator of defections can be responses to the market
research question, “Would you purchase this product again?” In an early application by the
author, 10.5 percent of a sample of current customers of washing machines said they would not
repurchase; the reason was dissatisfaction with the machine, not with the dealer or price. At
$50 profit per machine, the lost profit due to likely customer defections was then estimated.
Progress has been made in quantifying the benefits of an effort to reduce defections. The parameters
include the economic effect of losing customers over the “customer life,” the level of
quality to retain present customers (see Section 18 under Customer Satisfaction versus
Customer Loyalty), and the effect on retention of the quality of handling customer complaints
(see Section 18 under Linking Customer Satisfaction Results to Customer Loyalty and to
Processes). Additional discussion is provided by Rust, Zahorik, and Keiningham (1994),
first-time work
Cost of
Cost of
Manage by more
new activities
first-time work
to improve quality
and profit
Much less
From current
To reduce cost
& improve quality
Applying Cost of Quality
& process improvement
FIGURE 8.4 Effects of identifying cost of quality. (AT&T 1990, p. 9.)
Chapter 6. Again, potential benefits should not be projected unless problem areas have been
identified and an action plan has been formulated.
Increase in new customers: This is a most difficult benefit to quantify and predict. Quality
improvements that make goods or services attractive to new customers will increase sales revenue
but the amount and the timing depend on many internal actions and external market forces. Note
that as the cost of poor quality is reduced, additional resources become available to finance new
features for the goods and services-without increasing the price. The result can be a dramatic
increase in market share.
The investments required to achieve the benefits may include diagnosis and other forms of analysis,
training, redesign of products and processes, testing and experimentation, and equipment.
Surprisingly, many improvement projects require little in costly equipment or facilities. The investment
is mainly in analysis work.
An issue in calculating an ROQ is the matter of assumptions and estimates. Both must be realistic
for the ROQ to be viewed as credible. Avoid assumptions which really represent ideal conditions;
such conditions never occur. Also, the impact of numerical estimates on the ROQ can be evaluated
(if necessary) using “sensitivity analysis.” This involves changing the estimates (say by ±20 percent)
and recalculating the ROQ to see the effect of the accuracy of the estimate on the ROQ. The ROQ
will change but if the amount is small, it adds credibility to the estimates.
Note that one source of benefits (reducing the cost of errors) is based on relatively “hard” data of costs
already incurred that will continue unless improvement action is instituted. A second source (reducing the
cost of defections) also represents a loss (of sales revenue) already incurred. Other sources (improvement
of process capability and gaining new customers) are not based on current losses but do represent important
opportunity costs. In the past, because of the difficulty of quantifying other sources of benefits, the
cost savings of a quality improvement program have been based primarily on the cost of errors. Advances
made in quantifying the impact of quality on sales revenue, however, are making it possible to add the
revenue impact to the return on quality calculation. At a minimum, the ROQ calculation can be based on
savings in the cost of errors. When additional data are available (e.g., process information or market
research data), then estimates for one or more of the other three sources should be included to calculate
the total benefits. A note of caution: this expanded view of the cost of poor quality could mean that “traditional”
quality improvement efforts (reducing cost of errors) will become entangled with other efforts
(increasing sales revenue), leading to a blurring of the traditional efforts on reducing errors. We need both
efforts—just as much as we need the sun and the moon.
The rate of return on an investment in quality activities translates into the ratio of average annual
benefits to the initial investment. (Note that the reciprocal—investment divided by annual savings—
represents the time required for savings to pay back the investment, i.e., the “payback period.”) But this
calculation of ROQ provides an approximate rate of return because it neglects the number of years
involved for the savings and also the time value of money. A more refined approach involves calculating
the “net present value” of the benefits over time. This means using the mathematics of finance to
calculate the amount today which is equivalent to the savings achieved during future years (see Grant,
Ireson, and Leavenworth1990, Chapter 6). Rust, Zahorik, and Keinubghan (1994) describe an approach
for calculating the ROQ incorporating savings in traditional losses due to errors, sales revenue enhancement
using market research information for customer retention, and the time value of money.
Wolf and Bechert (1994) describe a method to determine the payback of a reduction in failure
costs when prevention and appraisal expenditures are made. Bester (1993) discusses the concept of
net value productivity which addresses the cost of quality and the value of quality.
Those presenting the results of the cost study should be prepared to answer this question from management:
“What action must we take to reduce the cost of poor quality?”
The control of quality in many companies follows a recognizable pattern—as defects increase,
we take action in the form of more inspection. This approach fails because it does not remove the
causes of defects; i.e., it is detection but not prevention. To achieve a significant and lasting reduction
in defects and costs requires a structured process for attacking the main sources of loss—the failure
costs. Such an attack requires proceeding on a project-by-project basis. These projects in turn
require resources of various types (see Section 5, The Quality Improvement Process). The resources
must be justified by the expected benefits. For every hour spent to identify one of the vital few problems,
we often spend 20 hours to diagnose and solve the problem.
To gain approval from upper management for a quality improvement effort, we recommend the
following steps.
1. Establish that the costs are large enough to justify action (see, for example, Tables 8.1 and 8.2).
a. Use the grand total to demonstrate the need for quality improvement. This is the most significant
figure in a quality cost study. Usually, managers are stunned by the size of the total—
they had no idea the amount was so big. One memorable example was a leading manufacturer
of aircraft engines. When the total quality costs were made known to the managing director,
he promptly convened his senior executives to discuss a broad plan of action. Those presenting
the report should be prepared for the report to be greeted with skepticism. The cost may
be such that it will not be believed. This can be avoided if management has previously agreed
to the definition of the cost of poor quality and if the accounting function has collected the
data or has been a party to the data collection process. Also, don’t inflate the present costs by
including debatable or borderline items.
b. Relate the grand total to business measures. Interpretation of the total is aided by relating
total quality costs to other figures with which managers are familiar. Two universal languages
are spoken in the company. At the “bottom,” the language is that of objects and deeds: square
meters of floor space, output of 400 tons per week, rejection rates of 3.6 percent, completion
of 9000 service transactions per week. At the “top,” the language is that of money: sales,
profit, taxes, investment. The middle managers and the technical specialists must be bilingual.
They must be able to talk to the “bottom” in the language of objects and to the “top”
in the language of money. Table 8.4 shows actual examples of the annual cost of poor quality
related to various business measures. In one company which was preoccupied with meeting
delivery schedules, the quality costs were translated into equivalent added production.
Since this coincided with the chief current goals of the managers, their interest was aroused.
In another company, the total quality costs of $176 million per year for the company were
shown to be equivalent to one of the company plants employing 2900 people, occupying 1.1
million ft2 of space and requiring $6 million of in-process inventory. These latter three figures
in turn meant the equivalent of one of their major plants making 100 percent defective
work every working day of the year. This company is the quality leader in its industry.
Similarly, in an airplane manufacturing company, it was found useful to translate the time
spent on rework to the backlog of delivery of airplanes, i.e., reducing the rework time made
more hours available for producing the airplanes.
c. Show the subtotals for the broad major groupings of quality costs, when these are available.
A helpful grouping is by the four categories discussed above under Categories of Quality Costs.
Typically, most of the quality costs are associated with failures, internal and external. The
proper sequence is to reduce the failure costs first, not to start by reducing inspection costs. Then
as the defect levels come down, we can follow through and cut the inspection costs as well.
2. Estimate the savings and other benefits:
a. If the company has never before undertaken an organized program to reduce quality-related
costs, then a reasonable goal is to cut these costs in two, within a space of 5 years.
b. Don’t imply that the quality costs can be reduced to zero.
c. For any benefits that cannot be quantified as part of the return on quality, present these benefits
as intangible factors to help justify the improvement program. Sometimes, benefits can
be related to problems of high priority to upper management such as meeting delivery
schedules, controlling capital expenditures, or reducing a delivery backlog. In a chemical
company, a key factor in justifying an improvement program was the ability to reduce significantly
a major capital expenditure to expand plant capacity. A large part of the cost of
poor quality was due to having to rework 40 percent of the batches every year. The improvement
effort was expected to reduce the rework from 40 percent to 10 percent, thus making
available production capacity that was no longer needed for rework.
3. Calculate the return on investment resulting from improvement in quality. Where possible, this
return should reflect savings in the traditional cost of poor quality, savings in process capability
improvement, and increases in sales revenue due to a reduction in customer defections and
increases in new customers. See above under Return on Quality.
4. Use a successful case history (a “bellwether” project) of quality improvement in the company to
justify a broader program.
5. Identify the initial specific improvement projects. An important tool is the Pareto analysis
which distinguishes between the “vital few” and the “useful many” elements of the cost of
poor quality. This concept and other aids in identifying projects is covered in Section 5 under
Use of the Pareto Principle. An unusual application of the concept is presented in Figure 8.5.
Here, in this “double Pareto” diagram, the horizontal axis depicts (in rank order) within each
major category the cost of poor quality and the vertical axis shows the subcategories (in rank
order) of the major category. Thus, the first two major categories (patient care–related and
facility-related) account for 80 percent of the cost. Also, of the 64 percent which is patient
care related about 54 percent is related to variation in hospital practice and 46 percent to outpatient
thruput optimization. This graph has proven to be a powerful driver for action.
6. Propose the structure of the improvement program including organization, problem selection, training,
review of progress, and schedule. See Section 5, The Quality Improvement Process.
Justification is essential for an effective program of quality improvement. Approaches to justification
are discussed in more detail in Section 5 under Securing Upper Management Approval and
As formal quality improvement continues with projects addressed to specific problems, the measurement
of the cost of poor quality has several roles.
TABLE 8.4 Languages of Management
Money (annual cost of poor quality)
24% of sales revenue
15% of manufacturing cost
13 cents per share of common stock
$7.5 million per year for scrap and rework compared to a profit of $1.5 million per year
$176 million per year
40% of the operating cost of a department
Other languages
The equivalent of one plant in the company making 100% defective work all year
32% of engineering resources spent in finding and correcting quality problems
25% of manufacturing capacity devoted to correcting quality problems
13% of sales orders canceled
70% of inventory carried attributed to poor quality levels
25% of manufacturing personnel assigned to correcting quality problems
Roles to Support Improvement. These include:
1. Identify the most significant losses for an individual problem and the specific costs to be eliminated.
This helps to focus the diagnostic effort on root causes. Figure 8.6 shows an example of a
special report that dissects one element of cost (“penalties”) and summarizes data on the cost of
poor quality to identify “activity cost drivers” and “root causes.”
2. Provide a measure of effectiveness of the remedies instituted on a specific project. Thus, a project
quality improvement team should provide for measuring the costs to confirm that the remedies
have worked.
3. Provide a periodic report on specific quality costs. Such a report might be issued quarterly or
4. Repeat the full cost of poor quality study. This study could be conducted annually to assess overall
status and help to identify future projects.
5. Identify future improvement projects by analyzing the full study (see item 4 above) using Pareto
analysis and other techniques for problem selection.
Note that the emphasis is on using the cost of poor quality to identify improvement projects and
support improvement team efforts rather than focusing on the gloomy cost reporting.
Optimum Cost of Poor Quality. When cost summaries on quality are first presented to
managers, one of the usual questions is: “What are the right costs?” The managers are looking for a
standard (“par”) against which to compare their actual costs so that they can make a judgment on
whether there is a need for action.
Unfortunately, few credible data are available because (1) companies almost never publish such
data and (2) the definition of cost of poor quality varies by company. [In one published study, Ittner
(1992) summarizes data on the four categories for 72 manufacturing units of 23 companies in 5
industry sectors.] But three conclusions on cost data do stand out: The total costs are higher for complex
industries, failure costs are the largest percent of the total, and prevention costs are a small percent
of the total.
Out patient thruput optimization
Patient Cure Related Facility Related X-ray Admin. Billing
Variation in hospital practice
20 40 60 80 100% 0
FIGURE 8.5 Mayo Rochester improvement opportunity. (Adapted from Rider 1995.)
$198,714 Dollar impact
$19,871 Dollar impact
Incorrect shipments
10% Cost-driver
$19,872 Dollar impact
Late shipments
10% Cost-driver
percentage $994 Dollar impact
Poor workmanship
5% Root-cause percentage
$994 Dollar impact
Confusing documents
5% Root-cause percentage
$17,884 Dollar impact
Poor raw materials
90% Root-cause percentage
$23,846 Dollar impact
Confusing documents
15% Root-cause percentage
$33,185 Dollar impact
Poor workmanship
21% Root-cause percentage
$101,940 Dollar impact
Poor raw materials
64% Root-cause percentage
$19,871 Dollar impact
Confusing documents
100% Root-cause percentage
$44,711 Financial impact
Confusing documents
$34,179 Financial impact
Poor workmanship
$119,824 Financial impact
Poor raw materials
$158,971 Dollar impact
Partial shipments
80% Cost-driver
Financial Impact of
Common Root Causes
Root Causes Cost Drivers Key Cost of Poor
Quality Element
FIGURE 8.6 The completed cost-driver analysis. (Adapted from Atkinson, Hamburg, and Ittner, 1994, p. 142.)
The study of the distribution of quality costs over the major categories can be further explored
using the model shown in Figure 8.7. The model shows three curves:
1. The failure costs: These equal zero when the product is 100 percent good, and rise to infinity
when the product is 100 percent defective. (Note that the vertical scale is cost per good unit of
product. At 100 percent defective, the number of good units is zero, and hence the cost per good
unit is infinity.)
2. The costs of appraisal plus prevention: These costs are zero at 100 percent defective, and rise
as perfection is approached.
3. The sum of curves 1 and 2: This third curve is marked “total quality costs” and represents the
total cost of quality per good unit of product.
Figure 8.7 suggests that the minimum level of total quality costs occurs when the quality of conformance
is 100 percent, i.e., perfection. This has not always been the case. During most of the twentieth
century the predominant role of (fallible) human beings limited the efforts to attain perfection
at finite costs. Also, the inability to quantify the impact of quality failures on sales revenue resulted
in underestimating the failure costs. The result was to view the optimum value of quality of conformance
as less than 100 percent.
While perfection is obviously the goal for the long run, it does not follow that perfection is
the most economic goal for the short run, or for every situation. Industries, however, are facing
increasing pressure to reach for perfection. Examples include:
1. Industries producing goods and services that have a critical impact on human safety and wellbeing:
The manufacture of pharmaceuticals and the generation of mutual fund statements
provide illustrations.
2. Highly automated industries: Here, it is often possible to achieve a low level of defects by
proper planning of the manufacturing process to assure that processes are capable of meeting
Quality of conformance, %
Costs of appraisal
plus prevention
Failure costs
Cost per good unit of product
0 100
FIGURE 8.7 Model for optimum quality costs.
specifications. In addition, automated inspection often makes it economically feasible to perform
100 percent inspection to find all the defects.
3. Companies selling to affluent clients: These customers are often willing to pay a premium price
for perfect quality to avoid even a small risk of a defect.
4. Companies striving to optimize the user’s cost: The model depicted in Figure 8.7 shows the
concept of an optimum from the viewpoint of the producer. When the user’s costs due to product
failure are added to such models, those costs add further fuel to the conclusion that the optimum
point is perfection. The same result occurs if lost sales income to the manufacturer is included in
the failure cost.
The prospect is that the trend to 100 percent conformance will extend to more and more goods
and services of greater and greater complexity.
To evaluate whether quality improvement has reached the economic limit, we need to compare
the benefits possible from specific projects with the costs involved in achieving these benefits. When
no justifiable projects can be found, the optimum has been reached.
As structured quality improvement teams following the project-by-project approach have emerged
as a strong force, reporting on the cost of poor quality has focused on supporting these team activities.
These reports provide information which helps to diagnose the problem and to track the change
in costs as a remedy is implemented to solve the problem. Regularly issued reports with a fixed format
usually do not meet the needs of improvement teams. Teams may need narrowly focused information
and they may need it only once. What data are needed is determined by the team, and the
team often collects its own data (Winchell 1993).
Some companies use periodic reporting on the cost of poor quality in the form of a scoreboard.
Such a scoreboard can be put to certain constructive uses. To create and maintain the
scoreboard, however, requires a considerable expenditure of time and effort. Before undertaking
such an expenditure, the company should look beyond the assertions of the advocates; it should
look also at the realities derived from experience. (Many companies have constructed quality cost
scoreboards and have then abandoned them for not achieving the results promised by the advocates.)
After finalizing the categories for a cost scoreboard, the planning must include collecting
and summarizing the data, establishing bases for comparison, and reporting the results.
Summarizing the Data. The most basic ways are
1. By product, process, component, defect type, or other likely defect concentration pattern
2. By organizational unit
3. By category cost of poor quality
4. By time
Often, the published summaries involve combinations of these different ways.
Bases for Comparison. When managers use a scoreboard on the cost of poor quality, they are
not content to look at the gross dollar figures. They want, in addition, to compare the costs with some
base which is an index of the opportunity for creating these costs. A summary of some widely used
bases, along with the advantages and disadvantages of each, is presented in Table 8.5. The base used
can greatly influence the interpretation of the cost data.
It is best to start with several bases and then, as managers gain experience with the reports, retain
only the most meaningful. The literature stresses that quality costs be stated as percent of sales
income. This is a useful base for some, but not all, purposes.
Reporting the Results. The specific matters are the same as for those for other reports—
format, frequency, distribution, responsibility for publication. Atkinson, Hamburg, and Ittner
(1994) describe how reporting can help promote a cultural change for quality, Dobbins and Brown
(1989) provide “tips” for creating reports, and Onnias (1985) describes a system used at Texas
The likely trend is for cost of poor quality and other quality-related information to become integrated
into the overall performance reporting system of organizations. Kaplan and Norton (1996)
propose that such a system provide a “balanced scorecard.” Such a scorecard allows managers to
view an organization from four perspectives:
1. How do customers see us? (Customer perspective.)
2. What must we excel at? (Internal perspective.)
3. Can we continue to improve and create value? (Innovation and learning perspective.)
4. How do we look to shareholders? (Financial perspective.) The scorecard would include a limited
number of measures—both the financial result measures and the operational measures that drive
future financial performance.
Periodically (say annually), a comprehensive report on the cost of poor quality is useful to summarize
and consolidate results of project teams and other quality improvement activities. The format
for this report need not be identical to the initial cost of poor quality study but should (1) reflect
results of improvement efforts and (2) provide guidance to identify major areas for future improvement
The costs of poor quality affect two parties—the provider of the goods or service and the user. This
section discusses the impact on the provider, i.e., a manufacturer or a service firm. Poor quality also
increases the costs of the user of the product in the form of repair costs after the warranty period,
various losses due to downtime, etc. Gryna (1977) presents a methodology with case examples of
user costs of poor quality. The extent of these user costs clearly affects future purchasing decisions
of the user and thereby influences the sales income of the provider. This section stresses the potential
for profit improvement by reducing provider costs and by reducing loss of sales revenue due to
poor quality.
TABLE 8.5 Measurement Bases for Quality Costs
Base Advantages Disadvantages
Direct labor hour Readily available and understood Can be drastically influenced by automation
Direct labor dollars Available and understood; Can be drastically influenced by automation
tends to balance any inflation effect
Standard manufacturing More stability than above Includes overhead costs both fixed and variable
cost dollars
Value-added dollars Useful when processing costs Not useful for comparing different types of
are important manufacturing departments
Sales dollars Appeals to higher management Sales dollars can be influenced by changes in prices,
marketing costs, demand, etc.
Product units Simplicity Not appropriate when different products are made
unless “equivalent” item can be defined
An extension of this thinking is provided by applying concepts from the economics discipline.
Cole and Mogab (1995) apply economic concepts to analyze the difference between the “mass
production/scientific management firm” and the “continuous improvement firm” (CIF). A defining
feature of the CIF is the ability to add to the net customer value of the marketed product. The
net customer value is defined as the total value realized by the customer from the purchase and
use of the goods or service less that which must be sacrificed to obtain and use it. In terms of the
economy of countries, Brust and Gryna (1997) discuss five links between quality and macroeconomics.
From the birth of the cost of poor quality with the emphasis on the cost of errors in manufacturing,
the concept is now extended in the scope of cost elements and applies to manufacturing and service
industries in both the profit and nonprofit sectors.
AT&T (1990). AT&T Cost of Quality Guidelines. Document 500-746, AT&T’s Customer Information Center,
Indianapolis, IN.
Atkinson, Hawley, Hamburg, John, and Ittner, Christopher (1994). Linking Quality to Profits. ASQ Quality Press,
Milwaukee, and Institute of Management Accountants, Montvale, NJ.
Atkinson, John Hawley, et al. (1991). Current Trends in Cost of Quality: Linking the Cost of Quality and
Continuous Improvement. Institute of Management Accountants, Montvale, NJ.
Aubrey, Charles A., II (1988). “Effective Use of Quality Cost Applied to Service.” ASQC Quality Congress
Transactions, Milwaukee, pp. 735–739.
Bester, Yogi (1993). “Net-Value Productivity: Rethinking the Cost of Quality Approach.” Quality Management
Journal, October, pp. 71–76.
Bohan, George P., and Horney, Nicholas F. (1991). “Pinpointing the Real Cost of Quality in a Service Company.”
National Productivity Review, Summer, pp. 309–317.
Briscoe, Nathaniel R., and Gryna, Frank M. (1996). Assessing the Cost of Poor Quality in a Small Business.
Report No. 902, College of Business Research Paper Series, The University of Tampa, Tampa, FL.
Brust, Peter J., and Gryna, Frank M. (1997). Product Quality and Macroeconomics—Five Links. Report No. 904,
College of Business Research Paper Series, The University of Tampa, Tampa, FL.
Campanella, Jack, Ed. (1999). Principles of Quality Costs, 3rd ed., ASQ, Milwaukee.
Carr, Lawrence P. (1992). “Applying Cost of Quality to a Service Business.” Sloan Management Review,
Summer, pp. 72–77.
Cole, William E., and Mogab, John W. (1995). The Economics of Total Quality Management: Clashing
Paradigms in the Global Market. Blackwell, Cambridge, MA.
Dawes, Edgar W., and Siff, Walter (1993). “Using Quality Costs for Continuous Improvement.” ASQC Annual
Quality Congress Transactions, Milwaukee, pp. 810–816.
Dobbins, Richard K., and Brown, F. X. (1989). “Quality Cost Analysis—Q.A. Versus Accounting.” ASQC Annual
Quality Congress Transactions, pp. 444–452.
Esterby, L. James (1984). “Measuring Quality Costs by Work Sampling.” Quality Costs: Ideas and Applications.
Andrew F. Grimm, Louis A. Pasteelnick, and James F. Zerfas, eds., ASQC Quality Press, Milwaukee, pp.
Finnegan, John T., and Schottmiller, John C. (1990). “Cost of Quality—Multinational and Multifunctional.”
ASQC Quality Congress Transactions, Milwaukee, pp. 746–751.
Garrison, Ray H., and Noreen, Eric W. (1994). Managerial Accounting. Irwin, Boston.
Grant, Eugene L., Ireson, W. Grant, and Leavenworth, Richard S. (1990). Principles of Engineering Economy,
8th ed., John Wiley and Sons, New York.
Gryna, Frank M. (1977). “Quality Costs: User vs. Manufacturer.” Quality Progress, June, pp. 10–15.
Harrington, H. James (1986). Poor Quality Cost. ASQC Quality Press, Milwaukee, and Marcel Dekker, New
Hester, William F. (1993). “True Quality Cost With Activity Based Costing.” ASQC Annual Quality Congress
Transactions, Milwaukee, pp. 446–454.
Hou,Tien-fang (1992). “Cost-of-Quality Techniques for Business Processes.” ASQC Annual Quality Congress
Transactions, Milwaukee, pp. 1131–1137.
ISO 9004-1 (1994). Quality Management and Quality System Elements—Guidelines, Section 6, “Financial
Considerations of Quality Systems.” ISO, Geneva.
Ittner, Christopher Dean (1992). The Economics and Measurement of Quality Costs: An Empirical Investigation.
Doctoral dissertation, Harvard University, Cambridge, MA.
Kaplan, Robert S., and Norton, David P. (1996). The Balanced Scorecard, Harvard Business School Press,
Keaton, Willie E., et al. (1988). “White Collar Quality Pinpointing Areas for Breakthrough.” IMPRO 1988
Conference Proceedings, Juran Institute, Inc., Wilton, CT., pp. 3C-19 to 3C-24.
Krause, James D., and Gryna, Frank M. (1995). Activity Based Costing and Cost of Poor Quality—a Partnership.
Report No. 109, College of Business Research Paper Series, The University of Tampa, Tampa, FL.
Nickell, Warren L. (1985). “Quality Improvement in Marketing.” The Juran Report, no. 4 (Winter), pp.
29–35.O’Neill, Ann R. (1988). “Quality Costs In  Continuous Improvement Out.” ASQC Annual Quality
Congress Transactions, Milwaukee, pp. 180–187.
Onnias, Arturo (1985). “The Quality Blue Book.” IMPRO 1985 Conference Proceedings, Juran Institute, Inc.,
Wilton, CT, pp. 127–134.
Quevedo, Rene (1991). “Quality, Waste, and Value in White-Collar Environments.” Quality Progress, January,
pp. 33–37.
Reeve, J. M. (1991). “Variation and the Cost of Quality.” Quality Engineering, vol. 4, no. 1, pp. 41–55.
Rider, C. T. (1995). “Aligning Your Improvement Strategy for the Biggest Payback.” IMPRO 1995 Conference
Proceedings, Juran Institute Inc., Wilton, CT, pp. 7E. 1-1 to 7E.1-11.
Romagnole, Anthony J., and Williams, Mary A. (1995). “Costs of Poor Quality.” IMPRO 1995 Conference
Handout, Juran Institute, Inc., Wilton, CT.
Rust, Roland T., Zahorik, Anthony J., and Keiningham, Timothy L. (1994). Return on Quality. Probus, Chicago.
Schottmiller, John C. (1996). “ISO 9000 and Quality Costs.” ASQC Annual Quality Congress Proceedings,
Milwaukee, pp. 194–199.
Schrader, Lawrence, J. (1986). “An Engineering Organization’s Cost of Quality Program.” Quality Progress,
January, pp. 29–34.
Shepherd, Nick A. (1998). “Cost of Quality and ISO Implementation Using Cost of Quality as a Driver for
Effective Use of ISO 9000 and 14000.” ASQ Annual Quality Congress Transactions, pp. 776–783.
The Conference Board (1989). Current Practices in Measuring Quality. Research Bulletin No. 234, New York.
The University of Tampa (1990). “Costly Mistakes.” Visions, Fall, p. 7.
Wasserman, Gary S., and Lindland, John L. (1994). “Minimizing the Cost of Quality Over Time: A Dynamic
Quality Cost Model.” ASQC Annual Quality Congress Proceedings, pp. 73–81.
Wild, Cheryl, and Kovacs, Carol (1994). “Cost-of-Quality Case Study.” IMPRO 1994 Conference Proceedings,
Juran Institute, Inc., Wilton, CT, pp. 3B-19 to 3B-29.
Winchell, William O. (1991). “Implementation of Quality Costs—Making It Easier.” ASQC Annual Quality
Congress Transactions, Milwaukee, pp. 592–597.
Winchell, William O. (1993). “New Quality Cost Approach Gives Proven Results.” ASQC Annual Quality
Congress Transactions, Milwaukee, pp. 486–492.
Wolf, C., and Bechert, J. (1994). “Justifying Prevention and Appraisal Quality Cost Expenditures: A Benefit/Cost
Decision Model.” Quality Engineering, vol. 7, no. 1, pp. 59–70.
Thomas C. Redman
Managers Need Information 9.1
Systems of Measurement 9.1
About This Section 9.3
Make Decisions/Take Action 9.4
Understand Framework 9.5
Plan Measurements 9.7
Collect Data 9.9
Analyze, Synthesize, Formulate Results,
and Present Results and
Recommendations 9.9
Data Quality and Measurement
Assurance 9.11
Checklist 9.11
Objective Measurement versus
Subjective Measurement 9.12
Systems of Systems 9.14
Manager’s Intuition 9.14
Managers Need Information. The general business leader and quality manager share an
eternal lament first voiced by Alexander the Great: “Data, data, data! I’m surrounded by data. Can’t
anyone get me the information I really need?” Alexander needed answers to some rather basic questions.
“Where is the enemy? How large are his forces? How well will they fight? When will our supplies
arrive?” His spies, lieutenants, and others gathered data. But they could not always satisfy
Alexander’s need for information.
Today’s business manager also has a voracious appetite for information (see also Drucker 1995).
Obtaining needed information is time-consuming, expensive, and fraught with difficulty. And in the
end, the manager is often less informed than Alexander the Great. Basic questions like “What do customers
really want? How well are we meeting their needs? What is the competitor going to do next?”
are not easier to answer. This, despite an amazing proliferation of measuring devices, customer surveys,
statistical methods, and database and networking technology. Just as quality management is a
never-ending journey, so too is the task of learning of, obtaining, sorting through, synthesizing, and
understanding all the data and information that could productively be used. It seems that the manager’s
appetite for information will never be satisfied.
Systems of Measurement. We usually consider information in light of decisions that managers
must make and the actions they take. Information plays a similar role in science. The scientific
method is essentially a process by which hypotheses are proposed, experiments designed to test
aspects of those hypotheses, data collected and analyzed, and the hypotheses either advanced, discarded,
or modified. Time-tested and rigid standards apply. Most businesses cannot afford the required
rigor. The system of measurement advanced here parallels the scientific method, but the standards are
different. “Caution” is the watchword in science. “Success” is the watchword in business. (See
Section 47 for further reference to the scientific method and experimental design.)
In some cases, the manager is presented with well-defined choices. A simple example is the question:
“Is this process in a state of control?” The manager can either decide that the process is in control
and take no action, or decide that the process is not in control and take action to find and
eliminate a special cause. (See Section 4: The Quality Control Process.) In other situations, the range
of options is ill-defined and/or unbounded. In many cases (perhaps too many), the manager may even
choose to gather more data. In almost all cases, it appears axiomatic that the better the information,
the better the decision. Here better information may have any number of attributes, including more
complete, more accurate, more relevant, more current, from a more reliable source, more precise,
organized in a more convincing fashion, presented in a more appealing format, and so forth.
A critical step in obtaining needed information is measurement. To measure is “to compute, estimate,
or ascertain the extent, dimensions, or capacity of, especially by a certain rule or standard”
(Webster 1979). Measurement, then, involves the collection of raw data. For many types of measurements,
specialized fields have grown up and there is a considerable body of expertise in making
measurements. Chemical assays and consumer preference testing are two such areas. Data collection
may involve less formal means—searching a library, obtaining data originally gathered for other
purposes, talking to customers, and the like. For our purposes, all such data collection shall be considered
Of course there is more to developing the information the manager needs then simply collecting
data (indeed, therein lies Alexander’s lament). For examples we cited “more relevant” and “presented
in a more appealing format” as attributes of better information. It is evident that the choice of what
to measure and the analysis, synthesis, and presentation of the resultant information are just as
important as the act of measurement itself. High-quality information results only from high-caliber
and integrated design, data collection, and analysis/synthesis/presentation. Thus we distinguish
between the act of measurement, or data collection, and the measurement process, which includes
design, data collection, and analysis/synthesis/presentation. This process is presented in Figure 9.1
and described throughout this section.
But good information is not a decision. So Figure 9.1 goes further. The measurement process
is preceded by a step “Understand framework” and followed by a step “Make decision/take
action.” These steps put the measurement process in its proper context and represent the
suppliers to and customers of the measurement process. Finally, Figure 9.1 features one further
important subtlety. Decision makers in most organizations are often too busy to carefully consider
all the data and evaluate alternatives. What they require are recommendations, not just
clearly presented information. So the analysis/synthesis/presentation step is better described as
analysis/synthesis/recommendations/presentation of results and recommendations. It could be
FIGURE 9.1 The act of measurement is but one step in a larger measurement system. Herein we consider the
measurement process as consisting of steps needed to collect data and present results. The larger measurement
system also embraces the decisions that are made and the framework in which the process operates.
argued that formulation of results is better thought of as a part of decision making. But it is more
often the case that those analyzing data also are expected to make recommendations. Indeed, they
may be petitioning decision makers to adopt their recommendations.
As will be discussed, decision making is often a complex political process, in and of itself. Thus
the framework in which information is produced is critical. At a high level, the framework defines
the overall context in which decisions are to be made, including such diverse considerations as the
organization’s business goals, its competitive position, and its available resources; customer
requirements; the goals and biases of decision makers; and any relevant constraints on the measurement
process or decision making. For any particular decision, the framework includes the specific
issues to be addressed. Taken together, these five elements (framework, design, data collection,
analysis/synthesis/recommendations/presentation, and decision/action) compose a “measurement
Ultimately, the goal is to help the organization make consistently better decisions and take better
actions. Here, better decisions are defined in terms of results in achieving organizational objectives.
Good measurement systems support a number organizational goals, not just a few (Kaplan and
Norton 1992, 1993; Meyer 1994). It is usually true that “what gets measured, gets managed.” Most
organizations intend to pursue a variety of goals, satisfying customers, providing a healthy and fulfilling
workplace, meeting financial objectives, and so forth. In many organizations financial measurements
are the most fully developed and deployed. It is little wonder that financial considerations
dominate decision making in such organizations (Eccles 1991).
The most important point of this section is that those who define and operate a measurement
process (i.e., those who gather or analyze data or who recommend what the organization should do)
must consider the system as a whole, including the environment in which it operates. It is not sufficient
to be technically excellent or for one or two elements to be outstanding (see Futrell 1994 for a
review of the failures of customer satisfaction surveys). Overall effectiveness is as much determined
by how well the elements relate to one another and to other systems in the enterprise as by excellence
in any area. The following personal anecdote illustrates this point.
Very early in my career, I was asked to recommend which color graphics terminal our department
should buy. I spent a lot of time and effort on the measurement process. I talked to several users about
their needs, called a number of vendors, arranged several demonstrations, and read the relevant literature.
At the time (early 1980s) the underlying technology was in its infancy—it had many problems
with it and a terminal cost about $15,000. Further, many people anticipated dramatic and
near-term improvements to the technology and substantial price reductions. So I recommended that
we wait a year and then reconsider the purchase. I was proud of my work and recommendation and
presented it to my manager. He promptly informed me I had misunderstood his question. The question
was not “Should we buy a terminal?” but “Which terminal should we buy?” (and none was not
a permitted answer). In retrospect, I could have: explicitly defined the possible decisions in advance,
or thought through the framework. My manager was a forward thinker. He had seen the potential for
personal computing and clearly wanted to experiment. Even casual consideration of his objectives
would have made it clear that “wait a year” was an unacceptable recommendation.
About This Section. The primary audiences for this section are persons interested in helping
their organizations make better decisions. I have already noted that obtaining relevant information
can be time-consuming, expensive, and fraught with difficulty. On the other hand, there are many
cost-effective practices that everybody can take to reduce the gap between the information they
desire and the information they have at their disposal.
Readers may have several interests:
 Those whose measurement and decision processes are well defined may be especially interested
in comparing their processes to the “ideals” described here.
 Those who make measurements and/or decisions routinely, but whose processes are ad hoc, may
be especially interested in applying the techniques of process management to improve their measurements
and decisions.
 Those whose interests involve only a single “project” may be especially interested in learning
about the steps their project should involve.
This section does not consider many technical details. Entire volumes have been written on technical
details of measurement making (see Finkelstein and Leaning 1984 and Roberts 1979, for example),
statistical analysis (see Sections 4, 44, 45, and 47 of this Handbook, for example), graphical
data presentation, and the like.
In the next section, we consider measurement systems in greater detail. Outputs of each step are
defined and the activities needed to produce those outputs are described. Then, we consider more
complex measurement systems, involving hierarchies of measurements. The fourth section provides
practical advice for starting and evolving a measurement system. The final section summarizes good
practice in a list of the “Top 10 Measurement System Principles.”
Three examples, of escalating complexity, are used to illustrate the main points. At the so-called
“operational level,” we consider the measurement system for a single step for the billing process
summarized in Figure 9.2. At the tactical level, we consider the measurement system needed to support
changes to the feature set associated with the invoice (the result of the billing process). We also
consider measurement systems needed to support strategic decisions. Virtually everyone is involved
in some way or another at all three levels of decision making.
A middle manager may find himself or herself playing the following roles:
 A designer of the measurement system used at the operational level
 A decision maker at the tactical level
 A supplier of data to strategic decisions
Figure 9.3 expands on Figure 9.1 in two ways: It lists the work products produced at each step and
describes the work activities in more detail. The first two steps (understand framework and plan measurement)
are planning steps. There is no substitute for careful planning, so most of the discussion
is on planning. The most important work products of the system are decisions and actions. Other
work products aim to produce better decisions and actions. So we begin the discussion with the final
step: Make decision/take action.
Make Decisions/Take Action. The first step in defining a measurement system is to understand
who will make the decisions and how. Many decisions, and virtually all decisions of consequence, are
not made by an individual, but by a committee or other group. In some cases, this helps build support
for implementation. In others, it is more a vehicle for diffusing accountability. Some groups decide by
majority rule, others by consensus. Most groups have a few key individuals. Some are thoughtful leaders,
others are self-centered, domineering bullies. Few decision makers are completely unbiased. Most
are concerned, at least to some degree, with their careers. Individuals intuitively make decisions based
on different criteria. Some are risk takers, others are risk-averse. Some are concerned only with the near-
FIGURE 9.2 A hypothetical billing process.
term financial impacts, others consider the long-term. And those who may be saddled with responsibility
for implementation have other perspectives. Decision making is thus also a political process that the
designer of the measurement system is well advised to understand.
Understand Framework. Prior to determining what to measure and how to measure it, it is
important to understand the overall framework in which the measurement system operates. We’ve
already noted the political nature of decision making. Those who make decisions and take actions
are members of organizations, and all organizations have their own politics and cultures that define
acceptable directions, risks, behaviors, and policies that must be followed. These features of the
organization form much of the context or framework for the measurement system. Good measurement
systems usually work in concert with the organizational culture. But they are also capable of
signaling need for fundamental changes in the culture.
Defining the framework is somewhat akin to stakeholder analysis in strategic planning.
Stakeholders include at least three groups: customers, owners (and perhaps society), and employees.
Each has valid, and sometimes conflicting, goals for the organization. These impact the organization’s
business model and in turn, the measurement system. We consider each in turn. See
Figure 9.4.
1. Customers: One goal of most organizations is to maintain and improve customer satisfaction,
retention, and loyalty. Customer needs are usually stated in subjective terms. At the operational
level, a consumer may simply want “the bill to be correct.” At the tactical level, an important
requirement may be that the invoice feed the customer’s invoice payment system. Finally, at the
strategic level, business customers may want to establish single sources of supply with companies
they trust. It is important to recognize that there is an element of subjectivity in each customer
requirement. Technicians are often dismayed by customers’ lack of ability to give clear, objective
requirements. But customers and their communications abilities are explicitly part of the overall
FIGURE 9.3 Further detail about subprocesses of the measurement system. For each subprocess, the
principal work product and several steps are given.
2. Owners: Owners of the business are usually less concerned about day-to-day considerations
and more concerned about the economic viability, both long- and short-term, of the business. Their
interests are reflected in corporate direction, competitive position, and financial performance. They
usually wish, implicitly at least, to see costs kept to the lowest levels.
The impact of their interests on the measurement system is that certain things are more important
to measure than others. Consider a company that wishes to pursue a strategy of price leadership.
It wishes to be perceived as “just as good” as a major competitor, but wants to keeps costs
as low as possible. Such a company will design its measurement system around competitive intelligence.
It will not, for example, invest to learn of customers’ preferred bill formats. In contrast,
a company pursuing customer value and intimacy will design its measurement system to learn
more about customers, how they use the organization’s products and services, and how to use
these measurements in defect prevention. It will invest to learn how to make its invoices a source
of advantage.
3. Employees: Insofar as measurements systems are concerned, employees are stakeholders
because they depend on the organization for their livelihood. Many are also stakeholders because
they make decisions and take actions, others because they collect data, and so forth. Employees
may view the measurement system as a device of management control, but good measurement
systems are also empowering. In our operational example, the day-to-day decision maker could
be owner of the guiding function, the billing process owner, or even a product manager. There
may be good reasons for the process owner—local custom, required skill, union rules—to be the
decision maker. But it is usually preferable that decisions be made as close to the action as possible.
So unless there is a compelling reason otherwise, the owner of the guiding function is the
preferred decision maker. In contrast, poor measurement systems require much additional time
and are of dubious value.
A second aspect of understanding the framework involves the range of possible decisions and
actions. A list of such decisions and actions is called the “decision/action space.” In some cases creating
this list is a straightforward exercise and in others it is nearly impossible. At the lowest, or
operational, level it is usually possible to describe the decision space completely. Thus, in our first
example, there are only a few possible decisions:
FIGURE 9.4 The measurement system is impacted by and impacts virtually all systems within an
 The process is in control and performing at an acceptable level and should be left alone,
 The process is out of control and must be brought under control.
 The process is in control but not performing at an acceptable level. It must be improved.
At the tactical and strategic levels, the exercise of defining the decision/action space becomes
more difficult. The range of possible decisions may be enormous, many possible decisions may be
difficult to specify beforehand, and some decisions may be enormously complex. In our second (tactical)
example, one possible decision is to leave the invoice alone. But the invoice can also be
improved, possibly in a virtually unlimited number of ways. There are any number of incremental
improvements to the formatting and accounting codes. Or the invoice can be wholly redesigned. A
paper invoice may be replaced with an electronic one, for example. Finally the invoice may even be
replaced with a superior invoice based on electronic commerce on the Internet.
Experience suggests that the more carefully the decision/action space is defined, the better the
resultant decisions. This is just as true at the strategic level as it is at the operational level.
Unfortunately there are no complete templates for defining the framework. Any number of other
considerations may be pertinent. For examples, legal obligations, safety rules, or technical limitations
may be very important.
Framework Document. The end result should be a framework document that captures the major
points of the work conducted here. It should describe major business goals and strategies and customer
requirements, define the decision/action space and decision makers (by name in some cases,
by job classification in others), note important constraints (financial and other), and reference more
detailed business plans and customer requirements. And, as the business grows and changes, so too
should the framework document.
Plan Measurements. Once the planner understands decision space and the framework in
which the measurement system operates, plans for the remaining steps of the process are made. The
output of this step is a “measurement protocol”, a document that describes the “whats, whens,
wheres, hows, and how often” of data collection, storage, and planned analyses. Perhaps most importantly,
the protocol should also describe the “whos”—who is responsible for each step. Figure 9.5
portrays the landscape to be covered. The most important issues to be addressed are discussed below.
Data Collection: What to Measure. Above, we noted that most customer requirements are stated
in subjective terms. These requirements have to be translated into a set of objective measurements.
A good example involves the early days of telephony. Customers’ most basic requirement was “to
hear and be heard.” An almost unlimited number of problems can thwart this basic requirement.
And, except for the actual speech into the mouthpiece and sound emanating from the earpiece, a
phone call is carried electrically. A remarkable series of experiments helped determine that three
parameters, loss, noise, and echo, each measurable on any telephone circuit or portion thereof,
Data collection Data storage presentation
How often
FIGURE 9.5 The landscape to be covered by a measurement protocol.
largely determined whether the customer could hear and be heard (Cavanaugh, Hatch, and Sullivan
1976; AT&T 1982).
In recent years, Quality Function Deployment (Hauser and Clausing 1988) has proven to be an
invaluable tool in helping map subjective user requirements into objective criteria for process performance.
Figure 9.6 illustrates an ideal scenario in which the user requirement for a “correct bill”
is first translated into a small number of objective parameters that are further translated into requirements
on steps in the billing process (and, in particular, on guiding).
Naturally, many requirements will never lend themselves to objective measurement. Our second
example, involving changes to a feature set, is such an example. Here the primary sources of data
will be customer satisfaction surveys, customer feedback, observation of competitors’ features, and
views of likely technological innovations.
In some cases, it is pretty clear what you would like to measure, but you simply can’t measure
it. A famous story involves the vulnerability to enemy fire in World War II planes. The goal was
to have more aircraft complete their missions and return safely. And, ideally, one would like to
determine where aircraft that didn’t return were hit. But these aircraft were not available. Good
surrogate measurements are needed in such cases. The problem with World War II aircraft was
addressed by examining where planes that did return were hit and assuming that those that didn’t
return were hit elsewhere.
In almost all cases, literally dozens of possible measurements are possible. The planner is usually
well advised to select “a critical few” measurements. There are decreasing returns as measurements
are added, and too many measurements can overwhelm the measurement system. The planner should
list possible measurements and rank-order them. There will be an essential few that he/she must
select. Other measurements should only grudgingly be admitted. Reference to the framework is usually
most helpful in making the necessary selections.
Many planners fall into a trap by concentrating on getting a few good measures for each step of a
process and giving insufficient attention to overall measurement. There is a compelling logic that, in
billing for example, if each step performs correctly, then the overall process will perform correctly.
Unfortunately this logic is often incorrect. Too many problems can arise between steps, where
accountability is not clear. An overall measure of bill correctness and measures of correctness at each
step are needed. The principles of business process management are covered in Section 6.
Precise definitions of what is measured are essential, as slight changes in definition can produce
very different results, a fact that advertisers may exploit (see Schlesinger 1988 for one example).
Data Collection: Where. The planner must simultaneously determine where to make measurements.
In quality management, the usual evolution is from inspection at the end of a production
process to measurement of in-process performance. Immature systems place greater weight on
inspection, more mature ones on in-process measurement (see Ishiakawa 1990).
Data Collection: When, How, How Often. In some cases, no new data are collected, but rather
existing data from “customer accounts” or other databases are used. The planner is still advised
All purchases made by the
day before the billing date
must be billed.
All purchases must be on
the proper account.
All discounts and payments
must be properly applied.
On the billing date, all
purchases for the prior
day and before must be
associated with the
proper account.
All purchases that cannot
be properly guided are
sent back to order-entry
by the end of the business
“I want a
FIGURE 9.6 Customer requirements are usually subjective. They need to be translated into objective
measurable parameters. Here it is done for the requirement “I want a correct bill” and the guiding step of
the billing process.
to learn about the intricacies of data collection, as data collected for one purpose may not suit
The planner next specifies how, when, and how often measurements are to be made. Each should
be spelled out in full detail. “How” involves not only how a particular measurement is to be made,
but also how the measurement equipment is to be calibrated and maintained and how accurate data
are to be obtained. “When” and “how often” must be addressed to ensure that ample data are available.
The interested reader is referred to Sections 44, 45, and 47.
Data Storage and Access. Perhaps nothing is more frustrating than knowing “the data are in the
computer” but being unable to get them. Planners too often give insufficient attention to this activity,
and data storage and retrieval becomes the Achilles’ heel of the system. Suffice it to note that, despite
an explosion in database technology, especially in ease of use, data storage and retrieval are not easy
subjects and should be carefully planned.
Data Analysis, Synthesis, Recommendations, and Presentation. Finally, the planner must consider
the analysis, synthesis, formulation of recommendations, and presentation step. While all the analyses
that will be carried out cannot be planned, certain basic ones should be. Thus, in our operational example,
the addition of a point to a control chart at specified intervals should be planned in advance.
Who. Just as important as what, where, when, and how is who. Who collects data, who stores
them, who plots points on control charts, who looks at data in other ways. All should be specified.
Measurement Protocol. The output of this step is a measurement protocol that documents plans
for data collection and storage, analysis/synthesis and presentation. In effect, the measurement protocol
defines the sub-processes to be followed in subsequent steps. Written protocols appear to be
common in many manufacturing, service, and health-care settings. In many other areas, particularly
service areas, written protocols are less common. This is poor and dangerous practice. Protocols
should be written and widely circulated with those who must follow them. There are simply too
many ways to interpret and/or bypass verbal instructions. The protocol should be carefully maintained.
Like many useful documents it will be in constant revision.
Collect Data. When all goes well, data collection involves nothing more than following the measurement
protocol. All going well seems to be the exception rather than the rule, however. For this reason,
those making measurements should maintain careful logs. Good discipline in maintaining logs is
important. Logs should be kept even when calibration and measurement procedures go as planned. It
is most important that any exceptions be carefully documented. One topic, data quality, deserves special
attention (see also Redman 1996 and Section 34). Unfortunately, measuring devices do not always
work as planned. Operators may repeat measurements when, for example, values are out of range. Or
data analysts may delete suspect values. Detecting and correcting erred data goes by many names:
data scrubbing, data cleanup, data editing, and so forth. It is better to detect and correct readings as
they are made, rather than later on. And naturally it is best to control data collection so that errors are
prevented in the first place. But wherever errors are caught, changes in data must be carefully logged.
Protocols for tactical and strategic systems often call for literature scans, attendance at professional
conferences, discussion with consultants, and so forth. When data are gathered in this manner,
it is important that sources be documented. It is best to determine original sources.
Analyze, Synthesize, Formulate Results, and Present Results and
Recommendations. Once data are collected, they must be summarized and presented in a
form that is understandable to decision makers. This step is often called “data analysis”. But that
term is a misnomer. “Analysis” is defined as “separating or breaking up of any whole into its parts
so as to find out their nature, proportion, function, relationship, etc.” Analysis is absolutely essential,
but it is only one-fourth of the required activity. The other three-fourths are “synthesis”, “formulation
of results”, and “presentation”. Synthesis is “composition; the putting of two or more things
together so as to form a whole: opposed to analysis.” Alexander’s henchmen (and many others) seem
not to have heard of synthesis. Next, specific “recommendations” for decision/action are developed.
Finally, presentation involves putting the most important results and recommendations into an
easily understood format.
That said, we use “analysis” as a shorthand for analysis, synthesis, formulation
of results, and presentation. There are four steps:
 Completing planned analysis
 Exploratory data analysis (when appropriate)
 Formulation of results and recommendations
 Presentation of results and recommendations
For the operations example, the planned data analysis and presentation involves nothing more
than computing an average and control limits and plotting them and requirements lines on a chart.
Such a chart is presented in Figure 9.7. Simple as it is, the control chart is ideal:
1. It prescribes the proper decision for the manager.
2. It is graphical (and most people more readily interpret graphs) and visually appealing.
3. It is easy to create, as it does not require extensive calculation (indeed, much of its utility on the
factory floor stems from this point).
4. It empowers those who create or use it.
5. Perhaps most important, the control chart provides a basis for predicting the future, not just
explaining the past. And the whole point of decision making is to make a positive impact on
the future.
Unfortunately, in tactical and strategic situations, this step is not so simple. We have already
noted that many analyses should be planned in advance. Time should also be allotted for data
exploration (also called data analysis, exploratory data analysis, data visualization, etc.). There are
often “hidden treasures” data, waiting to be discovered. Indeed explosions in database and graphical
exploration technologies can increase any organization’s ability to explore data. More and
more, all data, including details of all customer transactions and operations, are available. The
most critical element, though, is not tools, but inquisitive people who enjoy the detective work
needed to uncover the treasures.
The output of this step is a “presentation package.” It may be nothing more than the control
chart. The new tools also make it possible for every organization to present results in a clear,
FIGURE 9.7 The control chart, quite possibly the best presentation vehicle in quality
understandable, and engaging manner to every organization. Experience suggests that good presentation
 Comprehensive: The presentation covers the points of view of all pertinent stakeholders. It is, in
effect, a “balanced scorecard” (Kaplan and Norton 1992; Eccles 1991).
 Presented in layers: High-level summaries that cover the landscape are presented in overview
and layers of additional detail are available, as needed.
 Graphical: Almost everyone prefers well-conceived graphical summaries to other forms of presentation.
See Tukey (1976), Tufte (1983), and Chambers et al. (1983) for good examples and
 Fair and unbiased.
 To the point: Recommendations should be specific and clear.
We conclude with a quotation: “We also find that common data displays, when applied carefully,
are often sufficient for even complex analyses…” (Hoaglin and Velleman 1995).
Data Quality and Measurement Assurance. Clearly, decisions are no better than the
data on which they are based. And a data quality program can help ensure that data are of the highest
possible quality. One component of a data quality program is measurement assurance. The
National Institute of Standards and Technology (NIST) defines a measurement assurance program
as “a quality assurance program for a measurement process that quantifies the total uncertainty of
measurements (both random and systematic components of error) with respect to national or other
standards and demonstrates that the total uncertainty is sufficiently small to meet the user’s
requirements.” (Carey 1994, quoting Belanger 1984). Other definitions (Eisenhart 1969; Speitel
1982) expanded, contracted, or refocused this definition slightly. All definitions explicitly recognize
that “data collection,” as used here, is generally itself a repeatable process. So the full range
of quality methods is applicable. Clearly measurement assurance is a component of a good measurement
But the data quality program should be extended to the entire system, not just data collection.
The measurement system can be corrupted at any point, not just at data collection.
Consider how easy it is for a manager to make an inappropriate decision. Figure 9.8 illustrates
a simple yet classic situation. Note that the control chart clearly indicates a stable process that
is meeting customer needs. However, an inexperienced manager, at the point in time indicated
on the chart, notes “deteriorated performance in the previous three periods.” The manager may
decide that corrective action is needed and take one, even though none is indicated. Such a manager
is “tampering.” Unless saved by dumb luck, the best he or she can accomplish is to waste
time and money. At worst, he or she causes the process to go out of control with deteriorated
In many organizations it is common to “manage the measurements, not the process.”
Departmental managers urged to reduce their travel costs may be facile at moving such costs to their
training budgets, for example. In good measurement systems, the measurement assurance program
helps all components function as planned.
Checklist. Figure 9.9 summarizes the elements of a good measurement system for the billing
example. The planner, moving from left to right on the figure, first defines the desired decision/action
space. Next, the overall context is defined. It consists of three components: the customer’s overall
requirements and specific requirements on this step, the choice of operator as the decision maker and
his/her understanding of control charts, and the budget allotted for quality control on this assembly
line. The measurement protocol specifies the data collection plan, and raw data is collected accordingly.
The control chart is the presentation vehicle.
To conclude this section, Figure 9.10 presents a measurement system checklist. It may be used to
plan, evaluate, or improve a system. Not all items on the list are of equal importance in all systems.
Objective Measurement versus Subjective Measurement. Texts on science, management,
and statistics often extol the virtues of objective measurement, made on a nominal scale,
over subjective “opinions,” especially those of the “Yes/No” variety. The reasons are simple:
Opinions are imprecise and subject to change.
These disadvantages aside, subjective measurements have certain advantages that should not be
dismissed. First and foremost, customer opinion is the final arbiter in quality management. Second,
FIGURE 9.8 The control chart misinterpreted. At point 417, the overzealous manager sees
evidence of degraded performance.
FIGURE 9.9 The major elements of a measurement system for the guiding step are summarized.
there are always requirements that customers are not able to state in objective terms. A fairly typical
customer requirement is: “I want to trust my suppliers.” “Trust” simply does not translate well into
objective terms. Of course, the customer may be able to give examples that illustrate how a supplier
can gain trust, but these examples do not form a solid basis for a measurement system. Finally, there
is a richness of emotion in customer opinion that does not translate well into objective parameters.
If the customer discussing trust above adds: “If this critical component fails, our product will fail.
We’ll lose customers and jobs if that happens,” there can be no mistaking this customer’s priorities
and the sense of urgency associated with such components.
Make decision/take action
 Is the decision/action space specified as clearly as possible?
 Are the planned actions consistent with the decision and the data that lead to them?
Define framework
 Are customer requirements clearly defined?
 Are the requirements of owners clearly defined?
 Are employee requirements clearly defined?
 Are any other major stakeholders identified? Are their requirements clearly defined?
 Are decision makers named?
 Is the framework documented?
Plan measurements
 Are plans for making measurement clearly laid out, including:
What is to be measured?
When are measurements to be made?
Where are measurements made?
How are measurements to be made, including calibration routines, data editing, and how a
measurement log is to be used?
How often are measurements to be made?
Who is responsible for measurement, including calibration?
 Are plans for data storage clearly laid out, including:
What data are to be stored?
When are they to be stored?
Where are they to be stored?
How is storage to be accomplished?
How often are data stored? How often are data backed up?
Who is responsible for data storage and data management?
 Are planned analyses and data presentations clearly defined, including:
What analyses are planned?
When are planned analyses conducted?
Where are analyses conducted (i.e., which analytic environment)?
How are planned analyses carried out?
How often are routine analyses made?
Who conducts the planned analyses?
 Is the measurement protocol written?
 Are those who make measurements familiar with the protocol?
 Is the protocol under change management?
Collect and store data
 Is the measurement protocol followed?
 Do data collection plans provide high-quality data?
 Is a measurement log of exceptions maintained?
Analysis, synthesis, present results
 Is the measurement protocol followed?
 Are presentation packages clear and understandable?
 Are results presented in a comprehensive and fair manner?
 Is sufficient time allotted for data exploration?
Data quality program
 Is the data quality program comprehensive? Does it cover all aspects of the measurement system?
 Is the data quality program documented?
FIGURE 9.10 A measurement system designer’s checklist.
While it will still be necessary to make objective measurements, the designer of the measurement
system should also work to ensure that everyone hears the “voice of the customer” (Davis 1987).
Systems of Systems. So far, we have treated our operational, tactical, and strategic examples
as if they were independent of one another. Obviously this is not the case. Ideally, measurements made
at a lower level can be integrated into higher-level measurements, and higher-level measurements provide
context (framework) to help interpret lower-level measurements. We call this property “integrability.”
The customer requirement for a “correct bill” should lead to overall measures of billing
process performance and of the performance of each step. It is possible to achieve a certain amount
of integrability. But as a practical matter, integrability is harder than one might expect, even with
financial measurements.
Manager’s Intuition. Measurement systems can take a manager only so far. Even the best
measurement systems cannot eliminate all risk in decision making. And Lloyd Nelson has noted
that “the most important figures are unknown and unknowable” (Deming 1986). Corporate traditions
are made of the fiercely independent leader who defied the numbers to take the organization
into new and successful directions (although I suspect that many organizations are also long gone
because of similar decisions). Morita’s decision to invest in the Walkman is a famous example of
such a tradition. So clearly, there are times when the decision maker’s intuition must take over.
Managers and organizations must recognize this situation. Certainly decisions supported by data
are preferred, but organizations must not require that all decisions be fully supported. The danger
that an opportunity will pass or a potential crisis will escalate while diligent managers seek data
to support a clear direction is simply too great. Organizations must support individual managers
who take prudent risks. And individual managers should train their intuition.
Just as organizations are adaptive, so too are their systems. Measurement systems must be among
the most adaptive. Day-to-day, they are integral to the functioning of the organization at all levels. It
is not enough for measurement systems to keep pace with change. They also must signal the needs
for more fundamental change. In this section we provide some practical advice for starting and
evolving measurement systems.
1. It is usually better to build on the existing system than to try to start from scratch. All
existing organizations have embedded measurement systems. Even start-ups have rudimentary
financial systems. So one almost never begins with a “clean sheet of paper.” Fundamental business
changes will require new types of measurements (indeed the quality revolution forced
many organizations to expand their focus to include customers) but they will only rarely eliminate
all established measurements or measurement practices. In addition, in times of great
change, the measurement system is an old friend that can be counted on to provide a certain
amount of security.
2. Experiment with new measurements, analyses, and presentations. Learn what others are doing
and incorporate the best ideas into your system.
3. Prototype. It is difficult to introduce new measures. Prototyping helps separate good ideas
from bad, provides an environment in which technical details can safely be worked out, and gives
people needed time to learn how to use them.
4. Actively eliminate measures that are no longer useful. This can be very difficult in some organizations.
But we have already noted that good measurement systems are not oppressive. Similarly,
we have noted the need to create new measures. So those that have outlived their usefulness must be
5. Expect conflicts. We noted in the previous section that it is not usually possible to fully integrate
measurements. Conflicts will arise. Properly embraced, they are a source of good ideas for
adapting a measurement system.
6. Actively train people about new measures, their meaning, and how to use them.
We conclude this section with the Top 10 Measurement System Principles:
1. Manage measurement as an overall system, including its relationships with other systems of the
2. Understand who makes decisions and how they make them.
3. Make decisions and measurements as close to the activities they impact as possible.
4. Select a parsimonious set of measurements and ensure it covers what goes on “between functions.”
5. Define plans for data storage and analyses/syntheses/recommendations/presentations in
6. Seek simplicity in measurement, recommendation, and presentation.
7. Define and document the measurement protocol and the data quality program.
8. Continually evolve and improve the measurement system.
9. Help decision makers learn to manage their processes and areas of responsibility instead of the
measurement system.
10. Recognize that all measurement systems have limitations.
The author wishes to thank Anany Levitin of Villanova University and Charles Redman of Nashville,
TN, whose comments on an early draft of this section led to substantial improvement.
AT&T Engineering and Operations in the Bell System. Members of Technical Staff, AT&T Bell Laboratories
(1982), Holmdel, NJ.
Belanger, B. (1984). Measurement Assurance Programs: Part 1: General Introduction. National Bureau of
Standards, NBS Special Publication 676-1, Washington, D.C.
Carey, M. B. (1994) “Measurement Assurance: Role of Statistics and Support from International Statistical
Standards.” International Statistical Review, vol. 61, pp. 27–40.
Cavanaugh, J. R., R. W. Hatch, and J. L. Sullivan (1976). “Models for the Subjective Effects of Loss, Noise and
Talker Echo on Telephone Connections.” Bell System Technical Journal, vol. 55, November, pp. 1319–1371.
Chambers, J. M., W. S. Cleveland, B. Kleiner, and P. A. Tukey (1983). Graphical Methods for Data Analysis.
Wadsworth International Group, Belmont, CA.
Davis, S. M. (1987). Future Perfect. Addison-Wesley, Reading, MA.
Deming,W. E. (1986). Out of the Crisis. Massachusetts Institute of Technology Center for Advanced Engineering
Study, Cambridge, MA.
Drucker, P. F. (1995). “The Information Executives Truly Need.” Harvard Business Review, vol. 73, no. 1,
January–February, pp. 54–63.
Eccles, R. G. (1991). “The Performance Measurement Manifesto.” Harvard Business Review, vol. 69, no 1,
January–February, pp. 131–137.
Eisenhart, C. (1969). “Realistic Evaluation of the Precision and Accuracy of Instrument Calibration Systems,” in
Ku, H. K., (ed.). Precision Measurements and Calibration—Statistical Concepts and Procedures. National
Bureau of Standards, NBS Special Publication 330-1, Washington, D.C.
Finkelstein, L., and M. S. Leaning (1984). “A review of fundamental concepts of measurement.” Measurement.
vol. 2, January–March, pp. 25–34.
Futrell, D. (1994). “Ten Reasons Why Surveys Fail.” Quality Progress, vol. 27, April, pp. 65–69.
Hauser, R., and D. Clausing (1988). “The House of Quality.” Harvard Business Review, vol. 66, no. 3, May–June,
pp. 17–23.
Hoaglin, D. C., and P. F. Velleman (1995). “A Critical Look at Some Analyses of Major League Baseball
Salaries.” The American Statistician, vol. 49, August, pp. 277–285.
Ishikawa, K. (1990). Introduction to Quality Control. 3A Corporation, Tokyo.
Kaplan, R. S., and D. P. Norton (1992). “The Balanced Scorecard—Measures that Drive Performance.” Harvard
Business Review, vol. 70, no. 1, January–February, pp. 71–79.
Kaplan, R. S., and D. P. Norton (1993). “Putting the Balanced Scorecard to Work.” Harvard Business Review,
vol. 71, no. 5, September–October, pp. 134–149.
Meyer, C. (1994). “How the Right Measures Help Teams Excel.” Harvard Business Review, vol. 72, no. 3,
May–June, pp. 95–103.
Redman, T. C. (1996). Data Quality for the Information Age. Artech, Norwood, MA.
Roberts, F. S. (1979). Measurement Theory with Applications to Decisionmaking, Utility and the Social Sciences.
Addison-Wesley, Reading, MA.
Schlesinger, J. M. (1988). “Ford’s Claims About Quality Show Difficulty of Identifying Top Autos.” Wall Street
Journal, March 14.
Speitel, K. F. (1982). “Measurement Assurance,” in G. Salvendy, Handbook of Industrial Engineering. Wiley,
New York.
Tufte, E. R. (1983). The Visual Display of Quantitative Information. Graphics Press, Cheshire, CT.
Tukey, J. W. (1976). Exploratory Data Analysis. Addison-Wesley, Reading, MA.
Webster (1979). Webster’s New Twentieth Century Dictionary, 2nd ed. William Collins Publishers, Simon &
Schuster, New York.
Fredric I. Orkin
Daniel Olivier
Communications and Computer Systems
Design and Development Plan 10.4
Benefits of Design Control 10.4
Testing Environment 10.5
Buyer Beware! 10.6
Simulation and Fuzzy Logic Quality Tools
Sources of Statistical Software 10.10
In-Cycle Gauging 10.10
Product Reliability and Safety 10.15
Microprocessor Reliability 10.15
Software Reliability 10.15
System Fail-Safe Design Using
Microprocessors 10.15
Fault Tree Analysis Input 10.16
Testing 10.17
The Reliability and Fail-Safe Design
Program 10.18
Everyone dealing with the quality function by now has either faced the computer or retired. The
exponential growth of computer systems will continue unabated for the foreseeable future and invade
all aspects of our jobs. Management and practitioners alike must devote significant effort supplementing
their computer skills and evaluating the risks to their companies associated with change.
The computer revolution has also been a primary force in the expansion of quality activity into
service businesses and functions. Breakthrough technology in communications and data storage has
already begun the transition from powerful stand-alone machines back to large, virtual machines of
unlimited potential. The Internet and its derivatives will spawn this revolution.
This section will include segments of the quality program affected by the computer: (1) communications
and computer systems, (2) design control, (3) testing and validation, (4) software
quality, (5) statistical application, (6) computer-aided inspection, (7) reliability and safety, and
(8) future trends.
Examples and checklists are provided to assist the reader in the application of computer and related
technology. The contributions of computers to quality-related activities are presented throughout this
Handbook. See particularly Section 28, Quality in a High Tech Industry.
Communications and Computer Systems. For purposes of this section, a computer
system is defined as a group of interacting hardware equipment and software that performs an independent
function. Computer systems can be as small as appliance- or automobile-type microprocessor
1In the Fourth Edition, material for the section on computers was supplied by Fredric I. Orkin.
chips with self-contained software or as large as mainframe computers running specialized database
management software. Personal computers running commercial word processing and spread
sheet software programs are generally considered as productivity tools rather than computer systems
unless the commercial software is specifically developed and tested to perform a unique function.
Examples are spread sheet programs that analyze financial, statistical, or engineering data, and word
processing programs that present forms for entry of specialized information on data. Those commercial
software applications are called “templates” and “macros.”
Computer systems have experienced a rapid evolution in recent years, from stand-alone processing
systems dominated by central mainframe processors to widely distributed architectures that provide
almost unlimited communication capabilities. The explosion in popularity of systems such as
the Internet has opened up tremendous new opportunities for information access. Almost any subject
can be researched through the use of the Internet search engines. Systems such as the Internet also
present new quality challenges.
Figure 10.1 lists some of the well-known quality societies and information sources that are currently
accessible via the Internet. Figure 10.2 is a copy of the logo on The World Wide Web site home
page for the American Society for Quality.
Of increasing popularity are “Intranet” applications. Intranet is a private corporate network using
Internet technologies. The Internet and Intranet, combined with the increasing power of on-line document
management systems, have transformed the way that companies communicate and perform
work. The future will be characterized by automated on-line services that provide unlimited access
to and rapid retrieval of information from huge databases of specifications and reference materials.
These new capabilities have several quality implications:
 Increased privacy concerns about access to personal information, such as salary and evaluation data
 Increased security concerns about external access to corporate information, such as financial and
product development data
 Accuracy and currency of information available
Also of concern is the potential for the introduction through these channels of programs that may
cause corruption or destruction of data. The most familiar corrupter is called a “virus.” A virus is
computer code that executes when an infected program is run. This has traditionally meant that only
stand-alone programs (executable files) can be infected. Recently, however, viruses have been inserted
into programs that can allow nonexecutable files to propagate a virus.
The threat of viruses is real. There were about 4000 known viruses at the end of 1994, and the
number of new viruses is growing at a steady pace (McAfee 1996). Viruses can infect data on computers
from mainframes to laptops. Viruses can be distributed across phone lines from bulletin
boards or can enter computers via floppy disks. A virus infection can destroy files; it can cost countless
hours to remove the virus and recover lost data from affected programs.
Quality practices must address the growing virus risk, which is a consequence of the increasing
interconnection of computer systems. Available protection techniques are shown in Figure 10.3.
ISO 9001 is the model for quality assurance in design/development, production, installation, and servicing
promulgated by the International Organization for Standardization. Subclause 4.4, Design Control, spells
out the elements which make a quality system acceptable from the standpoint of assurance of quality in
design and development. These elements of design control also apply to the development of computer
systems and are used by successful companies to assure efficient and high-quality development. Design
control elements include classical engineering activities but are addressed in the process-oriented terminology
consistent with ISO 9000. Design control phases are discussed in terms of process inputs
and outputs for each of the design and development phases: requirements definition; design; code and
debug; unit, integration, and acceptance testing; and operation and maintenance.
American National Standards Institute (ANSI)
American Society for Quality (ASQ)
Association for Computing Machinery (ACM)
Deming Electronic Network
Electronic Industries Association (EIA)
European Standards Institute
Institute of Electrical and Electronics Engineers (IEEE)
International Electrotechnical Commission (IEC)
International Organization for Standardization (ISO)
The Juran Institute
National Institute of Standards and Technology (NIST)
Software Engineering Institute (SEI)
Underwriters Laboratories, Inc. (UL)
FIGURE 10.1 Quality resources available via the Internet.
Password controls to restrict system access only to authorized users
Password procedures that force password uniqueness and periodic changes
Training of personnel on how a virus may be detected and on how controls are used for checking
the integrity of files and disks
Internal procedures to limit the programs that can be entered onto company workstations and networks
Use of virus checking programs to detect the presence of a virus
Diskless workstations to prevent the introduction of foreign files into critical computer systems
Periodic scanning of systems to detect the presence of viruses
Audits of system access and file activity to detect the presence of viruses
Regular system backups using different tapes (an infected tape if reused may be useless) to support
file recovery if data is lost
FIGURE 10.3 Techniques to prevent virus infections.
FIGURE 10.2 Logo for the World Wide Web site home page for the
American Society for Quality (http://www.asq.org).
Design and Development Plan. The first element of design control is an organized plan for
establishing design and development activities. The plan defines the activities to be conducted in
support of the design effort and the parties responsible for each.
Organizational and Technical Interfaces. Further, the plan defines the relationship between
departments, technical disciplines, and subcontractors who are contributing to the design. It also
establishes the documents to be produced by each group and responsibilities regarding review.
Design Input. During the design input phase, the specification of requirements is developed. This
document defines the essential elements of the product and becomes the basis for evaluating when
the product design and development is complete.
Design Output. The design output is the aggregate of all documents, design descriptions, analyses,
test procedures, and test results that are produced during the actual design activities. For hardware
development this may include functional block diagrams, schematics, drawings, and detailed component
specifications. For software the design output may include design descriptions, module hierarchy
charts, object-oriented diagrams, and other representations of the program. (See Section 20 for
a discussion of software development.)
Design Review. Informal reviews conducted by the project team provide quick feedback on specification
compliance, misunderstandings, and errors. Formal reviews that include independent peers are
scheduled milestone events that are often used to separate design phases. Reviews have been shown to
be a more effective method for improving product quality than relying on testing (Wheeler, Brykczynski,
and Meeson 1996). Reviews are important factors in increasing product quality and improving the efficiency
of the design and development process. The most significant benefit from reviews is the early
identification of errors. The earlier in the design process that errors are identified, the less costly they are
to correct (Boehm 1985).
Design Verification. Verification is the incremental checking performed during each phase of the
design and development process to ensure that the process is being followed and that the low-level
derived requirements have been correctly implemented. Verification activities include unit level testing,
analysis studies, integration level testing, and reviews.
Design Validation. Validation is assurance that the product conforms to the defined input requirements
and needs of the customer. Validation must also encompass stress testing in the actual or simulated
customer-use environment.
Design Changes. Changes to the approved design are accomplished with controls similar to those
applied for the initial development process. The changed design must be subjected to the same steps
in the process.
Benefits of Design Control. The addition of reviews during the early project stages will
require scheduled time. Some will complain that valuable code and prototype development time
will be lost. However, history has demonstrated that the investment in reviews is repaid in customer
satisfaction and in reduced rework during testing and reduced errors in maintenance, as shown in
Figure 10.4 (Olivier 1996). The chart, developed by the authors, shows two critical characteristics of
design control reviews.
First is a savings in the effort invested in the development project. The added time required up
front for reviews is represented by the shaded area on the left of the graph. The additional time spent
in rework and refining the product during later stages if reviews are not conducted is shown by the
shaded area on the right. It can be readily seen that the area on the left is smaller than the area on the
right, indicating less effort and, therefore, reduced total cost attributable to design reviews. Errors
found early in the development process are much easier to correct. Studies referenced by Humphrey
(1989) show that AT&T Bell Laboratories found third-party code review inspections to be 20 times
more effective in finding errors than testing for them.
A second benefit of design review as shown in the figure is an increased level of quality. The
quality improvement is demonstrated by the area under the curve shown for the maintenance phase.
It is smaller when design control reviews are conducted, representing fewer errors and less rework.
Freedman and Weinberg (1990) reference a 10-times reduction in the number of errors reaching
each phase of testing for projects with a full system of reviews.
With the increasing complexity and criticality of computer programs, traditional ad hoc methods for testing
have become inadequate. Current computer system design and development models incorporate validation
tests to ensure that system requirements have been satisfied. These models also include verification
review and test techniques as part of the development process (IEEE 1986). Figure 10.5 illustrates the
application of verification and validation to design and development phases (Olivier 1993).
Testing Environment. Testing must ensure that the system operates correctly in the actual
environment or, where such testing is not possible, in an environment that simulates the conditions
of actual use. Stress testing in the actual-use environment is very effective in identifying errors that
may otherwise remain undetected until after product release. Effective techniques to assure correct
operation in the user environment must include “beta”-type testing, where early product versions are
provided for customer-use testing to assure that the system functionality is consistent with the
actual use environment.
Requirements-based testing conditions might not be representative of operational scenarios, and
the requirements themselves might not be error free (Collins et al. 1994). The customers for most
FIGURE 10.4 Impact of design review effort investment on projects. (Source: Olivier 1996.)
computer systems are not knowledgeable enough, nor can they define requirements specifications
precisely enough for validation (Kaner et al. 1993). Empirical data show that the majority of system
failures arise from errors in the requirements themselves, not from failure to satisfy the defined requirements
(Leveson 1995). Systems ultimately may be used in ways not envisioned during the development
process. Errors are experienced as a result of user entry of unexpected values, the use of
unanticipated equipment interfaces, and misinterpretation of reports and displays. These conditions
create a marketing challenge for purveyors of software, who may ignore the challenge and resign
themselves to providing less than perfect product.
Buyer Beware! In purchasing software, the buyer would do well to keep a few caveats in mind:
 Testing for satisfaction of requirements supports the software quality model. However, this approach
alone is no longer practical. Each significant piece of new off-the-shelf commercial software can be
assumed to contain errors, even after thousands of millions of executions (Collins et al. 1994).
 While fitness for use recognizes the importance of the satisfaction of defined requirements, there
may be errors in the defined requirements themselves.
 Freedom from errors is only one of many attributes which add up to fitness for use.
The success of the Microsoft Corporation and other software developers is based on the realization
that customers will tolerate bugs in software as a trade-off for timely product enhancements,
increased functionality, and low cost. This is in direct contrast to the reactions of the same customers
to errors experienced in hardware quality where zero-defect performance is expected.
For many software developers, a result of these customer attitudes is that structured design is
subordinated to early coding and arduous testing. At Microsoft, there is an attempt to accelerate the
release of new products by investing tremendous hours in early coding and testing and subsequent
debugging. A reference to the Windows NT development process during the last week of April 1993
showed the fixing of 1132 bugs while 713 new bugs serious enough to warrant testing were also
discovered (Wall Street Journal 1993).
The unending pressure to increase the complexity of emerging computer and software systems is
forcing the industry to rethink the cost and risk of this “brute force” strategy. In the authors’ opinion,
this complexity will dictate the gradual development of new software development tools and a
shift by successful developers back to more structured design techniques.
Techniques available today that can be used to increase the effectiveness of the testing program
for computer systems include:
FIGURE 10.5 Verification and validation during design and development
phases. (Source: Olivier 1993.)
 Training development personnel on effective review and testing techniques. Testers must understand
that the goal of testing is to identify as many errors as possible and not to test a program to
show that it works (Meyers 1979).
 Analyzing errors to identify the root cause. Error analysis leads to more effective review techniques
and checklists that focus on avoiding historical error causes.
 Designing systems to support testability. This means including diagnostic routines and errorconditions
logging, as well as designing tools in conjunction with the design and development
of the core system functions.
 Tracking the number and severity of errors found per unit test time.
Quality software programs exhibit certain attributes across programming languages and applications.
A list of some of these key attributes is provided in Table 10.1 (McCall et al. 1977). These attributes
include both subjective and objective measures and are only a subset of the total attributes of software
programs. The applicable elements should be selected by the quality professional for application to
each specific program.
Many application programs today are purchased off the shelf. For these programs quality assessment
is different than for programs developed internally. It is often worthwhile to contact the vendor who
developed the off-the-shelf program and current users to gain insight into the quality of the program. The
authors have prepared a list of suggested steps for the evaluation and selection of off-the-shelf software
programs (see Figure 10.6).
If the purchased programs are to be modified by in-house personnel, other information should also
be requested, including: design documentation, source code, test procedures, and support utilities used
for development and testing.
If software programs are developed internally, the developers should prepare a quality assurance
plan to identify specific activities that are to be performed. Such actions should include those shown
in Figure 10.6, as well as:
TABLE 10.1 Software Quality Attributes
Quality attribute Description
Correctness Extent to which a program satisfies its specifications and fulfills the user’s mission
Reliability Extent to which a program can be expected to perform its intended function with required
Efficiency Amount of computing resources and code required by a program to perform a function
Integrity Extent to which access to software or data by unauthorized persons can be controlled
Usability Effort required to learn how to operate, prepare input, and interpret output of a program
Maintainability Effort required to locate and fix an error in an operational program
Testability Effort required to test a program to ensure that it performs its intended function
Flexibility Effort required to modify an operational program
Portability Effort required to transfer a program from one hardware configuration and/or software
system environment to another
Reusability Extent to which a program can be used in other application—related to the packaging and
scope of the functions that programs perform
Interoperability Effort required to couple one system with another
Source: McCall, Richards, and Walters (1977).
 Internal audits to assure that internal design procedures are being followed
 Process and product quality measures
 Oversight of subcontractor activities
 Management of the configuration of the code and documentation
 Tracking of open program bugs and corrective actions
 Analyzing program bugs to identify process improvement opportunities
 Final release approval
Simulation and Fuzzy Logic Quality Tools. Simulation modeling as a computer tool is
gaining acceptance in many fields. It involves the use of mathematics to replace and predict physical
test results. “By altering the input variables of a simulation model, different configurations can
be created and, through successive experiments, each can be evaluated by analyzing the output variables
designed to measure performance” (Smith 1994).
Finite element analysis software, combined with design of experiments (DOE), can result in a
powerful quality engineering tool. Rizzo reports that simulation models were used to reveal
a nonlinear buckling behavior of a short segment of the power conductor of an electronic system.
Engineers were fascinated to find that their initial judgment was proven wrong by the model (Rizzo
1994). Chalsma also reports that “computer modeling and simulation have pushed much automobile
testing back into the design stage” (Chalsma 1994).
“The effort to make computer-based systems more intelligent and more human certainly is centered
on the use of fuzzy logic” (Chen 1996). “Fuzzy logic” is the application of mathematics (set theory) to
represent and manipulate data that possess nonstatistical uncertainty.
Fuzzy logic represents a new way to deal with the vagueness of everyday life and has a major
future role in product and process development (Bezdek 1996).
Zadeh (1965) observed that, as a system becomes more complex, the need to describe it with precision
becomes less important. Subjective descriptors, such as hot, cold, near, far, soon, and late enter
the design equations of control theory via fuzzy logic.
Traditional control systems are open- or closed-loop (Figure 10.7). They collect input data,
perform computation, and provide an output. Closed-loop systems use feedback of the output function
in the computation. With fuzzy systems, the input is “fuzzified” by placing it into overlapping
groups or sets—for example, cool, warm, or hot—rather than into the set characterized by a specific
temperature or range of temperatures. A feedback control system then uses a controller made up of
fuzzy-logic components (if-then rules, variables, sets, and translation mechanisms) that describe the
desired control. The results are then “defuzzified” and processed to the output (Zadeh 1965).
Figure 10.8, adapted from Miller (1996), shows the block diagram of a typical fuzzy-logic control
system. Specialized electronic hardware can handle from 800 to 8,000,000 rules per second.
Fuzzy-logic controllers are found in such wide-ranging applications as home appliances, nondestructive
testing, machine tool control, and inspection systems (see Figure 10.9).
1. Identify present and future requirements.
2. Survey available packages.
3. Examine documentation and manuals.
4. Determine data and communication interoperability with existing programs.
5. Survey existing users as to product acceptability. (Internet queries provide an optimal way
to obtain this information.)
6. Request quality assurance and testing information from the developer.
7. Request a list of known bugs.
8. Review copyright and licensing requirements.
9. Obtain programs for internal trial execution.
10. Negotiate contract to include maintenance services and upgrades.
FIGURE 10.6 Steps in acquiring an application package.
Industrial application of fuzzy logic is on an exponential growth curve. Applications include feedforward
controls for industrial machine tools, robotic systems, medical devices, and process equipment.
Kierstan (1995) reports of application of fuzzy logic systems to temperature control in automated
food processing clean rooms with extreme temperatures, environment control in microbe hostile
atmospheres, and monitoring the size and shape of pizzas!
One of the first, and still important, uses of the computer in quality control is for statistical analysis
(Besterfield 1979). Most statistical techniques discussed in this handbook may be programmed
by using one of the standard programming languages. “The Shewhart Chart is a child
of the pre-computer age, when all record-keeping was done manually. With today’s off-the-shelf
Input Output
Input Output
FIGURE 10.7 Traditional control system. (Source: Zadeh 1965.)
Fuzzifier Defuzzifier
FIGURE 10.8 Fuzzy logic block diagram. (Reprinted with permission from Electronics Now magazine,
May 1996 issue. © Gernsback Publications, Inc., 1998.)
Washing machines that sense the degree of dirt in the wash water and optimize the cycle
Video cameras with electronic image stabilizers that remove jitter caused by hand movement
Air conditioners that sense the number of occupants and thermal load as well as environmental conditions
of temperature and humidity
Refrigerators that reduce heat shock when the door is opened and new food added
Vacuum cleaners that sense the dirt load and control the cleaning cycle
Wind power rotor pitch control
Transit system speed control
Soda can printing defect inspection
Automobile brake and transmission control
computing capabilities, the total technology system needed can readily be put together.”
(Marquardt 1984.)
Sources of Statistical Software. Quality Progress annually publishes commercial sources
for software. The 1996 issue lists 183 companies that supply statistical software products covering
(Struebing 1996):
 Capability studies
 Design of experiments
 Statistical methods
 Statistical process control
Rapid acceptance of computer-aided design (CAD), computer-aided engineering (CAE), and computeraided
manufacturing (CAM) provides the database and information structure for the introduction of
computer-aided inspection and computer-aided test. Cost incentives are also expected to accelerate the
development of inspection and test automation to keep pace with major gains in manufacturing
Reduction of defect levels to the parts-per-million range often requires computer-aided technology.
Many industries are increasingly accepting inspection systems that are integrated with automated
manufacturing systems. “This step completes the computer-integrated manufacturing (CIM)
loop” (Reimann and Sarkis 1993).
Generally, automatic inspection will couple a transducer to a computer. Transducers can take the
form of dimensional position indicators or indicators of physical effects such as force, flow, vibration,
electrical properties, and magnetic properties. An American National Standards Institute (ANSI)
standard for integrating the CAD and dimensional measuring instruments was published in 1990
(ANSI/CAM-I 1990).
The multitude of potential applications for automated inspection and the equipment and computer
functions related to them are detailed in Table 10.2. This table, developed by the authors, should prove
useful as a checklist for potential project ideas.
In-Cycle Gauging. The use of touch-trigger probes to take measurements during a numerically
controlled machine-tool cycle is called “in-cycle gauging.” As reported by Reimann and Sarkis
(1993), recent developments in this technology have led to the availability of new techniques that can
compensate for machine variances and develop parametric probe-path data.
FIGURE 10.9 Applications of fuzzy logic spark the imagination.
TABLE 10.2 Potential Applications for Automated Inspection
Industry applications Equipment type Transducer type Computer function
Dimensional gauging Automatic high-speed, noncontact video Optical, laser, video, solid-state camera Video image processing; autofocus; mass storage
inspection, and optical comparators for uninterrupted cycle execution; part and table
multiple-axis servo positioning; inspection of
unaligned parts
Coordinate measurement machine Touch probe Geometrical tolerance programming, tolerance
analysis, data handling, multiple probe calibration,
laser calibration, math processing, contouring,
operator prompting, editing, feedback, accept/
reject decision
Computer-assisted gauging (lab) Touch probe, electronic, air Supervised prompting, automatic mastering,
magnification set, zeroing display, statistics, counting,
spec comparison, diagnostic testing
Electronic gauges and measuring Calipers, micrometers, snap gauges, Direct digital output, gauges to host computer
systems with computer interface bore gauges, indicator probes, height through interface
gauges, air gauges, ultrasonic gauges,
magnetic gauges, etc.
In-cycle gauging on numerical Touch probe On machine measurements, tool wear
control (NC) machines compensation, temperature compensation automatic
check of tool offset, work location, table and spindle
Bench laser micrometer Laser Automatic laser scan, data handling, statistical
dimension calculations, part sorting, accept/reject
Holography Laser Automatic stress, strain, displacement, image
TABLE 10.2 Potential Applications for Automated Inspection (Continued)
Industry applications Equipment type Transducer type Computer function
Laser interferometer Laser Automatic temperature and humidity compensation
data handling and storage, math processing
3-D theodolite, coordinate, Optical Interactive operator prompting, automatic angular
measurement measurement, data handling
Scanning laser acoustic Laser, acoustic Beam scanning, data processing
microscope (SLAM)
Electrical and electronic Temperature measurement Thermocouple, thermistor, resistance Calibration; data acquisition, analysis, and processing
instrumentation temperature detector (RTD)
Robotic-printed circuit board test Electronic Robot control, fully automatic board test
Weight and balance, filling Electronic Automatic tare, statistical processing, data recording
and packaging, inspection
Circuit analyzers Electronic Special-purpose test systems
Automatic test equipment All Special-purpose test systems with complete
functional testers real-time input, processing and output data
Cable testers Electrical Automated harness continuity and high-potential
Sem/iconductor testers Automated test of standard and special-purpose
Lab devices and equipment Chromatographs Optical Fully automatic preprogrammed sampling and data
Strength of materials Probe, force, displacement, Preprogrammed cycle operation; data, chart, and
strain gauge graphic output records; multichannel recording;
on-line data processing
Hardness testing Probe Robotic, fully automatic testing and recording,
results analysis, and prediction
Analyzers All Automatic calibration, testing, and recording
Electron microscopes Electromagnetic Processing and materials analysis, preprogrammed
for failure analysis
Optical imaging Video borescope, fiber-optic inspection Optical Digital data image processing documentation
Photographic Optical Fully automatic strobe, photographic sequencing
and processing
Video microscopes Optical Video image processing data documentation
High-speed video recording Optical Automatic 200–12,000 frames per second
stop-motion recording of machine and manual
processes; motion analysis; data processing
Environmental and Test chamber controls Temperature, humidity, altitude Preprogrammed cycle controls, time and
functional test equipment data records
Leak detection Vacuum, gas, acoustic Automatic zeroing, built-in calibration, automatic
sequencing, tolerance checking, data processing
and display
Shock and vibration testing Accelerometer Automatic cycle control, built-in calibration, data
logging and display
Built-in equipment Electrical, electronic Preprogrammed part and system functional and
environmental cycling, recording
EMI measurement Electronic, magnetic Data processing, math analysis, recording
Materials testing equipment Surface and roughness measurement Stylus follower, air flow Operator prompting, data analysis
Coating thickness, sheeting thickness Electronic, video, ultrasonic, Calculation and math processing; display; selfbeta
backscatter calibration; automatic filter changing and positioning;
prompting self-diagnostics; feedback;
accept/reject decision
TABLE 10.2 Potential Applications for Automated Inspection (Continued)
Industry applications Equipment type Transducer type Computer function
TABLE 10.2 Potential Applications for Automated Inspection (Continued)
Industry applications Equipment type Transducer type Computer function
Paper, plastic, and coated product process Laser Automatic high-speed processing, feedback
inspection for holes, particulates, controls, data analysis, and alarms
streaks, thickness
Nondestructive test equipment Magnetic particle, eddy current Probe Self-regulation, calibration, data handling,
defect recognition
Ultrasonic flaw detection Sonic, vibration Automated quantitative analysis, curve
matching, automated procedures, graphics data
acquisition and storage
Scanning laser acoustic microscope Laser, acoustic Beam scanning, data processing, flow detection
(SLAM) flaw detection
X-ray, fluoroscopic Optical, electronic Automatic calibration, operator prompting, data handling,
statistics, stored programming, defect
Acoustic emission Acoustic Independent channel monitoring and display, linear,
zone location, tolerance comparison, preprogrammed
tests, graphics output, triangulation, source location
Infrared test systems Optical, video Calibration, system control
Radiographic, gamma Optical, gamma Programmable, automatic, self-diagnostic, safety
malfunction interrupts, automatic defect recognition,
robotic part handling, automatic detection of
missing parts
Computer-aided tomography (CAT) X-ray Data acquisition, processing, interpretation and
Nuclear magnetic resonance Magnetic Data acquisition, processing, interpretation and
(NMR) scanner imaging
Product Reliability and Safety. The incorporation of microprocessors in a product under
development changes the way in which the quality, reliability, and safety design objectives are defined,
managed, and analyzed. Products that utilize microprocessors require an additional category consisting
of program software, which adds a unique dimension to the problem of design analysis. While methods
have evolved to analyze, predict, and control the reliability of conventional systems, software as an
entity has an altogether different character, presents greater difficulties, and must be treated separately.
Even though the software aspect of the microprocessor is more difficult to analyze, the product
objectives of a fail-safe design and self-testing system are easier to achieve when the design incorporates
microprocessors. Some products, because of the type of application for which they will be
used, have higher criticality than do others. The level of design effort that goes into each type of
product must be commensurate with the application. The following discussion focuses on the reliability
and quality of the microprocessor itself as an electronic device, the reliability of the software
program associated with the processor, and the design advantages that the microprocessor allows.
Microprocessor Reliability. Microprocessors and peripheral devices are supplied by many
manufacturers. Typical device costs vary with the number of instructions, execution time, and underlying
technology. The quality level of these hardware devices has a direct effect on the reliability of
the computer system. They can be procured at specified levels of quality as defined by MIL-M-
38510D (1977) and MIL-STD-883C (1983), which are keyed to the models for predicting reliability
and failure rate that are given in MIL-HDBK-217F (1991). These specifications define levels of
qualification testing, screening tests, and burn-in.
Commercial microprocessors have also demonstrated excellent performance, but testing is not
documented. However, as a minimum, they generally undergo incoming inspection, visual inspection,
electrical parameter testing, and burn-in.
Software Reliability. Built-in software is an element of the total product that governs computational
or control functions. Each step of the program development process is inherently error-prone. Errors
can be introduced from misinterpretation, mistakes in translation, or mistakes in coding. Unless special
attention is given to each step, a large number of errors can be introduced. It is most cost-effective to detect
and eliminate errors as early in the program development process as possible.
Figure 10.10 shows the cost of correcting an error relative to the phase in which the error is
detected. The figure indicates that an error corrected during operation can cost over 75 times as much
as if it had been corrected during preliminary design (Anderson and Martin 1982).
Software reliability is more difficult to measure than the reliability of hardware. Hardware reliability
tends to follow a “bathtub curve” that represents a high initial failure rate and then a constant low
rate until reaching the end of product life. Software does not wear out and reliability is a function of error
removal. Software errors in the program may be coding errors or the result of misunderstanding of the
requirements. Reliability prediction measures include tracking of the rate at which discrepancies are
found and the number of discrepancies (bugs) that remain open. As long as the rate of bug removal
exceeds the rate of bug introduction, the rate of new bugs detected will decrease over time. This relationship
has led to prediction of software reliability through tracking of the error detection rate to predict
the level of software reliability. This relationship is graphically shown in Figure 10.11.
Although zero errors in software prior to release may be a desirable goal, for many large software
development efforts this may not be achievable with current tools. If all errors cannot be eliminated
prior to release, it is essential that safety and critical performance errors be corrected.
System Fail-Safe Design Using Microprocessors. The incorporation of a microprocessor
into the design of a product permits greater freedom in designing the system. Failures of
single circuit components can be detected by the microprocessor before the effects of those failures
produce a hazard to the user. Techniques such as watchdog timers, drift analysis, and redundant voting
circuits can place the system in a predicted, safe configuration before shutdown.
Quality analysis techniques such as hazards analysis and failure mode, effect, and criticality
analysis (FMECA) are important “preventive” tools (see Section 3, The Quality Planning Process;
Section 19, Quality in Research and Development).
Fault Tree Analysis Input. Fault tree analysis (FTA) is another technique that can be used to
identify potential safety risks. An FTA is a top-down analysis technique aimed at identifying causes
or combinations of causes that can lead to an unsafe condition. The analysis is primarily qualitative,
which makes FTA an excellent complement to FMECA. While FMECA is especially effective in
identifying potential failure conditions and predicting system reliability, the primary strength of the
FIGURE 10.10 Relative cost of error correction. (Source: Anderson and Martin
FIGURE 10.11 Predicting software reliability
on the basis of errors discovered.
FTA is the ability to predict potential unsafe conditions that may result due to operator errors, misuse,
or abuse conditions. Often, unsafe conditions may result from unforeseen use of the system
rather that from a system failure.
In addition to established analysis techniques such as FMECA and FTA, there are several expected
safety risk mitigation controls that should be implemented. These controls include the steps shown in
Figure 10.12 (Olivier 1995).
Application of these and other controls is not only part of good engineering practices but is also
an essential element of an effective safety program.
Several regulatory standards bodies, such as the International Electrotechnical Commission (IEC)
and United Laboratories (UL), have developed testing laboratories and published standards regarding
procedures and controls necessary to reduce personnel and property hazards. [See IEC 601-1
(1988), IEC 1025 (1990), and UL 544 (1993).] These standards take into consideration a survey of
known existing standards and are based on the inputs from a wide variety of users and professional
organizations. The standards are frequently updated to remain current with social and technological
advances. The standards cover a wide range of conditions that can affect product safety ranging from
electrical standards and electromagnetic compatibility requirements to cautions and warning notices.
Both IEC and UL publish a list of current standards and can be contacted via the Internet.
Compliance with IEC and UL standards presents a baseline for establishing levels of safeness for products.
Requirements for electrical systems, safe levels of radiation, and practices for labeling as defined in
these standards should be adopted as a minimum level for products such as medical devices. Verification
that the requirements defined by these standards have been realized can be tested through laboratories.
The safety risk analysis program should recognize the dynamic nature of the market. A system that
was intended for a very specific application may be used for more general purposes, and new potential
safety risk conditions may be experienced. As a consequence of the evolution of a system, the safety risk
analysis program must also evolve and continue throughout the life of the system.
Testing. Although somewhat tedious, it is also possible to verify, through test of the final design,
that certain undesirable failure modes cannot produce a hazardous result. Such test methods involve
the systematic simulation of an internal fault and observation of the response of the system.
Electronic systems including microprocessors are susceptible to environmental threats, which
require different types of protection. Electromagnetic interference (EMI) and radio frequency interference
(RFI) are present in most operating environments. These effects are not always easy to measure
and trace to their source. Both EMI and RFI can produce transient, unpredictable behavior,
which even the most thorough fail-safe circuit design cannot manage. Forethought must be given to
the methods by which the system will be shielded and protected from these effects.
Because of the often nonreproducible nature of EMI/RFI-induced failures, such problems could
easily be thought to exist in the software logic of the microprocessor. Therefore, EMI/RFI must be ruled
out before an unrewarding search through the software is begun. Electrostatic discharge, temperature,
humidity, and vibration also need to be considered.
Checking the status of hardware on start-up
Monitoring hardware equipment during run time
Data range checking to reduce the likelihood of operator entry errors
Defining system fail-safe states in the case of failures
Installing security controls to restrict access to authorized users
Providing safety margins for critical tolerances
Low-level testing and review of safety related functions
Complying with regulatory standards (UL and IEC)
External laboratory testing
FIGURE 10.12 Safety check controls. (Source: Olivier 1995.)
The Reliability and Fail-Safe Design Program. The amount of effort that goes into a
product to ensure that it is fail-safe and reliable depends on the criticality of its use and its inherent
complexity. A product can be graded into four categories, as shown in Figure 10.13.
Low Complexity and Noncritical. At the low end of the product scale are devices that are relatively
simple and noncritical in terms of their application. Many simple commercial products fall
into this category.
Low Complexity and Critical. Also at the low-complexity end of the product scale are those
devices that are relatively simple but absolutely critical. An example is a medical syringe, which
must be sterile and must be sharp; these are not complex characteristics, but they are critical. Other
examples are an electrical fuse or a lug nut for an automobile wheel.
High Complexity and Noncritical. For systems that have high complexity and low criticality,
emphasis must be placed on ensuring that the reliability is sufficient to the needs of the application.
The reliability emphasis for complex systems is due to the inverse relationship between complexity
and reliability. Unless adequate reliability procedures and programs are followed, complex systems
are likely to perform disappointingly with respect to mean time between failure.
High Complexity and Critical. At the high end of the scale are microprocessor systems used in
aircraft control systems and patient-connected medical devices. These types of device are both complex
and critical, and their development requires a great deal more attention to the details of design
and the procedures by which the designs are brought forward. For products that are both complex
and critical in operation, the use of a microprocessor simplifies some problems but introduces
others in the validation of the software.
The destruction of a European Space Agency Ariane 5 rocket only 39 seconds after launch on its
maiden flight illustrates the rigor required for this category. One line of code, without benefit of error
trapping and graceful recovery, shut down both of the guidance system computers. “One bug, one
crash. Of all the careless lines of code recorded in the annals of computer science, this one may stand
as the most efficient” (Gleick 1996).
Higher Safety Requirements
Higher Reliability Requirement
FIGURE 10.13 Evaluation matrix: complexity versus criticality.
Matrix of Categories. Table 10.3 displays a matrix of program approaches and procedures that can
be applied to enhance reliability for the four categories of complexity and criticality. (Note that Table
10.3 is divided into the same four categories as Figure 10.13). These procedures can be applied more
or less stringently and in appropriate combinations in order to meet the development effort objectives.
Although the future is impossible to predict precisely, one thing is certain: Computer systems will
continue to revolutionize the definition of quality practices. Some current trends include:
 Data from disparate quality tracking systems will be increasingly integrated to provide systemwide
TABLE 10.3 Recommended Tasks for Product’s Complexity and Criticality
 With the speed of new computer systems doubling every 18 months for the foreseeable future, the
availability of this steadily increasing computer system capability will lead to more extensive
automation within businesses of every type (Scientific American 1995).
 The cost of scrap, rework, warranties, and product liability will impart continuing importance to
monitoring of the system, the process, and the machines that assure quality of output (McKee 1983).
 Paperless factories and automated document management systems will emphasize automated
analysis tools and document control. [The cost of copying paper is approximately $0.05 per page
as against $0.00021 per page for electronic copies (Gates 1995).]
 Evaluation of the effectiveness of software quality systems will become an increasing responsibility
of the quality professional.
 “As documents become more flexible, richer in multimedia content, and less tethered to paper, the ways
that people collaborate and communicate will become richer and less tied to location” (Gates 1995).
 Company networks will reach beyond employees into the world of suppliers, consultants, and customers
(Gates 1995).
A 1995 survey by IEEE asked experts in academia and industry to share their vision of software’s
future. Significant to the quality professional are predictions of:
 “Parallel processing and networked computing,” which will continue the exponential expansion of
computer power into the foreseeable future. Factory automation will follow a parallel path.
 “Object-oriented technology integrated into a `Business Information Navigator’ application,”
which will tie internal and worldwide external databases together. Fuzzy logic search routines will
retrieve data and answer questions posed in sentence form.
 “Conferencing of data and video as the work force moves to telecommuting.” Paperless systems
and video communications will allow much of your work to be done at home.
 “Information overload.”
The reliability and safety engineering work of James Wingfield, Ph.D., is recognized and appreciated.
Anderson, R. T., and Martin, T. L. (1982). Guidebook for Management of Software Quality. Reliability
Technology Associates, Lockport, IL, P. 8.
ANSI/CAM-I (1990). Standard 101-1990, Dimensional Measuring Interface Specification. Computer Aided
Manufacturing—International, Inc., Arlington, TX.
Besterfield, Dale H. (1979). Quality Control. Prentice-Hall, Englewood Cliffs, NJ.
Bezdek, James C. (1996). “A Review of Probabilistic, Fuzzy, and Neural Models for Pattern Recognition,” in
Fuzzy Logic and Neural Network Handbook, Chen, ed. McGraw-Hill, New York, pp. 2.1–2.33.
Boehm, Barry W. (1985). “Verifying and Validating Software Requirements and Design Specifications.” IEEE
Software, January, pp. 471–472.
Chalsma, Jennifer K. (1994). “Better tests mean better cars.” Machine Design, vol. 66, no. 1, January 10, pp. 36–42.
Chen, ed. (1996). Fuzzy Logic and Neural Network Handbook. McGraw-Hill, New York, pp. 2.1–2.33.
Collins,W. Robert, Miller, Keith W., Spielman, Bethany J., Wherry, Phillip (1994). Communications of the ACM,
vol. 37, no. 1, pp. 81–91.
Freedman, Daniel P., and Weinberg, Gerald M. (1990). Handbook of Walkthroughs, Inspections, and Technical
Reviews: Evaluating Programs, Projects, and Products. Dorset House, New York.
Gates, Bill (1995). The Road Ahead. Viking Penguin, pp. 63, 119.
Gleick, James (1996). “Little Bug, Big Bang,” New York Times Magazine, December 1.
Humphrey, Watts S. (1989). Managing the Software Process. Addison-Wesley, Reading, MA, p. 187.
IEC 601-1 (1988). IEC 601-1: 1988 Medical Electrical Equipment, Part 1: General Requirements for Safety, 2nd
ed. International Electrotechnical Commission, Geneva.
IEC 1025 (1990). Fault Tree Analysis, 1st ed., 1990-10, p. 11. International Electrotechnical Commission,
IEEE (1986). IEEE Standard 1012-1986, Software Verification and Validation Plans. Institute of Electrical and
Electronic Engineers, New York, pp. 9–12.
Kaner, Cem, Falk, Jack, and Nguyen, Hung Quoc (1993). Testing Computer Software, 2nd ed. International
Thomson Computer Press, p. 59.
Kierstan, Mark (1995). “Food Hygiene, Quality and Safety: Toward the Year 2000.” British Food Journal, vol.
97, no. 10, pp. 8–10.
Leveson, Nancy G. (1995). Safeware: System Safety and Computers. Addison Wesley, Reading, MA, p. 157.
Marquardt, D. (1984). “New Technical and Educational Directions for Managing Product Quality.” The American
Statistician, vol. 38, no. 1, February, pp. 8–13.
McAfee (1996). “An Introduction to Computer Viruses and other Destructive Programs.” McAfee,
McCall, J., Richards, P., and Walters, G. (1977). “Factors in Software Quality—Concepts and Definitions of
Software Quality.” Joint General Electric and U.S. Air Force Report No. RADC-TR-77-369, vol. 1, November,
pp. 3–5.
McKee, Keith E. (1983). “Quality in the 21st Century.” Quality Progress, vol. 16, no. 6, June, pp. 16–20.
Meyers, Glenford J. (1979). The Art of Software Testing, Wiley Interscience, New York, p. 5.
MIL-HDBK-217F (1991). Reliability Prediction for Electronic Equipment. U.S. Department of Defense,
Washington, DC. Available from Department of Defense and NASA, Washington, DC 20301.
Miller, Byron, (1996). “Fuzzy Logic.” Electronics Now, vol. 67, no. 5, pp. 29–30.
MIL-M-38510D (1977). General Specifications for Microcircuits. U.S. Department of Defense,Washington, DC.
Available from Department of Defense and NASA, Washington, DC 20301.
MIL-STD-883C (1983). Test Methods and Procedures for Microcircuits. U.S. Department of Defense,
Washington, DC. Available from Department of Defense and NASA, Washington, DC 20301.
Olivier, Daniel (1993). “Required Documentation for Software Validation.” Medical Device and Diagnostic
Industry, July.
Olivier, Daniel P. (1995). “Software Safety: Historical Problems and Proposed Solutions.” Medical Device and
Diagnostic Industry, July.
Olivier, Daniel P. (1996). “Implementation of Design Controls Offers Practical Benefits.” Medical Device and
Diagnostic Industry, July.
Reimann, Michael D., and Sarkis, Joseph (1993). “An Architecture for Integrated Automated Quality Control.”
Journal of Manufacturing Systems, vol. 12, no. 4, pp. 341–355.
Rizzo, Anthony (1994) “Diagrams.” Mechanical Engineering, vol. 116, no. 5, pp. 76–78, May.
Scientific American (1995). “Microprocessors in 2020.” Scientific American, September, pp. 62–67.
Smith, D. L. (1994). “The Leisure Industry.” Service Industries Journal, vol. 14, no. 3, pp. 395–408, July (ISSN
Struebing, Laura (1996). “Quality Progress’ 13th Annual QA/QC Software Directory.” Quality Progress, vol. 29,
no. 4, pp. 31–59, April.
UL 544 (1993). Standard for Medical and Dental Equipment, UL 544, 3rd ed. Underwriters Laboratories, Inc.,
Northbrook, IL, August 31.
Wall Street Journal (1993). “Climbing the Peak: 200 Code Writers Badgered by a Perfectionist Guru, Bring Forth
Windows NT.” Wall Street Journal, May 26, p. A12.
Wheeler, David A., Brykczynski, Bill, and Meeson, Reginald N., Jr., (1996). Software Inspections: An Industry
Best Practice. IEEE Computer Society Press, pp. 10–11.
Zadeh, L. A. (1965). Information Control, vol. 8, pp. 338–352.
Donald W. Marquardt
Role in Facilitating International
Trade 11.1
External Driving Forces 11.2
Internal Response to the External
Forces 11.4
The “Separate and Complementary”
Concept 11.4
Characteristics of ISO 9000 Standards
The Clauses of ISO 9001 and Their
Typical Structure 11.5
VISION 2000 11.8
Generic Product Categories 11.8
Acceptance, Compatibility, and
Flexibility 11.9
Avoiding Proliferation 11.9
BEYOND VISION 2000 11.10
Origin of the Need 11.11
Activities 11.11
Accreditation-Level Activities 11.12
Accreditation and Registration Flowchart,
Including Related Activities 11.12
Mutual International Acceptance 11.13
Formal International Mutual
Recognition 11.13
FAMILY 11.15
Problems and Opportunities 11.15
The Scope of Standardization 11.15
The Concept of Continuous
Improvement 11.15
The Role of Continuous Improvement in
ISO 9001 11.16
The Role of Statistical Techniques 11.18
Interpretations of the Standards 11.19
Defining the Scope of
Certification/Registration 11.19
Alternate Routes to
Certification/Registration 11.21
Industry-Specific Adoptions and
Extensions of ISO 9000
Standards 11.21
Other Areas of Application 11.24
(TQM) 11.24
Relationship to ISO 9001 11.25
Relationship to ISO 9004 11.25
Role in Facilitating International Trade. The ISO 9000 standards exist principally to
facilitate international trade. In the pre-ISO 9000 era there were various national and multinational
quality system standards. These were developed for military and nuclear power industry needs, and,
to a lesser extent, for commercial and industrial use. These various standards had commonalities
and historical linkages. However, they were not sufficiently consistent in terminology or content for
widespread use in international trade.
The ISO 9000 standards have had great impact on international trade and quality systems implementation
by organizations worldwide. These international standards have been adopted as national
standards by over 100 countries and regional groups of countries. They are applied in a wide range
of industry/economic sectors and government regulatory areas. The ISO 9000 standards deal with
the management systems used by organizations to design, produce, deliver, and support their products.
The standards apply to all generic product categories: hardware, software, processed materials,
and services. Specific ISO 9000 family standards provide quality management guidance, or quality
assurance requirements, or supporting technology for an organization’s management system. The
standards provide guidelines or requirements on what features are to be present in the management
system of an organization but do not prescribe how the features are to be implemented. This nonprescriptive
character gives the standards their wide applicability for various products and situations.
The ISO 9000 family does not deal with any technical specifications for a product. The ISO 9000
standards for an organization’s management system are complementary to any technical specifications,
standards, or regulations applicable to the organization’s products or to its operations.
The standards in the ISO 9000 family are produced and maintained by Technical Committee 176
of the International Organization for Standardization (ISO). The first meeting of ISO/TC176 was
held in 1980. ISO 8402, the vocabulary standard, was first published in 1986. The initial ISO 9000
series was published in 1987, consisting of:
 The fundamental concepts and road map guideline standard ISO 9000
 Three alternative requirements standards for quality assurance (ISO 9001, ISO 9002, or ISO 9003)
 The quality management guideline standard ISO 9004
Since 1987, additional standards have been published. The ISO 9000 family now contains a variety
of standards supplementary to the original series, some numbered in the ISO 10000 range. In particular,
revisions of the basic ISO 9000 series, ISO 9000 through ISO 9004, were published in 1994.
This section is written in relation to the 1994 revisions. Table 11.1 lists the standards published as of
the beginning of 1996. Additional standards are under development.
ISO 9001, ISO 9002, and ISO 9003 have been adopted and implemented worldwide for quality
assurance purposes in both two-party contractual situations and third-party certification/registration
situations. ISO 9001 and ISO 9002 together have predominant market share in this segment. Their
use continues to grow, as does the infrastructure of certification/registration bodies, accreditation
bodies, course providers, consultants, and auditors trained and certified for auditing to these standards.
Mutual recognition arrangements between and among nations continue to develop, with the
likelihood of ISO-sponsored quality system accreditation recognition in the near future. The number
of quality systems that have been certified/registered worldwide now exceeds 100,000 and continues
to grow.
The periodic surveillance audits that are part of the third-party certification/registration arrangements
worldwide provide continuing motivation for supplier organizations to maintain their quality
systems in complete conformance and to improve the systems to continually meet their objectives for
The market for quality management and quality assurance standards is itself growing, partly in
response to trade agreements such as European Union (EU), General Agreement on Tariffs and Trade
(GATT), and North American Free Trade Association (NAFTA). These agreements all are dependent
upon standards that implement the reduction of nontariff trade barriers. The ISO 9000 family occupies
a key role in the implementation of such agreements.
Certain industry/economic sectors are developing industry-wide quality system standards, based
upon the verbatim adoption of ISO 9001, together with industry-wide supplemental requirements.
The automotive industry, the medical devices industry, government regulatory agencies, and military
procurement agencies are adopting this approach in many places worldwide.
External Driving Forces. The driving forces that have resulted in widespread implementation
of the ISO 9000 standards can be summed up in one phrase: the globalization of business.
Expressions such as the “post-industrial economy” and “the global village” reflect profound changes
during recent decades. These changes include:
 New technology in virtually all industry/economic sectors
 Worldwide electronic communication networks
 Widespread worldwide travel
 Dramatic increase in world population
 Depletion of natural resource reserves
Arable land, fishing grounds, fossil fuels
 More intensive use of land, water, energy, air
Widespread environmental problems/concerns
 Downsizing of large companies and other organizations
Flattened organizational structure
Outsourcing of functions outside the core functions of the organization
 Number and complexity of language, culture, legal, and social frameworks encountered in the
global economy
Diversity a permanent key factor
 Developing countries becoming a larger proportion of the total global economy
New kinds of competitors and new markets
These changes have led to increased economic competition, increased customer expectations for
quality, and increased demands upon organizations to meet more stringent requirements for quality
of their products.
TABLE 11.1 The ISO 9000 Family of International Standards
ISO 8402 Quality Vocabulary (1994)
ISO 9000 Quality Management and Quality Assurance standards
Part 1: Guidelines for Selection and Use (1994)
Part 2: Generic Guidelines for the Application of ISO 9001, ISO 9002, and ISO 9003 (1993)
Part 3: Guidelines for the Application of ISO 9001 to the Development, Supply, and
Maintenance of Software (1991, reissue 1993)
Part 4: Application for Dependability Management (1993)
ISO 9001 Quality Systems—Model for Quality Assurance in Design, Development, Production,
Installation and Servicing (1994)
ISO 9002 Quality Systems—Model for Quality Assurance in Production, Installation, and Servicing
ISO 9003 Quality Systems—Model for Quality Assurance in Final Inspection and Test (1994)
ISO 9004 Quality Management and Quality System Elements
Part 1: Guidelines (1994)
Part 2: Guidelines for Services (1991, reissue 1993)
Part 3: Guidelines for Processed Materials (1993)
Part 4: Guidelines for Quality Improvement (1993)
ISO 10005 Quality Management—Guidelines for Quality Plans (1995)
ISO 10007 Guidelines for Configuration Management (1994)
ISO 10011 Guidelines for Auditing Quality Systems
Part 1: Auditing (1990, reissue 1993)
Part 2: Qualification Criteria for Quality Systems Auditors (1991, reissue 1993)
Part 3: Management of Audit Programs (1991, reissue 1993)
ISO 10012 Quality Assurance Requirements for Measuring Equipment
Part 1: Management of Measuring Equipment (1992)
ISO 10013 Guidelines for Developing Quality Manuals (1994)
Source: Marquardt, D. W., et al. (1991). “Vision 2000: The Strategy for the ISO 9000 Series Standards in the ‘90s,” Quality
Progress, May, pp. 25–31.
The globalization of business is a reality even for many small- and medium-size companies.
These smaller companies, as well as large companies, now find that some of their prime competitors
are likely to be based in another country. Fewer and fewer businesses are able to survive by considering
only the competition within the local community. This affects the strategic approach and the
product planning of companies of all sizes.
Internal Response to the External Forces. Companies everywhere are dealing with the
need to change. There is greater focus on human resources and organizational culture, on empowering
and enabling people in their jobs. Dr. W. Edwards Deming often said that many workers do not
know what their job is. ISO 9000 implementation involves establishing policy, setting objectives for
quality, designing management systems, documenting procedures, and training for job skills. All of
these are parts of clarifying what people’s jobs are.
Companies are adopting the process perspective. This concept is emphasized in the 1994 revision
of the ISO 9000 standards. In implementing the ISO 9000 standards, companies are using flowcharts
and other devices to emphasize work-process diagnosis and to find opportunities for process simplification
and improvement. Metrics are being used increasingly to characterize product quality and
customer satisfaction more effectively.
Companies are implementing better product design and work-process design procedures, and
improved production strategies. Benchmarking and competitive assessment are used increasingly.
Enterprise models, electronic data exchange, and other information technology approaches are growing
in scope and impact.
It may be asked: In this world of rapid change, how can a single family of standards, ISO 9000,
apply to all industry/economic sectors, all products, and all sizes of organizations?
The “Separate and Complementary” Concept. The ISO 9000 standards are founded
on the concept that the assurance of consistent product quality is best achieved by simultaneous
application of two kinds of standards:
 Product Standards (technical specifications)
 Quality system (management system) Standards
I call this the “separate and complementary” concept because the two types of standards are separate
from each other and they are complementary. The two types of standards are needed to provide
confidence that products will meet consistently the requirements for quality.
Product standards provide the technical specifications that apply to the characteristics of the product
and, often, the characteristics of the process by which the product is produced. Product standards
are specific to the particular product: both its intended functionality and the end-use situations the
product may encounter.
The management system is the domain of the ISO 9000 standards. It is by means of the distinction
between product specifications and management system features that the ISO 9000 standards
apply to all industry/economic sectors, all products, and all sizes of organizations.
The standards in the ISO 9000 family, both guidance and requirements, are written in terms of
what features are to be present in the management system of an organization but do not prescribe
how the features are to be implemented. The technology selected by an organization determines how
the relevant features will be incorporated in its own management system. Likewise, an organization
is free to determine its own management structure.
Comment. In regard to terminology, three terms are in current use, all of them having the same
essential meaning. Quality system is the formal term currently defined internationally in ISO 8402,
the ISO/TC176 vocabulary standard. Management system is the term frequently used in the daily language
of business. Quality management system is the term coming into increasing use for discussing
an organization’s management system when the focus is upon the overall performance of the organization
and its results in relation to the organization’s objectives for quality. A benefit of the term
“quality management system” is its effectiveness in emphasizing both:
 The commonalities in management system features
 The differences in the objectives for the results of an organization’s management system, for
various areas of application (e.g., quality management systems and environmental management
Characteristics of ISO 9000 Standards. Some of the ISO 9000 family standards contain
requirements, while others contain guidelines.
ISO 9001, ISO 9002, and ISO 9003 are requirements standards. They are quality system models
to be used for quality assurance purposes for providing confidence in product quality. A requirements
standard becomes binding upon a company or organization wherever:
 It is explicitly called up in a contract between the organization and its customer
 The organization seeks and earns third-party certification/registration
The text of a requirements standard is phrased in terms of the verb “shall,” with the meaning that the
stated requirements are mandatory.
ISO 9004 is an example of a guideline standard. Guideline standards are advisory documents.
They are phrased in terms of the word “should,” with the meaning that they are recommendations.
The scope of ISO 9004 is broader than the scope of ISO 9001, because it covers not only quality
system features necessary to provide customer confidence in product quality, but also quality system
features for organizational effectiveness.
All of the ISO 9000 family standards are generic, in the sense that they apply to any product or
any organization. All of the ISO 9000 family standards are nonprescriptive in the sense that they
describe what management system functions shall or should be in place; but they do not prescribe
how to carry out those functions.
The Clauses of ISO 9001 and Their Typical Structure. The ISO 9000 family is best
known for ISO 9001, the most comprehensive of the quality assurance requirements standards. As
indicated by their titles listed in Table 11.1, ISO 9002 is identical to ISO 9001 except that ISO 9002
does not contain requirements for the design function (clause 4.4). Between them, ISO 9001 and ISO
9002 account for the largest current market share of use of the ISO 9000 family documents. The third
quality assurance requirements standard, ISO 9003, is much less comprehensive and is based on final
product inspection only. Its current market share is very small, less than 2 percent in most parts of
the world.
The clause titles of ISO 9001 are shown in Table 11.2. The actual quality management system
requirements are spelled out in clause 4, specifically in subclauses 4.1 through 4.20. The scope of
ISO 9001 is focused on management system features that directly affect product quality. This
emphasis is consistent with the most fundamental purpose of ISO standards: to facilitate international
To illustrate the structure, content, and style of ISO 9001, two brief subclauses are quoted below:
Quality systems, general (clause 4.2.1)
The supplier shall establish, document, and maintain a quality system as a means of ensuring that
product conforms to specified requirements. The supplier shall prepare a quality manual covering the
requirements of this International Standard. The quality manual shall include or make reference to
the quality system procedures and outline the structure of the documentation used in the quality system.
Document and data control, general (clause 4.5.1)
The supplier shall establish and maintain documented procedures to control all documents and data
that relate to the requirements of this International Standard, including, to the extent applicable, documents
of external origin such as standards and customer drawings.
Some key words and their meanings are
Supplier: The organization to which the standard is addressed; namely the organization that
will supply the products to the customer organization
Establish: To institute permanently
Document: To record in readable form
Maintain: To keep up-to-date at all times
Documents: Examples are overall quality manual, quality system procedures, work instructions
for specific jobs, etc. (as distinct from records of actions completed, measurements made, etc.)
As a result of these clauses not being prescriptive as to how the requirements are to be implemented,
it is expected that there may be wide variations from one supplier to another. The appropriate
method of implementation will depend upon such characteristics as the type of product, its
complexity, regulatory requirements that must be satisfied for legal reasons, and size of supplier
One important benefit of the nonprescriptive character of the ISO 9000 standards—in particular,
ISO 9001—is across-the-board applicability to all organizational structures. The requirements of ISO
9001 are equally relevant whether the supplier organization is large or small; has one site or many; is
downsized, bereft of middle management; is heavily networked and/or based on joint ventures; uses
contract subsuppliers, part-time and/or temporary personnel; or has multinational legal and/or economic
arrangements. The only organizational requirement is that the organization shall have “management
with executive responsibility” and that the executive management “appoint a member of the
TABLE 11.2 International Standard ISO 9001:1994(E)
Quality systems—Model for quality assurance in design,
development, production, installation, and servicing
Clause titles
1 Scope
2 Normative reference
3 Definitions
4 Quality system requirements
4.1 Management responsibility
4.2 Quality system
4.3 Contract review
4.4 Design control
4.5 Document and data control
4.6 Purchasing
4.7 Control of customer-supplied product
4.8 Product identification and traceability
4.9 Process control
4.10 Inspection and testing
4.11 Control of inspection, measuring, and test equipment
4.12 Inspection and test status
4.13 Control of nonconforming product
4.14 Corrective and preventive action
4.15 Handling, storage, packaging, preservation, and delivery
4.16 Control of quality records
4.17 Internal quality audits
4.18 Training
4.19 Servicing
4.20 Statistical techniques
supplier’s own management” to be the management representative with responsibility for the establishment,
implementation, maintenance, and reporting on the performance of the quality system.
The guideline standard ISO 9000-1:1994 explains many concepts that are fundamental to the ISO
9000 family. Among these is the concept of the four facets of product quality:
1. Quality due to definition of needs for the product
 Defining and updating the product to meet marketplace requirements and opportunities
2. Quality due to product design
 Designing into the product the characteristics that enable it to meet marketplace requirements
and opportunities:
Features that influence intended functionality
Features that influence the robustness of product performance under variable conditions of production
and use
3. Quality due to conformance to product design
4. Quality due to product support throughout the product life cycle
Facets 1, 2, 3, and 4 encompass all stages of the product life cycle.
The publication of the ISO 9000 series in 1987 brought necessary harmonization on an international
scale. As expected, the initial emphasis of ISO 9000 standards application was primarily on
facet 3, eliminating nonconformities in product supplied to customers. But, to many people’s surprise
there was little use of ISO 9003. By the late 1980s suppliers were recognizing that the preventive
approach of ISO 9001 and ISO 9002 was more effective than final-inspection alone as the means
to achieve quality due to conformance to product design. In the years before the ISO 9000 standards
were developed and put into commercial and industrial uses, the national predecessor counterparts
of ISO 9003 had the predominant market share; they focused on a final-inspection-only approach to
facet 3.
ISO 9002 implementation now has the largest market share of the three requirements standards.
With the growing worldwide emphasis on quality, suppliers involved in international trade have continued
to gain maturity of understanding about quality systems. Consequently, the market share of
ISO 9001 has increased, reflecting the widening appreciation of facet 2 by customers and suppliers.
The 1994 revisions of the ISO 9000 standards include significant changes in many features of the
requirements and guidelines for a quality management system. These changes tend to strengthen the
quality contributions from all of facets 1, 2, 3, and 4. However, primary emphasis still remains on
facet 3 and the first group of features under facet 2. The next revision is likely again to have some
broadening of the emphasis, reflecting the continually changing needs of international trade.
One of the most pressing needs in the early years of ISO/TC176 work was to harmonize internationally
the meanings of terms such as “quality control” and “quality assurance.” These two terms,
in particular, were used with diametrically different meanings among various nations, and even within
nations. In my role as convener of the working group that wrote the ISO 9000:1987 standard, I
proposed early in the 1980s that the term “quality management” be introduced into the ISO 9000
standards as the umbrella term for quality control and quality assurance. The term “quality management”
was defined, included in ISO 8402, adopted internationally, and is now used worldwide. This,
in turn, enabled agreement on harmonized definitions of the meanings of each of the terms “quality
control” and “quality assurance.”
However, discussions in TC176 during 1995 revealed that the essential commonalities and distinctions
between quality management and quality assurance are still not universally understood.
This may be a result of the expansion of ISO 9000 standards use to many more countries than participated
in the early 1980s, or lack of widespread reference to ISO 8402, or deficiencies in the ISO
8402 definitions. Undoubtedly, all these reasons have contributed. In any event, the meanings of the
terms “quality management” and “quality assurance” need careful articulation to achieve clarity.
Table 11.3 describes the essence and is the same as the meanings intended in ISO 8402:1986 and
ISO 8402:1994. The quality control aspects of the umbrella term “quality management” are focused
on the word “achieving,” but all bullet points in the left-hand column of Table 11.3 relate at least
indirectly to quality control. The right-hand column of Table 11.3 shows that the quality assurance
aspects of the umbrella term “quality management” have primary focus on the notions of demonstrating
and providing confidence through objective evidence.
“Vision 2000” refers to the report of the ISO/TC176 Ad Hoc Task Force (Marquardt et al. 1991). It
outlines the strategy adopted by TC176 for the ISO 9000 standards in the 1990s. Several key concepts
and strategies from that report are essential to any discussion of the ISO 9000 standards.
Generic Product Categories. The task force identified four generic product categories:
 Processed materials
Table 11.4, from Marquardt et al. (1991), provides descriptors of the four generic product categories.
Several of these now have formal definitions in ISO 8402:1994. These categories encompass
TABLE 11.3 The Prime Focus of Quality Management and Quality Assurance
The Prime Focus of
Quality management Quality assurance
 Achieving results that satisfy the requirements for  Demonstrating that the requirements for
quality quality have been (and can be) achieved
 Motivated by stakeholders internal to the  Motivated by stakeholders, especially
organization, especially the organization’s customers, external to the organization
 Goal is to satisfy all stakeholders  Goal is to satisfy all customers
 Effective, efficient, and continually improving,  Confidence in the organization’s products is the
overall quality-related performance is the intended intended result
 Scope covers all activities that affect the total  Scope of demonstration covers activities that
quality-related business results of the organization directly affect quality-related process and
product results
all the kinds of product that need explicit attention in quality management and quality assurance standardization.
The initial 1987 standards were acknowledged to have inherited some of the hardware
bias of the predecessor standards. To remedy this, supplemental standards for each of the other three
generic product categories were developed and published (ISO 9000-3; ISO 9004-2; ISO 9004-3);
see Table 11.1.
One of the principal strategies in Vision 2000 was stated as follows:
We envision that, by the year 2000, there will be an intermingling, a growing together, of the terminology,
concepts, and technology used in all four generic product categories. This vision implies that, by
the year 2000, the need for separate documents for the four generic product categories will have diminished.
Terminology and procedures for all generic product categories will be widely understood and used
by practitioners, whatever industry/economic sector they might be operating in.
Consequently, our Vision 2000 for TC176 is to develop a single quality management standard (an
updated ISO 9004 that includes new topics as appropriate) and an external quality assurance requirements
standard (an updated ISO 9001) tied together by a road map standard (an updated ISO 9000). There would
be a high degree of commonality in the concepts and architecture of ISO 9004 and ISO 9001. The requirements
in ISO 9001 would continue to be based upon a selection of the guidance elements in ISO 9004.
Supplementary standards that provide expanded guidance could be provided by TC176 as needed.
This strategy continues to guide TC176 in its work on the next revisions.
Acceptance, Compatibility, and Flexibility. Vision 2000 proposed four goals that relate
to maintaining the ISO 9000 standards so that they continually meet the needs of the marketplace.
These goals are universal acceptance, being adopted and used worldwide; current compatibility,
facilitating combined used without conflicting requirements; forward compatibility, with successive
revisions being accepted by users; and forward flexibility, using architecture that allows new features
to be incorporated readily.
TC176 continues to use these goals as guides, recognizing as in Vision 2000 that “Proposals that
are beneficial to one of the goals might be detrimental to another goal. As in all standardization, compromises
and paradoxes might be needed in specific situations.”
Avoiding Proliferation. Vision 2000 recognized that the role of the ISO 9000 standards to
facilitate international trade could be maintained only if the remarkable, rapid, worldwide success in
replacing national standards with harmonized ISO 9000 international standards did not itself lead to
new rounds of proliferation. The issue was stated as follows:
TABLE 11.4 Generic Product Categories*
Generic product category Kinds of product
Hardware Products consisting of manufactured pieces, parts, or assemblies thereof
Software Products such as computer software, consisting of written or otherwise
recordable information, concepts, transactions, or procedures
Processed materials† Products (final or intermediate) consisting of solids, liquids, gases, or
combinations thereof, including particulate materials, ingots, filaments, or
sheet structures
Services Intangible products which may be the entire or principal offering or
incorporated features of the offering, relating to activities such as planning,
selling, directing, delivering, improving, evaluating, training, operating, or
servicing a tangible product
*All generic product categories provide value to the customer only at the times and places the customer interfaces with and
perceives benefits from the product. However, the value from a service often is provided primarily by activities at a particular
time and place of interface with the customer.
†Processed materials typically are delivered (packaged) in containers such as drums, bags, tanks, cans, pipelines, or rolls
Source: Marquardt et al. (1991).
If the ISO 9000 series were to become only the nucleus of a proliferation of localized standards
derived from, but varying in content and architecture from, the ISO 9000 series, then there would be little
worldwide standardization. The growth of many localized certification schemes would present further
complications. Once again, there could be worldwide restraint of trade because of proliferation of standards
and inconsistent requirements.
Vision 2000 emphatically discourages the production of industry/economic-sector-specific generic
quality standards supplemental to, or derived from, the ISO 9000 series. We believe such proliferation
would constrain international trade and impede progress in quality achievements. A primary purpose of the
widespread publication of this article is to prevent the proliferation of supplemental or derivative standards.
It is, however, well understood that product-specific standards containing technical requirements for
specific products or processes or describing specific product test methods are necessary and have to be
developed within the industry/economic sector.
Proliferation has been virtually eliminated worldwide in terms of national standards because of
the withdrawal of prior national standards and adoption of the ISO 9000 standards. Moreover, the
fundamental role of the ISO 9000 standards in relation to other areas of international standardization
has been incorporated into the ISO/IEC Directives, which govern the operations of all Technical
Committees of ISO and IEC. (ISO and IEC together coordinate and publish international voluntary
consensus standards for all sectors of the economy and all technical fields. IEC, the International
Electrotechnical Commission, deals with standards in industries related to electrical and electronic
engineering; ISO deals with all other areas of standardization.) Clause 6.6.4 of the ISO/IEC
Directives, Part 2, reads:
6.6.4 When a technical committee or sub-committee wishes to incorporate quality systems requirements
in a standard for a product, process, or service, the standards shall include a reference to the relevant
quality systems standard (ISO 9001, ISO 9002 or ISO 9003). It shall not add to, delete, change or
interpret the requirements in the quality systems standard.
Any requests for additions, deletions, changes or interpretations shall be submitted to the secretariat
of ISO/TC176/SC2: Quality systems.
When the industry or sector terminology is sufficiently different, a document explaining the relationship
between the quality assurance terminology and the sector terminology may be prepared.
This clause may be viewed as an operational definition of avoiding proliferation. It is being
applied to good effect within ISO/IEC, and as a result a number of other ISO TCs have not prepared
proposed new standards that would have represented unnecessary proliferation within specific industry/
economic sectors. However, in one area of application, the ISO Technical Management Board has
ruled that a new Technical Committee, TC207, on Environmental Management Systems should be
set up. This is discussed further in this section (see Other Areas of Application), and should be
viewed as one of the necessary “compromises and paradoxes” quoted above from Vision 2000.
Even before publication of the 1994 “Phase 1” revisions, TC176 and its subcommittees began
explicit planning for the next revisions, which had been referred to in Vision 2000 as the Phase 2
revisions. In addition, TC176 appointed a Strategic Planning Advisory Group (SPAG). The SPAG
study included a formal strategic planning effort examining the TC176 products, markets, benefits
to users, beliefs about the value of such standardization, external trends, competitive factors, and
unmet market needs. From these examinations emerged a number of strategic opportunities and,
ultimately, strategic goals.
The essential concepts and strategies of Vision 2000 were reaffirmed. However, the study concluded
that ISO 9004 should and could have more impact in guiding practitioners of quality management.
To do so requires an expansion of scope and change of approach. During 1995 TC176 also
reexamined, in various meetings and study groups, the developing plans for the next revisions of the
ISO 9000 standards. At the Durban, South Africa, meeting in November 1995, work was completed
on specifications for the revision of ISO 9000, ISO 9001, ISO 9004, and a proposed new document
on quality management principles. The specifications were prepared for formal comment by the
member bodies representing the various nations. Also TC176 achieved tentative consensus on the
architecture and content of the ISO 9000 family for the year 2000. One of the guiding themes is to
avoid unnecessary proliferation of standards within the ISO 9000 family itself, as well as external to
TC176. The leaders of the delegations of the more than 40 countries represented at Durban spent several
days on these issues, including the detailed texts of vision and mission statements and of key
strategies for TC176 activities and products.
Origin of the Need. The earliest users of quality assurance requirements standards were large
customer organizations such as electric power providers and military organizations. These customers
often purchase complex products to specific functional design. In such situations the quality assurance
requirements are called up in a two-party contract, where the providing organization (i.e., the
supplier) is referred to as the “first party” and the customer organization is referred to as the “second
party.” Such quality assurance requirements typically include provisions for the providing organization
to have internal audits sponsored by its management to verify that its quality system meets
the contract requirements. These are first-party audits. Such contracts typically also include provisions
to have external audits sponsored by the management of the customer organization to verify
that the supplier organization’s quality system meets the contract requirements. These are secondparty
audits. Within a contractual arrangement between two such parties, it is possible to tailor the
requirements, as appropriate, and to maintain an ongoing dialogue between customer and supplier.
When such assurance arrangements become a widespread practice throughout the economy,
the two-party, individual-contract approach becomes burdensome. There develops a situation
where each organization in the supply chain is subject to periodic management system audits by
many customers and is itself subjecting many of its subsuppliers to such audits. There is a lot
of redundant effort throughout the supply chain because each organization is audited multiple
times for essentially the same requirements. The conduct of audits becomes a significant cost element
for both the auditor organizations and auditee organizations.
Certification/Registration-Level Activities. The development of quality system certification/
registration is a means to reduce the redundant, non-value-adding effort of these multiple
audits. A third-party organization, which is called a “certification body” in some countries, or a “registrar”
in other countries (including the United States), conducts a formal audit of a supplier organization
to assess conformance to the appropriate quality system standard, say, ISO 9001 or ISO 9002.
When the supplier organization is judged to be in complete conformance, the third party issues a certificate
to the supplying organization and registers the organization’s quality system in a publicly
available register. Thus, the terms “certification” and “registration” carry the same marketplace
meaning because they are two successive steps signifying successful completion of the same process.
To maintain its registered status, the supplier organization must pass periodic surveillance audits
by the registrar. Surveillance audits are often conducted semiannually. They may be less comprehensive
than the full audit. If so, a full audit is performed every few years.
In the world today, there are hundreds of certification bodies/registrars. Most of them are private,
for-profit companies. Their services are valued by the supplier organizations they register, and by
the customer organizations of the supplier organizations, because the registration service adds value
in the supply chain. It is critical that the registrars do their work competently and objectively and that
all registrars meet standard requirements for their business activities. They are, in fact, supplier organizations
that provide a needed service product in the economy.
Accreditation-Level Activities. To assure competence and objectivity of the registrars, systems
of registrar accreditation have been set up worldwide. Accreditation bodies audit the registrars for conformity
to standard international guides for the operation of certification bodies. The quality system of
the registrar comes under scrutiny by the accreditation body through audits that cover the registrar’s documented
quality management system, the qualifications and certification of auditors used by the registrar,
the record keeping, and other features of the office operations. In addition, the accreditation body
witnesses selected audits done by the registrar’s auditors at a client supplier organization’s facility.
Accreditation and Registration Flowchart, Including Related Activities. This
process as it operates in the United States is depicted graphically in Figure 11.1. The three columns
in the figure depict the three areas of activity of the Registrar Accreditation Board: accreditation of
registrar companies, certification of individuals to be auditors, and accreditation of the training
courses which are part of the requirements for an auditor to be certified. The relevant ISO/IEC international
standards and guides that apply in each of these activities are shown in Figure 11.1. The
ISO criteria documents for auditing quality systems are the ISO 10011 standard, Parts 1, 2, and 3.
See Table 11.1. Part 2 of ISO 10011 deals specifically with auditor qualifications.
In the United States the registrar accreditation is carried out by the Registrar Accreditation Board
(RAB) under a joint program with the American National Standards Institute (ANSI). This joint program
is called the American National Accreditation Program for Registrars of Quality Systems. The auditor
certification and training course accreditation portions of the entire scheme shown in Figure 11.1 also
are carried out by RAB.
Materials governing accreditation procedures for the American National Accreditation Program
for Registrars of Quality Systems are available from the Registrar Accreditation Board, P.O. Box
3005, Milwaukee, WI 53201-3005.
FIGURE 11.1 Accreditation and registration process.
Mutual International Acceptance. Various other countries have implemented these three
areas of activity, too:
 Accreditation of certification bodies/registrars
 Certification of auditors
 Accreditation of auditor training courses
The systems in the Netherlands and in the United Kingdom have been in place longer than most. At
this time, various bilateral mutual recognition agreements are in place between certain countries whereby,
for example, the certification of an auditor in one country carries over into automatic recognition of
that certification in another country. In other situations, a memorandum of understanding has been
negotiated between, say, the accreditation bodies in two countries, whereby they enter into a cooperative
mode of operation preliminary to entering into a formal mutual recognition agreement. Under a
memorandum of understanding, the accreditation bodies may conduct jointly the audit of a registrar,
and the auditors may jointly document the results of the audit. However, each of the accreditation bodies
would make its own decision whether to grant or continue, as the case may be, the accreditation.
In principle, there should be no need for a supplier organization to obtain more than one certification/
registration. A certificate from a registrar accredited anywhere in the world should, in principle,
be accepted by customer organizations anywhere else in the world. In practice, it takes time to
build infrastructure comparable to Figure 11.1 in any country. It takes additional time (measured in
years) for that infrastructure to mature in its operation and for confidence to build in other countries.
Of course, not all countries decide to set up their own infrastructure but may choose to have their
supplier organizations who wish to become registered do so by employing the services of an accredited
registrar from another country.
Indeed, many registrar companies have established operations internationally and provide services
in many countries. Such registrars often seek accreditation in multiple countries because their
customers (the supplier organizations) look for accreditation under a system with which they are
familiar and have developed confidence.
At the present time, there are a multiplicity of arrangements involving single or multiple accreditations
of registrars, single or multiple certifications of auditors, and single or multiple accreditations
of training courses. The overall system is moving toward widespread mutual recognition, but
the ultimate test of credibility is the marketplace willingness to accept a single certification and a single
The ISO/IEC are themselves sponsoring development and implementation of a truly international
system of mutual recognition for the accreditation of certification bodies/registrars. Called
QSAR, it has been worked out in detail through international negotiations and is starting to be
implemented. It is expected that the QSAR arrangement will accelerate the process of mutual
recognition internationally.
The current status where registrars and course providers may have multiple accreditations, and
auditors may have multiple certifications, may seem to have more redundancy than necessary. If we
step back and compare the current situation to the alternative of widespread second-party auditing of
supplier organizations’ quality systems, it must be acknowledged that the present situation is better
in the following ways:
 Much less redundancy of auditing
 Much improved consistency of auditing
 Potential for even less redundancy and further improved consistency through the use of international
standards and guides as criteria and through mutual harmonization efforts driven by the marketplace.
Formal International Mutual Recognition. For the United States, there is one further
complication. Almost alone among the countries of the world, the U.S. standards system is a privatesector
activity. American National Standards Institute (ANSI), a private sector organization, is the
coordinating body for standards in the United States. Under the ANSI umbrella many organizations
produce and maintain numbers of American National Standards. Most of these standards relate to
product technical specifications. Among the largest U.S. producers of standards are organizations
such as the American Society of Testing and Materials (ASTM), the American Society of
Mechanical Engineers (ASME), and the Institute of Electrical and Electronics Engineers (IEEE), but
there are many other organizations that produce American National Standards applicable to specific
products or fields of activity. The ANSI system provides a consistent standards development process
that is open, fair, and provides access to all parties that may be materially affected by a standard. The
success of the U.S. system is attested to by the predominance of the U.S. economy internationally
and the widespread adoption of U.S. standards for multinational or international use.
However, there are three levels of activities and infrastructure in relation to conformity assessment
in international trade. Two of these levels have already been discussed: the certification/registration
level and the accreditation level. The third level is the recognition level. At the recognition level, the
national government of country A affirms to the government of country B that A’s certification and
accreditation infrastructure conforms to international standards and guides. In most countries of the
world, where the standards system is run by a government or semigovernment agency and the accreditation
activities are carried out by that agency, the recognition level is virtually automatic. In the
United States, various government agencies may be called upon to provide the formal recognition.
For example, in dealing with the European Union (EU) on products that fall under one of the
EU Directives that regulate products that have health, safety, and environmental risks, the EU
insists upon dealing through designated government channels. The relevant U.S. government
agency varies from one EU Directive to another. In many areas, the recognition responsibility will
come under the recently authorized National Voluntary Conformity Assessment System
Evaluation (NVCASE) program to be run by the Department of Commerce, through the National
Institute of Standards and Technology. The NVCASE program had not come into operation at the
time of this writing.
The conformity assessment approach of the European Union typifies what is happening in many
parts of the world. For a regulated product to be sold in any EU country, it must bear the “CE” mark.
Under the Modular Approach of the EU, to qualify for use of the mark the supplier organization must
produce evidence of conformity in four areas:
 Technical documentation of product design
 Type testing
 Product surveillance (by samples, or by each product)
 Quality assurance surveillance.
Depending on the directive, the EU will offer suppliers various routes (modules) to satisfy the
requirements. These routes range from “Internal Control of Production,” which focuses on the product
surveillance aspects, to “Full Quality Assurance,” which typically focuses on certification/registration
to ISO 9001 and relies upon the ISO 9001 requirements for capability in product design. In most modules
the manufacturer must submit product units, and/or product design technical information, and/or
quality system information to a certification body that has been designated by the government as a
“notified body.” The notified body must, in some modules, also provide for product tests where
required. Several modules involve certification to ISO 9001, ISO 9002, or ISO 9003.
The implementation of this modular approach to conformity assessment for regulated products
by the European Union (then called the European Community) was the largest, single, early impetus
to the rapid spread of certification/registration to ISO 9001 or ISO 9002 worldwide. For example,
about half of the dollar volume of U.S. trade with Europe is in regulated products.
Nevertheless, the global trends in technology and in requirements for quality, and the cost savings
of third-party versus widespread second-party auditing, as discussed previously in this section,
are powerful additional incentives and staying power for sustained international use and growth of
third-party quality system certification/registration.
Moreover, for a supplier organization it is not effective to attempt to have two quality management
systems, one for regulated products and another for nonregulated products. Consequently, there
are multiple incentives for large numbers of supplier organizations, engaged directly or indirectly in
international trade, to operate a quality management system that conforms to ISO 9001 or ISO 9002,
as appropriate.
The rapid worldwide adoption and implementation of the ISO 9000 standards, and the rapid growth
of the infrastructure of certification/registration bodies, accreditation bodies, course providers, consultants,
auditors, trade magazines, books, and ISO 9000 journalists is virtually unprecedented in any
field of standardization. This can appropriately be regarded as a remarkable success story for the ISO
9000 standards. Accompanying this success story are several problems and opportunities that need
careful attention.
Problems and Opportunities. In all important endeavors, “problems” and “opportunities”
abound. It is important to understand that these are two sides of the same coin. Every problem is a
doorway to opportunities; and every opportunity carries problems with it. This universal truth certainly
applies in relation to the ISO 9000 family. Coming to grips with these problem-opportunities
deepens our understanding of the ISO 9000 standards themselves. These problem-opportunities will
be with us for a long time because they come from fundamental economic issues that are always
present in national and international trade. Some, or all, of them could be make-or-break issues
for the success of ISO 9001, and with it the ISO 9000 family. As a preliminary to discussing a
number of problem-opportunities, it is important to understand what is implied by the concept of
The Scope of Standardization. “Standardization” encompasses activities in two interrelated
 The conception, planning, production, promotion, and selling of standards
 The conception, planning, establishment, control, promotion, and maintenance of standards
In practice, the standardization problems and opportunities for the ISO 9000 family relate to implementation
activities at least as much as to the standards themselves. The topics discussed in the following
illustrate the predominance of problem-opportunities that arise from implementation activities.
The Concept of Continuous Improvement. The philosophy of using quality assurance
standards has changed over the years. In the 1960s the virtually universal perspective of business managers
was “If it ain’t broke, don’t fix it.” In that philosophical environment of maintaining the status
quo, the characteristics shown at the top of Table 11.5 prevailed in most of industry and commerce.
Today, as illustrated by the minimal use of ISO 9003, the philosophy of 30 years ago is giving way to
a philosophy of continuous improvement. Continuous improvement is increasingly necessary for economic
survival in the global economy and is becoming a widely pursued goal. It is the only reliable
route to sustaining marketplace advantage for both customer and supplier.
As shown at the bottom of Table 11.5, the focus of quality assurance standardization is now on
prevention of nonconformities. This requires a “process” focus, which is reflected in the documentation
and many other features.
The Role of Continuous Improvement in ISO 9001. In the first years of use since
1987 an unfortunate mind-set has been adopted by many registrars/certifiers and their auditors. This
mind-set can be called the “status quo mind-set.” Reminiscent of the 1960s, it is characterized by
the ditty:
“Say what you do.
Do what you say.”
This simple ditty is correct as far as it goes, but is far short of the requirements in the ISO 9000 standards
set in 1987 and revised in 1994. It focuses only on adherence to established procedures. For
example, the status quo mind-set ignores the linked requirements to demonstrate continuing adequacy
of the quality system:
 For business objectives
 For customer satisfaction
Intrinsic to the 1994 revision of ISO 9001 is a reshaped mind-set as summarized in Table 11.6.
The reshaped mind-set is a cycle of continuous improvement, which can be depicted in terms of the
classic plan-do-check-act management cycle. The P-D-C-A cycle is expected to be used explicitly in
the next revisions of the ISO 9000 standards.
Comment. Various people, in both the United States and Japan (e.g., Shewhart, Deming,
Mizuno), have been associated with the early evolution of the P-D-C-A management cycle. The
history is traced by Kolesar (1994, pp. 14–16). In its form and application to depict the management
cycle, “P-D-C-A plays a central role in Japanese thought” (Kolesar 1994, p. 16). In recent
years, P-D-C-A has been widely accepted in the Western world as a simple but robust embodiment
of management as an activity. The P-D-C-A management cycle is compatible, in particular,
with the contemporary concept that all work is accomplished by a process (ISO 9000-1:1994,
clause 4.6).
Continuous improvement is a necessary consequence of implementing ISO 9001 (and ISO 9002).
There are two groupings of linked clauses in ISO 9001 that work together to ensure continuous
TABLE 11.5 Philosophy of Using Quality Assurance Standards
30 years ago
 Quality goal: Maintain status quo.
 Business goal: The best deal for this
 Methods: Final-inspection oriented;
sort good product from
bad. “Records” paperwork
was featured.
 Customer/supplier relationship: Adversarial.
 Quality goal: Continuous improvement.
 Business goal: Mutual marketplace
 Methods: Prevention oriented; don’t
make bad product.
“Process” documentation is
 Customer/supplier relationship: Partnership.
The linkages among the clauses are really quite clear in ISO 9001 if your mind-set does not block
them out. The intention of the ISO 9000 standards has always been that the clauses are elements of
an integrated quality system. In implementing any system, the interrelationships among the elements,
that is, the linkages, are as important as the elements themselves.
The linkages among clauses in ISO 9001 can be recognized in three ways:
 Explicit cross references between linked clauses shown by parenthetic expressions “(see z.zz)” in
the text of certain clauses
 Use of key words or phrases in clauses that have linked requirements, for example, “objectives”
for quality in clauses 4.1.1 (Quality Policy) and 4.1.3 (Management Review)
 Content interrelationships which cause linkages among the activities required by two or more
clauses, for example, clauses 4.1.1 (Quality Policy), 4.1.3 (Management Review), 4.14 (Corrective
Action), and 4.17 (Internal Quality Audits).
The two groupings of linked clauses that work together to ensure continuous improvement are
1. Continuous improvement via objectives for quality and ensuring the effectiveness of the quality
system: The 1994 revision of ISO 9001 expands and strengthens the requirements for executive
management functions and links a number of clauses by requirements to define “objectives” for
quality [clauses 4.1.1 (Quality Policy) and 4.1.3 (Management Review)] and to “ensure the effectiveness
of the quality system” [clauses 4.1.3 (Management Review), 4.2.2 (Quality System),
4.16 (Quality Records), 4.17 (Internal Quality Audits)]. In today’s competitive economy, the
objectives for quality must continually be more stringent in order to maintain a healthy business
position. More stringent objectives for quality translate inevitably into the need for an increasingly
effective quality system.
2. Continuous improvement via internal audits and management review: The 1994 revision of
ISO 9001 expands and strengthens the requirements of four clauses in the 1987 standard that
link together for continuous improvement in the supplier organization. These are internal quality
audits (clause 4.17), corrective and preventive action (clause 4.14), management representative
(clause, and management review (clause 4.1.3). The interplay of activities required
by these four clauses provides a mechanism to institutionalize the pursuit of continuous
It is instructive to display simultaneously the similarities of the three managing processes that
have had great influence on the practice of quality management. These are the P-D-C-A cycle, the
ISO 9001/ISO 9002 requirements, and the Juran Trilogy. These are summarized in Table 11.7.
Viewed from this perspective, the similarities are striking.
If the ISO 9000 implementation community fails to embrace adequately the continuous improvement
requirements in the standards, the competitiveness of the standards themselves will be seriously
compromised in the global marketplace.
TABLE 11.6 The Reshaped Mind-Set
A cycle of continuous improvement (built into ISO 9001 by the linked requirements)
 Plan your objectives for quality and the processes to achieve them.
 Do the appropriate resource allocation, implementation, training, and documentation.
 Check to see if
You are implementing as planned
Your quality system is effective
You are meeting your objectives for quality
Your objectives for quality are relevant to the expectations and needs of customers
 Act to improve the system as needed.
The Role of Statistical Techniques. From the earliest days of the quality movement, statistical
techniques have been recognized as having an important role. In fact, during the 1940s and
1950s, statistical techniques were viewed as the predominant aspect of quality control. During succeeding
decades, the management system increasingly took center stage. In the 1987 version of ISO
9001, clause 4.20 on statistical techniques paid only lip service to its subject. The implementation of
quality assurance standards worldwide has reinforced the deterioration of emphasis on statistical
techniques to the point that important technical advances in statistical techniques for quality have
been deprecated by many practitioners who have claimed that only the simplest, most primitive, and
most ancient statistical techniques are needed, provided they are conscientiously implemented. Then
they neglect to implement even those!
It is critical that this situation be remedied and that statistical techniques and management systems
each be given an important place in quality.
Fortunately, the 1994 version of ISO 9001 contains explicit, meaningful requirements relating to
statistical techniques (clause 4.20, Statistical Techniques), citing their relation to “process capability”
and “product characteristics.” Recalling the ISO 9001 mechanisms to link clauses, there is direct
linkage between clause 4.20 and process control (clause 4.9, especially 4.9g), as well as indirect linkages
to other clauses. These requirements, if conscientiously implemented, would guarantee that statistical
techniques are enabled to have their important place in quality under the ISO 9001 umbrella.
Unfortunately, a large majority of personnel in the existing infrastructure of auditors, registrars,
and accreditation bodies—and their supporting consultants, training course providers, etc.—have little
knowledge or experience in statistical techniques. Consequently, despite the requirements in
clause 4.20 of ISO 9001:1994, the clause still is receiving little emphasis.
Like all problems, this creates opportunities for entrepreneurial supplier companies, registrars,
auditors, consultants, and course providers. I hope many will seize these opportunities (which really
are obligations) in the near future. Moreover, this problem places responsibilities on accreditation
bodies. They have the responsibility to ensure that the third-party registration system is implemented
in conformance to the applicable international standards and guides. In particular, this includes ISO
9001 itself, and clause 4.20.
TABLE 11.7 Similarities of Three Managing Processes
P-D-C-A cycle ISO 9001/9002 requirements Juran Trilogy
Plan Say what you will do Planning
 “Plan, define, establish, document”  Determine needs
 “Objectives for quality”  Establish product
 “Provide resources”  Establish process
 Develop process
 Set goals
Do Do what you say Control (remove sporadic deficiencies)
 “Implement, maintain”  Run process
Record what you did  Evaluate performance
 “Quality records”  Compare to goals
 Act on differences
Check Check results versus expectations Improvement (remove chronic deficiencies)
 Management review  Nominate projects
 Internal audits  Establish teams
 External audits  Use improvement process
 Provide resources
Act Act on any deficiencies
 Quality system revision
 Preventive action
 Corrective action
Source: Kolesar (1994).
In the United States, the Registrar Accreditation Board (RAB) has taken some initiatives in this
direction. The RAB has, among these initiatives, issued to all ANSI/RAB-accredited registrars a bulletin
stating RAB’s intention to monitor the operations of registrars with respect to the linked
requirements and the implementation of clause 4.20. I have emphasized this issue in various national
and international presentations and publications (e.g., Marquardt 1995, 1996a), encouraging the
accreditation bodies and certification/registration bodies in other countries to take similar steps.
There are opportunities for ISO/TC176, too. TC176 is cooperating with ISO/TC69 (Application of
Statistical Methods) to provide guidance documentation on the use of statistical techniques when
implementing the standards in the ISO 9000 family.
Interpretations of the Standards. In actual application ISO standards are published by
ISO in English and French (and often in Russian), which are the official ISO languages. ISO
requires that “the texts in the different official language versions shall be technically equivalent and
structurally identical” (ISO/IEC Directives, Part 1, 1.5, 1989). Sometimes ISO itself publishes standards
in languages other than the official languages; then “each is regarded as an original-language
version” (ISO/IEC Directives, Part 1, F.3, 1995). “However, only the terms and definitions given in
the official languages can be considered as ISO terms and definitions” (ISO/IEC Directives, Part 3,
B2.2, 1989).
When a nation “adopts” an ISO standard, the standard is first translated by the national body into
the national language, and processed through the official national procedures for adoption. In the
United States, the translation issue is minimal, consisting, when deemed necessary, of replacement
of British English (the ISO official English) with American English spellings or other stylistic editorial
details. In the United States, the adoption process is under the American National Standards
Institute procedures, which ensure the objectivity, fairness, and lack of bias that might favor any constituency;
these are requirements for all American National Standards.
In situations where the national language is not one of the ISO official languages, ISO has, at present,
no formal procedure for validating the accuracy of the national body translation. Translation
from one language to another always presents challenges when great accuracy of meaning should be
preserved. There are many ways in which the meaning may be changed by a translation. These
changes can be a troublesome source of nontariff trade barriers in international trade.
The problem-opportunity relating to interpretations of the ISO 9000 standards goes beyond
problems of translation into languages other than the official ISO languages. In the global economy
many situations of use are encountered; the intended meaning of the standard is not always
clear to those applying the standard in some situations of use. For such situations each member
body of ISO is expected to set up interpretation procedures. There will, nevertheless, be cases
where an official ISO interpretation is required. ISO has, at present, no formal procedure for
developing and promulgating such official interpretations. ISO/TC176 has taken the initiative
with ISO Central Secretariat to establish an official procedure; ultimately ISO/TC176 should be
the point of the final interpretation of the ISO 9000 standards which it is responsible to prepare
and maintain.
When the situation of use is a two-party contractual situation between the supplier organization
and the customer organization, differences of interpretation should normally be revealed and
mutually resolved at an early stage (e.g., during contract negotiation and contract review). Official
international interpretations become more necessary in third-party certification/registration situations.
In third-party situations negotiations between supplier and customer tend to focus on the
technical specifications for the product, plus only those quality system requirements, if any, that
go beyond the scope of the relevant ISO 9000 requirements standard.
Defining the Scope of Certification/Registration
Background. There is great variability in the documented definitions of scope of registration of
suppliers’ quality systems. This variability is observed from supplier to supplier for a given registrar,
from registrar to registrar in a given nation, and from one nation to another. Greater consistency in
defining and documenting this scope is an essential prerequisite for:
 Establishing marketplace credibility of quality system certification/registration to ISO 9001, ISO
9002, or ISO 9003
 Negotiating meaningful mutual recognition arrangements among nations
Beyond the benefits of marketplace credibility, there are important benefits to registrars if the
ground rules for defining scope are consistent for all parties. This topic is an important problemopportunity
for the ISO 9000 standards.
To describe adequately the scope of certification/registration of a supplier’s quality system, four
questions must be asked:
 Which standard?
 Which geographic sites or operating units?
 Which products?
 Which portions of the supply chain?
The first three elements of scope are dealt with in ISO/IEC Guides in generic terms. The last
(supply-chain boundaries) is not dealt with in the Guides but is equally important.
Certificates of registration are the original records from which other records (e.g., lists of registered
quality systems) are derived. Examination of samples of certificates and registers shows that
even the first three elements are not universally or uniformly documented today.
Selection of Standard. The procedure should provide confidence to the customer that the selection
of the appropriate standard jointly by the supplier and the registrar has taken adequately into consideration
the amount and nature of product design activity that is involved in the products produced
by the supplier, as well as the nature of the production processes through which the supplier adds
value to the product. In some cases ISO 9003 or ISO 9002 has been selected when it appears that a
more comprehensive quality assurance model would be more appropriate. In some cases, the mismatch
may not be readily apparent to the supplier’s customer.
An example where clarity is important is distributor operations. Many are registered to ISO 9003
under the rationale that the distributor does not produce the (tangible) products themselves. However,
a distributor’s product is the service products of acquiring, stocking, preserving, order fulfilling, and
delivery. Hence ISO 9002 is appropriate to cover the production of these services. Distributors who
design their service products should be registered to ISO 9001.
Specification of Boundaries in Terms of Geographic Location or Operating Unit. The procedure
should inform the customer whether the product the customer receives is processed within the registered
quality system, even in situations where the supplier may have multiple sites or operating
units dealing with the same product, not all of which may be registered. The lack of consistent procedures
for scope description in regard to geographic locations or operating units included sets the
stage for misrepresentation.
Specification of Boundaries in Terms of Product Processed. The procedure should inform the
customer whether the product the customer receives is processed within the registered quality system,
even in situations where the supplier may deal with multiple products at the same site or operating
unit and not all of the products may be processed within the registered quality system. The lack
of consistent procedures in regard to product processed sets the stage for misrepresentation.
Specification of Boundaries in Terms of Supply-Chain Criteria. The procedure should inform the
customer regarding:
 The starting points of the supplier’s registered operations (e.g., the raw materials, parts, components,
services, and intermediate products that are provided by subsuppliers)
 The ending points of the supplier’s registered operations (i.e., the remaining steps on the way to
the ultimate consumer that are excluded from the supplier’s registered operations)
 The nature of the value that has been added by the supplier’s registered operations
Where the registered quality system represents only a fraction of the supplier’s operations, or a
fraction of the total value added in the product, this should be stated in registration documentation
so that, as a consequence, customers may be aware of this fact.
The procedures should not invite suppliers who wish to be registered, but want to exclude portions
of their operations from scrutiny by the registrar, to declare the excluded portions to be subcontractor
operations. It does not matter whether the excluded portions are in another nation,
elsewhere in the same nation, or simply another part of the same production site.
Procedures for this element of scope would apply also to support functions that are critical to
product quality, such as a test laboratory, which may not be included in the supplier’s quality system
as registered to ISO 9001 or ISO 9002.
Guiding Principle. There are many registrars; each is registering many supplier quality systems.
Each supplier is dealing with many customers. It is impractical to monitor adequately the operations
of such a system solely by periodic audits conducted by an accreditation body. Consequently the
guiding principle should be
Primary reliance must be placed on the concept of “truth in labeling,” by means of which every
customer has routine, ready access to the information upon which to judge all four elements of
scope of a supplier’s registered quality system.
Alternate Routes to Certification/Registration. Organizations differ in regard to the
status of their quality management efforts. Some are at an advanced state of maturity and effectiveness.
Others have hardly begun. Most are at some intermediate state. The ISO 9000 standards have
as their primary purpose the facilitation of international trade. They are, therefore, positioned to
ensure a level of maturity and effectiveness that meets the needs for reducing nontariff trade barriers
in international trade. This required level of maturity and effectiveness will change with the
passage of time (compare the minimal use of ISO 9003 and the growth of use of ISO 9001).
At any point in time, there will be some organizations that have well-established, advanced quality
management systems based on an approach that may go beyond the requirements of ISO 9001.
For such organizations, the cost of registration/certification by the usual third-party route is perceived
to be high compared to the incremental value added to their quality management system. This
is, and will continue to be, a significant problem-opportunity.
In the United States a number of such companies that have major international presence, especially
ones in the electronics and computer industry, have been working with organizations involved
in the implementation of third-party certification/registration to devise an approach that would gain
international acceptance. The approach would have to take cognizance of their existing quality management
maturity and reduce the cost of certification/registration, while supporting their international
trade by providing the assurance conferred by certification/registration.
Industry-Specific Adoptions and Extensions of ISO 9000 Standards
Industry-Specific Situations. In some sectors of the global economy there are industry-specific adoptions
and extensions of the ISO 9000 standards. These situations are a classic example of a problemopportunity.
As problems, such adaptations and extensions strain the goal of nonproliferation. As
opportunities, they have been found effective in a very few industries where there are special circumstances
and where appropriate ground rules can be developed and implemented consistently. These special
circumstances have been characterized by
1. Industries where the product impact on the health, safety, or environmental aspects is potentially
severe; as a consequence most nations have regulatory requirements regarding the quality management
system of a supplier
2. Industries that have had well-established, internationally deployed industry-specific or supplierspecific
quality system requirements documents prior to publication of the ISO 9000 standards
Fortunately, in the very few instances so far, the operational nonproliferation criteria of the
ISO/IEC Directives have been followed.
Medical Device Industry. Circumstance 1 relates to the medical device manufacturing industry.
For example, in the United States, the Food and Drug Administration (FDA) developed and promulgated
the Good Manufacturing Practice (GMP) regulations. The GMP operates under the legal
imprimatur of the FDA regulations, which predate the ISO 9000 standards. The FDA regularly
inspects medical device manufacturers for compliance with the GMP requirements. Many of these
requirements are quality management system requirements that parallel the subsequently published
ISO 9002:1987 requirements. Other GMP regulatory requirements relate more specifically to
health, safety, or environmental aspects. Many other nations have similar regulatory requirements
for such products.
In the United States, the FDA is in late stages of developing and promulgating revised GMPs that
parallel closely the ISO 9001:1994 standard, plus specific regulatory requirements related to health,
safety, or environment. The expansion of scope to include quality system requirements related to
product design reflects the recognition of the importance of product design and the greater maturity
of quality management practices in the medical device industry worldwide. Similar trends are taking
place in other nations, many of which are adopting ISO 9001 verbatim for their equivalent of the
GMP regulations.
In ISO, a new technical committee, ISO/TC210, has been formed specifically for medical device
systems. TC210 has developed standards that provide supplements to ISO 9001 clauses. These supplements
primarily reflect the health, safety, and environment aspects of medical devices and tend to
parallel the regulatory requirements in various nations. These standards are in late stages of development
and international approval at this time.
Automotive Industry. Circumstance 2 relates to the automotive industry. In the years preceding
publication of the 1987 ISO 9000 standards, various original equipment manufacturers (OEMs) in
the automotive industry had developed company-specific proprietary quality system requirements
documents. These requirements were part of OEM contract arrangements for purchasing parts, materials,
and subassemblies from the thousands of companies in their supply chain. The OEMs had large
staffs of second-party auditors to verify that these OEM-specific requirements were being met.
Upon publication of ISO 9001:1994, the major U.S. OEMs began implementation of an industrywide
common standard, labeled QS-9000, that incorporates ISO 9001 verbatim plus industry-specific
supplementary requirements. Some of the supplementary requirements are really prescriptive
approaches to some of the generic ISO 9001 requirements; others are additional quality system requirements
which have been agreed on by the major OEMs; a few are OEM-specific.
QS-9000 is being deployed by these OEMs in their worldwide operations. Part of this deployment
involves separate registrations to QS-9000, through existing registrars who have been accredited
specifically to the QS-9000 system. These QS-9000 registrars must use auditors who have had
specific accredited training in those QS-9000 requirements which are more prescriptive than, or go
beyond, ISO 9001 requirements. Accreditations are provided through specially designated accreditation
bodies, including RAB in the United States.
The QS-9000 system, by removing the redundancy of multiple second-party audits to multiple
requirements documents, is providing cost reductions for both the OEMs and the large number of
organizations in their supply chain. Assuming that credibility is maintained by continuous improvement
to meet marketplace needs and requirements, the goals of improved quality industry-wide and
worldwide, together with reduced costs, can be attained.
Computer Software. The global economy has become permeated with electronic information technology
(IT). The IT industry now plays a major role in shaping and driving the global economy. As in
past major technological advances, the world seems fundamentally very different, and paradoxically,
fundamentally the same. Computer software development occupies a central position in this paradox.
First, it should be noted that computer software development is not so much an industry as it is a
Second, many IT practitioners emphasize that computer software issues are complicated by the
multiplicity of ways that computer software quality may be critical in a supplier organization’s business.
For example:
 The supplier’s product may be complex software whose functional design requirements are specified
by the customer.
 The supplier may actually write most of its software product, or may integrate off-the-shelf packaged
software from subsuppliers.
 The supplier may incorporate computer software/firmware into its product, which may be primarily
hardware and/or services.
 The supplier may develop and/or purchase from subsuppliers software that will be used in the supplier’s
own design and/or production processes of its product.
However, it is important to acknowledge that hardware, processed materials, and services often
are involved in a supplier organization’s business in these same multiple ways, too.
What, then, are the issues in applying ISO 9001 to computer software development? There is general
consensus worldwide that:
 The generic quality management system activities and associated requirements in ISO 9001 are
relevant to computer software, just as they are relevant in other generic product categories (hardware,
other forms of software, processed materials, and services).
 There are some things that are different in applying ISO 9001 to computer software.
There is at this time no worldwide consensus as to which things, if any, are different enough to
make a difference and what to do about any things that are different enough to make a difference.
ISO/TC176 developed and published ISO 9000-3:1991 as a means of dealing with this important,
paradoxical issue. ISO 9000-3 provides guidelines for applying ISO 9001 to the development, supply,
and maintenance of (computer) software. ISO 9000-3 has been useful and widely used. ISO 9000-3
offers guidance that goes beyond the requirements of ISO 9001, and it makes some assumptions about
the life cycle model for software development, supply, and maintenance. In the United Kingdom a
separate certification scheme (TickIT) for software development has been operated for several years,
using the combination of ISO 9001 and ISO 9000-3. The scheme has received both praise and criticism
from various constituencies worldwide. Those who praise the scheme claim that it
 Addresses an important need in the economy to provide assurance for customer organizations that
the requirements for quality in software they purchase (as a separate product, or incorporated in a
hardware product) will be satisfied
 Includes explicit provisions beyond those for conventional certification to ISO 9001 to assure competency
of software auditors, their training, and audit program administration by the certification
 Provides a separate certification scheme and logo to exhibit this status publicly
Those who criticize the scheme claim that it
 Is inflexible and attempts to prescribe a particular life cycle approach to computer software development
which is out of tune with current best practices for developing many types of computer
 Includes unrealistically stringent auditor qualifications in the technology aspects of software
development, qualifications whose technical depth is not necessary for effective auditing of management
systems for software development
 Is almost totally redundant with conventional third-party certification to ISO 9001, under which
the certification body/registrar already is responsible for competency of auditors, and accreditation
bodies verify the competency as part of accreditation procedures
 Adds substantial cost beyond conventional certification to ISO 9001 and provides little added
value to the supply chain
In the United States a proposal to adopt a TickIT-like software scheme was presented to the
ANSI/RAB accreditation program. The proposal was rejected, primarily on the basis that there was
not consensus and support in the IT industry and the IT-user community.
At this writing:
 ISO/TC176 is revising ISO 9000-3 for the short term to bring it up-to-date with ISO 9001:1994
and to remedy some technical deficiencies.
 ISO/TC176 is planning the next revision of ISO 9001 with the long-term intention of incorporating
ISO 9001 quality assurance requirements stated in a way that will meet the needs of all four
generic product categories without supplementary application guideline standards such as ISO
 Various national and international groups, conferences, and organizations are discussing whether
there is enough of a difference to warrant a special program, and if so, what such a program should
look like.
The one thing that is currently clear is that no worldwide consensus exists.
Other Areas of Application. The special case of environmental management systems and
their relation to quality management systems has been discussed earlier in this section. This situation,
too, is a classic example of a problem-opportunity from the perspective of the ISO 9000 standards.
Companies are likely to have to do business under both sets of requirements: the ISO 9000
standards from ISO/TC176 and the ISO 14000 standards from ISO/TC207. The opportunity for
mutually beneficial consistency promises important benefits. These benefits relate to the operational
effectiveness of having one consistent management approach in both areas of the business activities
and can translate also into cost benefits of such a single approach. The ISO Technical Management
Board has mandated that TC176 and TC207 achieve compatibility of their standards.
In the United States and other nations, the compatibility of the ISO 9000 standards and the ISO
14000 standards is one part of the standardization job. The implementation part requires that similar
harmonization and compatibility be established in each nation in the infrastructure of accreditation
bodies, certification/registration bodies, and auditor certification bodies, operating under internationally
harmonized guidelines. At this writing the ISO 14000 infrastructure is in its infancy.
Various nations and regional bodies have established quality awards. The most widely known of
these are the Deming Award in Japan; the Malcolm Baldrige National Quality Award (MBNQA) in
the United States; and the European Quality Award, a European regional award. These awards incorporate
concepts and principles of Total Quality Management (TQM).
TQM means different things to different people. ISO 8402:1994 defines TQM as follows:
…management approach of an organization centered on quality, based on the participation of all its
members and aiming at long-term success through customer satisfaction, and benefits to all members of
the organization and to society.
For purposes of this section the criteria of the Baldrige Award or the Deming Award or the
European Quality Award can be considered to be an operational definition of full-scale TQM
Questions often are asked about the relationships between the criteria upon which these awards
are based and the content of the ISO 9000 standards. This discussion is in two parts: the relationship
to ISO 9001 and the relationship to ISO 9004.
Overall, it is important to understand that the purpose of the ISO 9000 standards is to facilitate
international trade. To achieve that purpose, the ISO 9000 standards focus on the supplier organization
functions that most directly affect product quality. The ISO 9000 standards are intended for
implementation by the large majority of supplier organizations. By contrast, the purposes of the
award criteria are (1) to select, from among all the supplier organizations in a nation or region, those
few organizations that exemplify the very best level of achievement in quality management and
(2) to provide criteria and guidelines for other organizations that may wish to improve in the direction
of becoming best of the best and are willing to make the substantial investment to achieve that
lofty level of quality performance.
Relationship to ISO 9001. ISO 9001 is a requirements standard for two-party contractual or
third-party registration use in support of international trade. Commensurate with this role, ISO 9001
focuses only on the functions that most directly affect product quality. It does not, therefore, deal
with questions of economic effectiveness and cost efficiency. It deals only with specific personnel
aspects and specific sales and marketing aspects that directly affect product quality. Thus, the scope
of ISO 9001 is narrower than the scopes of the cited national awards. For example, the MBNQA criteria
examine many specific items in seven broad categories of an organization’s activities. These
seven categories are: leadership, information and analysis, strategic planning, human resource development
and management, process management, business results, and customer focus and satisfaction.
The ISO 9001 requirements give greatest emphasis to the process management category of
MBNQA, and have lesser emphasis on the other categories.
In view of the differing purposes of ISO 9001 and the award criteria, there is a difference also in
the depth of examination of the supplier organization’s quality management system. The MBNQA
and ISO 9001 both embrace the concept that all work is accomplished by a process and that an organization’s
activities can be viewed as a network of processes. Both MBNQA and ISO 9001 recognize
the need to examine the approach, the deployment, and the results (Marquardt 1996b) in the
examination of a process.
In the late 1980s the author developed Table 11.8 to describe the relative depth of expectations in
terms of the appropriate assessment questions at various levels, including ISO 9001 and award criteria.
Questions are shown for approach, deployment, and results.
It is instructive to compare ISO 9000 registration (specifically ISO 9001 or ISO 9002) to the
achievement of an MBNQA award. As described in Table 11.9, ISO 9000 registration has in many
ways more modest requirements, but it does emphasize to a greater degree the necessity of a consistent,
disciplined, documented quality system, with periodic internal and external audits that serve
to hold the gains and institutionalize continuous improvement.
Relationship to ISO 9004. ISO 9004 is the standard in the ISO 9000 family that provides
to organizations quality management guidelines that cover a wider scope and greater depth than the
requirements in ISO 9001. The additional scope and depth go part way toward the scope and depth
of award criteria such as the MBNQA. In keeping with the purpose of the ISO 9000 standards, the
scope and depth is at a level that is achievable by a large proportion of organizations in the global
economy. Thus, ISO 9004 is deliberately positioned in an intermediate range, between ISO 9001
and the award criteria. The marketplace uniqueness of ISO 9004 is that it offers to organizations a
framework for building a quality management system that will be effective and efficient and that
will focus on features that have a direct effect on product quality, features that are fully consistent
with ISO 9001, ISO 8402, and other standards in the ISO 9000 family. This enables the organization
to use one consistent set of terminology that is internationally standardized and one consistent
Many organizations worldwide have adopted the strategic perspective that ISO 9001 provides
for them minimum adequate criteria for effective operation and for meeting the marketplace
requirements for quality. ISO 9004 provides a guideline to enrich and enhance the ISO 9001 baseline,
and to take deliberate, planned initiatives that build upon the baseline in the direction of TQM.
The award criteria, such as those of MBNQA, provide a comprehensive operational statement of
full-scale TQM. With this insight, the standards and the award criteria are compatible and complementary.
Both are necessary in the global economy.
Is there a defined process?
Is the process appropriate to the
needs of the function?
Is there documentation appropriate
to each person’s needs at
each organizational level?
Is the documentation controlled
(accurate, up-to-date, available
when and where needed)?
Is the process state of the art,
world class, best of the best for
this function?
Is there innovation, technology
advantage, cost advantage,
functional superiority in this
Does the excellence in this
process provide superior value
that is perceived by the
customer as exceeding
Is the process fully deployed?
Is a process used wherever this
generic function is implemented
in the organization?
Is every involved person trained
(understands the process, why it
exists, how to use it)?
Does every person have value
for the function?
Is the process deployed
consistently and universally to
accomplish this function in a
standard way with consistent,
transferable training
Does the consistency of the
results of this process provide
superior value that is perceived
by the customer as exceeding
Are the results meeting
Have quantitative metrics been
Are the metrics understood and
used by those involved?
Do the values of the metrics
show that the process is
appropriate to the function for
both quality management and
quality assurance purposes?
Are the metrics designed and
implemented in a way that
elicits value-adding behavior
Do the values of the metrics
clearly portray continuous
improvement and world-class
Do the financial results
demonstrate that customers
perceive superior value?
TABLE 11.8 Assessment Questions (for Any Process in a Quality Management System)
Approach Deployment Results
First level of depth
Second level of depth (e.g., ISO 9001 requirements)
Third level of depth (National or International Award Criteria)
TABLE 11.9 ISO 9000 Registration in Relation to Malcolm Baldrige Award
ISO 9000 requires:
 Adequate quality systems
 Objective evidence for every requirement
 Complete, controlled, up-to-date documentation
 Periodic surveillance audits that verify continuing compliance to requirements
MBNQA looks for:
 Best-of-the-best quality systems
 Clear evidence of product quality superiority
 Clear evidence of customer perception of superiority
 Historic trends that lend credence to one-time audit
Kolesar, P. J. (1994). “What Deming Told the Japanese in 1950.” Quality Management Journal, vol. 2, issue 1,
pp. 9–24.
Marquardt, D. W. (1994). “Credibility of Quality Systems Certification: How to Deal with Scopes of
Certification, Conflicts of Interest and Codes of Conduct.” In Peach, R. W. (ed.), ISO 9000 Handbook, 2d ed.
Irwin Professional Publishing, Fairfax, VA, chap. 17.
Marquardt, D. W.(1995). “The Missing Linkage in ISO 9001: What’s Being Done About It?” ASQC 49th Annual
Quality Congress Proceedings, American Society for Quality Control, Milwaukee, pp. 1056–1061.
Marquardt, D. W. (1996a). “The Importance of Linkages in ISO 9001.” ISO 9000 News, vol. 5, no. 1, pp. 11–13.
Marquardt, D. W. (1996b). “The Functions of Quality Management in Relation to the P-D-C-A Management
Cycle.” Manuscript submitted for publication.
Marquardt, D. W., et al. (1991). “Vision 2000: The Strategy for the ISO 9000 Series Standards in the ‘90s.”
Quality Progress, May, pp. 25–31. Reprinted in ISO 9000 Quality Management (an ISO standards compendium
containing all standards in the ISO 9000 family, plus the Vision 2000 paper), 5th ed. (1994), ISO Central
Secretariat, Geneva. Also reprinted in a number of other countries and languages, and in the ISO 9000
Handbook, op cit., chap. 11.
Robert C. Camp
Irving J. DeToro
Benchmarking: Definition 12.2
Benchmarking: Objectives 12.2
Practices and Performance Levels 12.3
Phase 1: Planning 12.3
Phase 2: Analysis 12.3
Phase 3: Integration 12.3
Phase 4: Action 12.3
Phase 5: Maturity 12.4
Benchmarking Triggers 12.4
Benchmarking Teams 12.5
Documentation 12.5
Internal Benchmarking 12.6
Competitive Benchmarking 12.6
Functional Benchmarking and World-
Class Leaders 12.7
Partnering 12.7
Internal 12.7
External 12.7
Original 12.8
GAP 12.9
Decide Who Needs to Know 12.12
Select the Best Presentation Vehicle
Organize Findings 12.12
Present Recommendations 12.13
Revise Operational Goals 12.13
Analyze the Impact on Others 12.13
Secure Management Approval 12.13
Set Implementation Priorities 12.14
Show Revisions to the Performance Gap
Develop Action Plans 12.15
Behavioral Benefits 12.19
Competitiveness 12.19
This section defines benchmarking and outlines the 10-step benchmarking process as developed in
Camp (1989 and 1994). It summarizes the activities of typical benchmarking teams, including their
objectives, tasks, and responsibilities.
The hottest and least understood new term in the quality field is “benchmarking.” Xerox does it. Ford
does it. GTE, AT&T, DEC, TI, duPont, HP, J&J, IBM, and Motorola do it. Just what is it?
Benchmarking is an ongoing investigation and learning experience. It ensures that the best practices
are uncovered, adopted, and implemented. Benchmarking is a process of industrial research that
enables managers to perform company-to-company comparisons of processes and practices to identify
the “best of the best” and to attain a level of superiority or competitive advantage.
Benchmarking is a method of establishing performance goals and quality improvement projects
based on industry best practices. It is one of the most exciting new tools in the quality field.
Searching out and emulating the best can fuel the motivation of everyone involved, often producing
breakthrough results.
The Japanese word dantotsu—striving to be the best of the best—captures the essence of benchmarking.
It is a positive, proactive process to change operations in a structured fashion to achieve
superior performance. The purpose of benchmarking is to gain competitive advantage.
Benchmarking: Definition. The formal definition of benchmarking is “The continuous
process of measuring products, services, and practices against the company’s toughest competitors
or those companies renowned as industry leaders.” (Camp 1994).
Benchmarking Objectives. The purpose of benchmarking is derived primarily from the
need to establish credible goals and pursue continuous improvement. It is a direction-setting process,
but more important, it is a means by which the practices needed to reach new goals are discovered
and understood.
Benchmarking legitimizes goals based on an external orientation instead of extrapolating from
internal practices and past trends. Because the external environment changes so rapidly, goal setting,
which is internally focused, often fails to meet what customers expect from their suppliers.
Customer expectations are driven by the standards set by the best suppliers in the industry as well
as by great experiences with suppliers in other industries. Thus, the ultimate benefit of benchmarking
is to help achieve the leadership performance levels that fully satisfy these ever-increasing customer
Benchmarking is an important ingredient in strategic planning and operational improvement. To
remain competitive, long-range strategies require organizations to adapt continuously to the changing
marketplace. To energize and motivate its people, an organization must:
 Establish that there is a need for change
 Identify what should be changed
 Create a picture of how the organization should look after the change
Benchmarking achieves all three. By identifying gaps between the organization and the competition,
benchmarking establishes that there is a need. By helping understand how industry leaders do
things, benchmarking helps identify what must be changed. And by showing what is possible and
what other companies have done, benchmarking creates a picture of how the organization should
look after the change.
Embarking on a benchmarking activity requires acceptance of the following fundamentals:
 Know the operation. Assess strengths and weaknesses. This should involve documentation of work
process steps and practices as well as a definition of the critical performance measurements used.
 Know industry leaders and competitors. Capabilities can be differentiated only by knowing the
strengths and weaknesses of the leaders.
 Incorporate the best and gain superiority. Adapt and integrate these best practices to achieve a leadership
Practices and Performance Levels. Benchmarking can be divided into two parts: practices
and performance levels. From experience, most managers now understand that benchmarking should
first focus on industry best practices. The performance levels that result from these practices can be
analyzed and synthesized later. Having identified the best practices of several companies, the lessons
learned can be integrated to create world-class work processes. At that stage, the expected performance
from these work processes can be determined so that service levels that are superior to the
best of the competitors’ can be delivered.
When preparing for benchmarking, it is important to engage line management so that the findings
are understood and accepted and result in a commitment to take action. This requires concerted management
involvement and carefully designed communications to the organization that must implement
the action plans.
The 10-step process for conducting a benchmarking investigation consists of the following five
essential phases (see Figure 12.1).
Phase 1: Planning
 Decide what to benchmark. All functions have a product or output. These are priority candidates
to benchmark for opportunities to improve performance.
 Identify whom to benchmark. World-class leadership companies or functions with superior work
practices, wherever they exist, are the appropriate comparisons.
 Plan the investigation, and conduct it. Collect data sources. A wide array of sources exists, and a
good starting point is a business library. An electronic search of recently published information on
an area of interest can be requested. Begin collecting. Observe best practices.
Phase 2: Analysis
 It is important to have a full understanding of internal business processes before comparing them
to external organizations. After this, examine the best practices of other organizations. Then measure
the gap.
 Project the future performance levels. Comparing the performance levels provides an objective
basis on which to act and helps to determine how to achieve a performance edge.
Phase 3: Integration
 Redefine goals and incorporate them into the planning process.
 Communicate benchmarking findings and gain acceptance from upper management.
 Revise performance goals.
 Remember, the competition will not stand still while organizations improve. Thus, goals that
reflect projected improvement are necessary.
 On the basis of the benchmarking findings, the targets and strategies should be integrated into business
plans and operational reviews and updated as needed.
Phase 4: Action
 Best practices are implemented and periodically recalibrated as needed.
 Develop and implement action plans.
 Monitor progress.
 Recalibrate the benchmarks.
Phase 5: Maturity
 Determine when a leadership position is attained. Maturity is achieved when best practices are
incorporated in all business processes; when benchmarking becomes a standard part of guiding
work; and when performance levels are continually improving toward a leadership position. Assess
benchmarking as an ongoing process.
Benchmarking Triggers. Events that cause a benchmarking project to be initiated usually fall
into one of three groups: problem, innovation, or policy.
FIGURE 12.1 The formal 10-step benchmarking process. (Quality
Resources, a division of The Kraus Organization Limited, White Plains, NY,
through ASQC Quarterly Press.)
Problem. If a crisis occurs within an organization, such as a major cost overrun or a major customer
threatening to cancel an existing contract, the topic for an improvement project may be apparent.
Innovation. If the organization becomes aware of some innovative technology, practice, or process
employed by another organization, this information may well cause an organization to commission
a benchmarking study.
Policy. If the organization does not have a significant problem or need to understand an innovative
practice, then selecting a benchmarking project may be difficult. This is especially true for an organization
employing a total quality management (TQM) philosophy. However, a well-established
quality management effort includes a strategic planning process. In organizations with such efforts,
it is common for one of the outputs of the planning process to be a list of nominations for appropriate
benchmarking projects.
Benchmarking Teams. Benchmarking is conducted by teams consisting of individuals with
direct operational experience and knowledge of the process. Members should possess analytical,
research, process documentation, and team facilitation skills. These requirements favor candidates
with engineering or technical backgrounds, and those with research experience. Benchmarking
teams are typically commissioned by the process champion. Teams rarely function effectively if they
consist of more than 9 to 12 members. Team size of 3 to 6 is preferred. Large teams can be considered
but will most likely break down into small subgroups to do their work.
The first step in determining what to benchmark is identifying the product or output of the business
process or function. Fundamental to this is the development of a clear mission statement detailing
the reason for the organization’s existence, including key outputs expected by its customers and critical
to fulfilling the mission successfully. Next, each function’s broad purposes should be broken
down into specific outputs to be benchmarked. Outputs should be documented to a level of detail
necessary for analyses of key tasks, handoffs, and both in-process and end results measurements, and
for quality, cost, and delivery analyses.
One good way to determine which outputs are most in need of benchmarking is to pose a set of
questions that might reveal current issues facing the function. Questions might focus on customer
care (including service), cost, or perception of product offerings. Another way to identify key outputs
is to convert the problems, issues, and challenges faced by the function into problem statements
and then to develop these into a cause-and-effect Ishikawa diagram. The causals in the diagram are
candidates for benchmarking.
Successfully completing a benchmarking project is dependent on selecting a worthwhile topic. It
should not be too large, too trivial, or one that would not secure a performance advantage. To avoid
these problems, the topic should be selected after some analysis to ensure that the organization’s
resources are being justifiably expended. The intent of step 1 is to confirm a topic already selected,
to narrow or broaden the scope of a project, or to select a topic that could best contribute to the organization’s
If resources are constrained, consideration should be given to improving the earliest possible area
in a process since, if that is improved, there may be a beneficial effect on all subsequent activities.
The candidate process should be tested for reasonableness by asking such questions as: Is this the
area customers complain about the most? Are there areas with major cost overruns that need attention?
Is there something that if not fixed immediately will be affected in the marketplace?
Documentation. The team must describe how the work is currently performed by preparing
detailed flowcharts. This is essential because it helps the team gain consensus on how the work is
actually performed, the time it takes to perform the work, the cost, and the errors created in the current
work flow. This understanding is essential because comparison to a superior system will not
reveal deficiencies in the current system unless such understanding and documentation exists.
A brief, two- or three-page, description of the benchmarking project should be prepared and circulated
among sponsors, managers, stakeholders, and other interested parties. This document captures
all the thinking that has gone into selecting the project, the potential resources required, and
the expected outcome. As more information is gathered and the team completes other steps in the
10-step benchmarking model, the project description can be updated and used as a means of keeping
sponsors informed.
The difficulty is in identifying which leading-edge companies possess processes that truly have best
practices. Determining whom to benchmark against is a search process that starts with consideration
of, in broad terms, an operation’s primary competitors and then extends to leading companies that
are not com