International Journal of Lean Six Sigma

Emerald Article: Design for Six Sigma: caveat emptor

Gregory H. Watson, Camille F. DeYong

Article information:
To cite this document: Gregory H. Watson, Camille F. DeYong, (2010),"Design for Six Sigma: <IT>caveat emptor</IT>", International
Journal of Lean Six Sigma, Vol. 1 Iss: 1 pp. 66 - 84
Permanent link to this document:
http://dx.doi.org/10.1108/20401461011033176
Downloaded on: 04-11-2012
References: This document contains references to 69 other documents
Citations: This document has been cited by 2 other documents
To copy this document: [email protected]
This document has been downloaded 921 times since 2010. *

Access to this document was granted through an Emerald subscription provided by INSTITUTO TECNOLOGICO Y DE ESTUDIOS SUPERIO
MONTERREY
For Authors:
If you would like to write for this, or any other Emerald publication, then please use our Emerald for Authors service.
Information about how to choose which publication to write for and submission guidelines are available for all. Please visit
www.emeraldinsight.com/authors for more information.
About Emerald www.emeraldinsight.com
With over forty years' experience, Emerald Group Publishing is a leading independent publisher of global research with impact in
business, society, public policy and education. In total, Emerald publishes over 275 journals and more than 130 book series, as
well as an extensive range of online products and services. Emerald is both COUNTER 3 and TRANSFER compliant. The organization is
a partner of the Committee on Publication Ethics (COPE) and also works with Portico and the LOCKSS initiative for digital archive
preservation.
*Related content and download information correct at time of download.
 The current issue and full text archive of this journal is available at
www.emeraldinsight.com/2040-4166.htm

IJLSS
1,1 Design for Six Sigma:
caveat emptor
Gregory H. Watson and Camille F. DeYong
66 School of Industrial Engineering and Management, Oklahoma State University,
Stillwater, Oklahoma, USA

Abstract
Purpose – The purpose of this paper is to describe the historical approach to concurrent engineering
(CE) which has resulted in product line management (PLM) and then evaluates the theoretical models
that have been proposed for design for Six Sigma (DFSS) in order to determine which model is able to
provide the most consistent approach with historical development of PLM.
Design/methodology/approach – The approach begins with an overview of the approach taken by
the Union of Japanese Scientists and Engineers ( JUSE) in the development of a coherent quality
methodology for structured analysis and problem solving – the Deming Wheel of plan-do-check-act
(PDCA) which has become the standard model in Japanese total quality management to define a logical
decomposition in process management. In Japan, PDCA is the single logical model which has been
broadly accepted as the construct for understanding how to develop both strategic and operational
quality methods. The second step in the approach is to examine a similar American development of the
model for statistical problem solving that is applied in the Six Sigma method for statistical problem
solving: define-measure-analyze-improve-control (DMAIC). Next, the paper examines the historical
sequence in the way the product development process has developed over the past forty years, with
emphasis on its military origins (especially CE) and which resulted in the generic model for PLM.
The final part of this paper examines the models that have been proposed to implement DFSS over
the past ten years and evaluate their logical congruence with the engineering community’s design
process.
Findings – Problems in alignment with the engineering design process were identified with all of the
DFSS models and with the non-structured or “heuristic” approach to developing a coherent body of
knowledge related to DFSS.
Originality/value – This paper provides a challenge to the quality community as well as to the
academic community. The paper points out the need for rigorous examination of logical models that
are proposed for guiding the thinking of practitioners in the use of quality methods for both the
engineering of products and business systems. An expose of lack of rationality in the way an approach
to DFSS has been investigated calls for more responsibility in the management of the development of
this body of knowledge.
Keywords - Six sigma, Total quality management, Research and development, Production management
Paper type General review

Introduction
The need for agreed-upon logical models for the coordination of human work is a
phenomenon of the twentieth century. It was first recognized as a result of scientific
management, which has led to the Japanese plan-do-check-act (PDCA) model and more
recently to the development of a model for improvement methodologies in the American
International Journal of Lean Six
Sigma
Vol. 1 No. 1, 2010 Assistance from Soichi Shimizu in providing details of Japanese quality history is gratefully
pp. 66-84 acknowledged. In addition, this research has its origins from a challenge by Noriaki Kano to
q Emerald Group Publishing Limited
2040-4166
investigate the intellectual roots of PDCA and DMAIC. The authors thank him warmly for his
DOI 10.1108/20401461011033176 insight.
Six Sigma movement and creation of the define-measure-analyze-improve-control Design for Six
(DMAIC) model. Efforts over the past ten years have been pursuing a similar Sigma: caveat
development for what was generically referred to as design for Six Sigma (DFSS) since
the late 1990s. emptor
Systems thinking and concept of mental models has evolved out of the body of work
that was founded on the pragmatic philosophy of William James (1842-1910) and
expanded upon with the logical approach to thinking of John Dewey (1859-1952). 67
Systems thinking commenced in the mid-1940s and expanded greatly during the late
1960s with the work of Ludwig von Bertalanffy (1901-1972) (von Bertalanffy, 1968),
Kenneth E. Boulding (1910-1993), C. West Churchman (1913-2004), Russell L. Ackoff
(1919-2009) (Ackoff and Emery, 1972), and Jay W. Forrester and his school of systems
dynamics (Forrester, 1961, 1968). Most recently systems thinking evolved into a
popular approach under the guidance of Senge (1990). The purpose of this paper is not
to investigate this chain of development, but to make use of its approach to examine the
models for quality thinking. However, we should note that the development of PDCA
preceded development of the general systems method and is therefore an important
source of information for learning about both the systems approach and the
methodology that it promotes.
From the observations of modeling history, it is clear that when the model for a
concept is first developed that there is an initial divergence caused by the different
structures proposed until the scientific method prevails and there is a convergence upon
an accepted model, borne out of rigorous examination of the alternatives. To establish
this history in the quality movement, we shall first examine the evolution of the
Japanese PDCA model as an instructive example of this phenomenon and then use this
perspective to understand what lessons that should be learned in the early evolutionary
stages of DFSS.

Lessons learned from Japanese development of PDCA model

The Japanese PDCA methodology progressively evolved from the scientific
method originated by Sir Francis Bacon (1561-1626) in his masterpiece Novum
Organum (Bacon, 1620) in which he identified three steps of investigation: hypothesis,
experiment, and evaluation. Adam Smith (1723-1790) in An Inquiry into the Nature
and Causes of the Wealth of Nations (Smith, 1776, 2000) added to the dialog by
separating work according to a division of labor. A further distinction between
“planning” (management’s specialization) and “doing” (the job of workers) was made
by Frederick W. Taylor (1856-1915) in his The Principles of Scientific Management
(Taylor, 1911, 1990) whereby managers apply scientific management principles to
planning the work and the workers perform the tasks. The Japanese had summarized
Taylor’s method using a model that they described as “plan, do, see” to evaluate the
outcome of the execution of the scientific plans (Kano, 2005). This development
occurred before the model of Walter A. Shewhart (1891-1967) was introduced during the
early lectures of W. Edwards Deming (1900-1993) (Kano, 2005). Kaoru Ishikawa
(1915-1989) had previously credited Taylor’s work with influencing the Japanese
development of the PDCA methodology (Ishikawa, 1985).
Shewhart (1939) reinterpreted Taylor’s model for mass production as a three step
process: specification (making a hypothesis); production (carrying out an experiment);
and inspection (testing the hypothesis) and thereby linked it to the work of
IJLSS Francis Bacon. This was the logical basis for the models proposed by Dr Deming. In a
1,1 paper relating the origins of PDCA, Noriaki Kano (b. 1940) of the Tokyo Science
University traced the development of Japanese PDCA from its 1951 conceptual origin in
Deming’s lecture (Kano, 2005). The following description draws heavily upon the
findings of Kano and his investigation of the original source material in the Japanese
Scientists and Engineers (JUSE) archives.
68 The starting point of this history is the initial lecture by Deming to JUSE in 1951.
In the lecture notes prepared by Kenichi Koyanagi (1903-1965) of Deming’s 1951 lecture,
he describes an eight-step model proposed for viewing production as a system: quality
consciousness; quality responsibility and infrastructure requirements of design and
production; study of various elements such as raw materials, machines workers and
their attributes; development and design; manufacturing; testing; redesign; and sales.
Ishikawa simplified this model to include just four steps: design, production, sales, and
survey. In 1952, according to the lecture notes of Shigeru Mizuno (1910-1989), a member
of the original JUSE Research Committee on Quality Control (1948), Deming changed
his model and used the following steps to define his systems model as a Shewhart Cycle
(his observations were repeated when he took notes on Deming’s 1954 lectures):
(1) design the product (with appropriate tests);
(2) make it; test it on the production line and in the laboratory;
(3) put it on the market;
(4) test it in service through market research; find out what the user thinks of it and
why the non-user has not bought it; and
(5) redesign the product in light of consumer reactions to price and quality and
continue around and around the cycle.

Up until this point in time, there is no Japanese record of a “PDCA” model in Deming’s
lectures. Joseph M. Juran (1904-2008) also lectured Japan on quality management in
1954 and one student of his lecture believes that he may have helped to stimulate the
thinking on the PDCA model (Kolesar, 2008) In any case, Kano observed that following
the 1959 lecture by Deming, Mizuno recommended that the PDCA model was a more
meaningful way to describe Shewhart’s Model and afterwards this PDCA model was
named “the Deming model” in honor of Deming’s introduction of the core of this logical
concept.
At this point in time PDCA became the Japanese standard model for improvement
and problem solving. The model was promoted by Ishikawa and used by the
JUSE Research Committees that developed the theory of quality management and
investigated the best way to apply this theory in the quality practices of organizations
(for an example of how this approach was achieved relative to planning processes, see
Akao, 1991).
While the PDCA cycle is called the Deming Cycle, it is noteworthy that Deming did
not make a specific claim to originate this cycle in his major works (Deming, 1986,
1993). In his book, Out of the Crisis, Deming expounded upon a model that presented a
six-step process:
(1) What could be the most important contributions of the 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.
 (2) Carry out the change or test decided upon, preferably on a small scale. Design for Six
(3) Observe the effects of the change or test. Sigma: caveat
(4) Study the results. What did we learn? What can we predict? emptor
(5) Repeat Step (1) with knowledge accumulated.
(6) Repeat Step (2) and onward (Deming, 1986).
69
Those familiar with the literature on quality management will recognize that Deming
was expounding upon the “spiral” approach to the iterative application of PDCA or
“turning the PDCA wheel” as the continuous improvement process operates over time
to deliver ever enhanced performance results (creation of this “spiral” concept is most
appropriately credited to Juran (1974).
Later in his book On the New Economics (Deming, 1993) he redefined PDCA by
calling it plan-do-study-act which he claimed was a better description of the method:
.
plan a change or test aimed at improvement;
.
do – carry out the change or test (preferably on a small scale);
.
study the results (What did we learn? What went wrong?); and
.
act – adopt or abandon the change or test and continuously improve by
completing the cycle again.

Of course, by this time the PDCA cycle was widely referred to as the Deming Cycle and
his suggestion of an alternative wording to the model can be cynically viewed as a way
to more formally attach his name to the origins of the model.
However, while Japan uses PDCA as a general approach to describing how
processes operate, the Japanese have also amplified the application of the PDCA method
specifically for problem solving in conjunction with a set of basic quality tools and
statistical methods (paralleling the development of DMAIC in the United States during the
1980s). This is an important historical precedent that is noteworthy for considering
the development of the Six Sigma decision models for both problem solving (DMAIC) and
process and product development (DFSS).

Improvements extend PDCA to problem-solving project

Invention credit for this development should be given to Juran (1973) who suggested
using a problem-solving quality control (QC) story in his initial lectures to Japanese
managers (Kolesar, 2008). This was the approach was standardized for QC circles and
broadly disseminated as the quality circle movement grew after its establishment in
1962 by Ishikawa. Later, Tatsuo Ikezawa (b. 1928) revised Juran’s model and
transformed it into QC Story for Management (Ikezawa, 1970) and Hitoshi Kume edited
a textbook, Statistical Methods for Quality Improvement, for the Association for
Overseas Technical Scholarship an organization that taught the methods of Japanese
total quality control (TQC) and problem solving to business leaders and scholars
coming to Japan from around the world (Kume, 1985). It was this methodology that was
transferred to Florida Power & Light (FPL) as part of its application for the JUSE
Deming Application Prize (1985-1987) (Kano, 2005). Katsuya Hosotani (b. 1938) referred
to this approach as the as the QC seven-step formula (Hosotani, 1989). Further
developments of this idea within a JUSE team lead by Kano extended the QC story from
IJLSS problem solving to project improvement management in the book Task Achieving QC
1,1 Story for QC Circles (Kano, 1993). In 1997, Kano proposed application of this method for
the purpose of achieving breakthrough and creativity in his book Task Achieving QC
Story for Management (Kano et al., 1997).
There is a most valuable lesson to learn in this modern history of Japanese quality
development. Across a 50-year period the Japanese have been able to maintain
70 consistency in development of their logic. Today, the Japanese standard approach to
total quality management (TQM) still uses PDCA. By maintaining a single logical
model, the thinking behind the model has evolved to become more and more explicit
and clear and the fundamental logic of the PDCA model has not changed. The initial
contributions of Shigeru Mizuno and Kaoru Ishikawa were preserved across a
generation of Japanese quality thinkers who used a foundation generated from the ideas
of Taylor, Shewhart, and Deming. In the end, they honored Deming by attaching his
name to the cycle as it was his wheel model that gave them the idea that there could be
such a generalized improvement model. After having described the intellectual history
of PDCA, it is instructive to compare and contrast the logical development of the Six
Sigma logical model that has largely replaced the use of PDCA as an American
development for a logical model used for Six Sigma problem solving: DMAIC.

Examining the origins of DMAIC

Faced with this substantial history for the logical progress in advancement of PDCA,
creation of a unique Six Sigma DMAIC model seems strange. Indeed, Kano’s point
regarding a shift away from PDCA is a criticism about the Six Sigma method that should
not be taken lightly. Indeed this causes problems on two levels: it shows a potential lack
of respect for the tradition of PDCA and it also demonstrates ignorance of the legacy of
Japanese quality development. Indeed, the generation of a new model without bridging
from the previous history appears to illustrate a “not invented here” syndrome or the
practice of “quality du jour” – switching to a new quality method, slogan or “catch
phrase” for the sake of its “newness.” Such change is not adding value, but in reality these
decisions destroy intellectual value by grabbing something new and rejecting historical
lessons without applying critical examination to alignment the new understanding with
past lessons learned and thereby form an improved system for quality management.
It seems that a major opportunity was missed. So, how did DMAIC process evolve?
In the 1984, Bill Smith (1929-1993) joined Motorola after a long career with the
General Systems Corporation where he worked for Armand V. Feigenbaum, inventor of
the concept of TQC which employs the PDCA model and was a stimulus for Ishikawa’s
(1985) TQC approach. Thus, the PDCA model should have been inherent as a core
element in the development of an approach to Six Sigma. This was not to be the case. In
1987, Smith became vice president of quality for the communications sector which
launched a critical element in the Motorola quality program – the Bandit project
(Watson, 1994). At the start we should note that the DMAIC process has not always been
the process used in Six Sigma. The initial Six Sigma process that was defined in the
Motorola University Design for Manufacturing training program launched in March
1988 was defined by Smith as “six steps of Six Sigma” (Motorola Corporation, 1990):
(1) Identify the product created or the service provided.
(2) Identify the customer(s) for the product or service, and determine what is
important to them (requirements in measurable terms).
 (3) Identify your needs in order to provide the product or service. Design for Six
(4) Define the process for doing your work. Sigma: caveat
(5) Mistake proof the process and eliminate defects and waste. emptor
(6) Ensure continuous improvement by measuring, analyzing and controlling the
improved process.

Clearly, this definition contains elements of what would become known as DMAIC, but
71
its approach is actually close to the original process that Deming defined which
generated PDCA in Japan. This “six-step” process also integrates elements of lean
production from Toyota and was also based on the influence of the thinking of Juran
who was one of the quality gurus that Motorola consulted ( Juran, 1968).
It should be noted that during the second half of the 1980s decade, Motorola was
fighting a desperate commercial battle to revive its pocket pager business and the
communications sector where Smith lead quality initiative for the Bandit Project. This
initiative benchmarked and “stole” the best ideas from across the world to incorporate
them into a totally new business model that was capable of resurrecting market share
and turning around Motorola’s losses to Japanese competitors. In the light of this intent,
it becomes a little clearer why a “fresh” approach to problem solving was desired –
especially one that broke-away and ignored Japanese-dominated quality developments.
To stimulate thinking about what to do differently, the Motorola Communication
Sector’s quality department requested that its Japanese subsidiary provide it with the
details of the JUSE QC Story publications so they could understand the improvements
that were underway to the PDCA method. Kano reports that information was provided
by Kazuko Nishizaki in 1989 who reported that Motorola incorporated this information
into its initiative to deploy the Six Sigma improvement methodology (Kano, 2005).
Thus, it is clear that Motorola was aware of the latest developments in Japanese
thinking at the time that it developed the first generation of the DMAIC model.
However, strong energy associated with the need to resurrect American business
following its startling loss to Japan of computer memory chip business in the early
1980s caused a fervor that is best labeled as “Buy American.” Thus, the solution to
quality improvement had to have a uniquely “American” flavor that could be seen as
“better” than the Japanese developments.
In the period from its founding in 1990 to 1993, the Six Sigma Research Institute of
Motorola University, under the direction of its founder Mikel Harry re-examined a
variety of corporate measures and models. The original “six steps of Six Sigma” model
was considered as were ideas on machine and process control studies from Perez-Wilson
(1989) and Harry’s doctoral research on the application of sequential logic filters for
problem solving which were paired to specific methods and tools. To distinguish
their approach from the American use of PDCA and to focus teams on the statistical
aspects, the Six Sigma Research Institute initially named their problem-solving
process measure-analyze-improve-control (MAIC), based on the work of Harry, and
supplemented these steps with the logical elements from Perez-Wilson.
The MAIC problem-solving model is based on these roots from as well as a number
of external sources. This approach defined a sequence of methods to reduce the number
of variables contained in a scientific experiment by processing data using series of four
logical filters: recognition, classification, analysis (these steps are roughly analogous to
the DMA steps of DMAIC), and control (Harry, 1985). Research conducted by the
IJLSS Motorola Corporate Quality Group into the ideas of the major quality gurus (Deming,
1,1 Juran, Crosby, and Feigenbaum) lead to incorporation of the “best ideas” from these
“leading thinkers.”
In the end, the MAIC method added to the work of Harry and Perez-Wilson: Joseph
M. Juran’s quality journey along with Dorian Shainin’s advanced diagnostic tools;
Genichi Taguchi’s loss function; basic graphical analysis tools (Rumbler-Brache
72 process diagrams); basic QC tools found in the Japanese QC Story and PDCA model;
team-based problem solving as found in the Japanese Quality Circles: cycle time
reduction and mistake proofing from the Toyota Production System; and more
advanced tools such as statistical process control and designed experiments (Harry,
1993, 1997). A formulation of the sequence of analyses contained in MAIC created a more
rigorous analysis process than had been applied in the American TQM up to this point
in time; however, it was not greatly different from the description of how to apply PDCA
with statistical methods in QC story methodology as promoted by JUSE (Kume, 1985)
which had formed the basis of the approach to quality improvement taken by FPL at this
same time period and was disclosed through collaboration among leading American
companies in their support of the FPL bid to gain the Deming Application Prize.
Evidence in literature suggests that up to the time Six Sigma was implemented at
AlliedSignal in 1993 that MAIC was the process used for analysis (Harry, 1993).
However, MAIC did not stay stationary for a long time. During the 1997, deployment of
Six Sigma in General Electric (GE) management recognized that the problem-solving
process did not do an adequate job to establish a business reason to improve. So, GE
encouraged adding a “define” step to precede the MAIC steps. Thus, by 1997 the
consultant-recommended approach to Six Sigma evolved to DMAIC (Harry, 1997).
(When the author was working with Nokia mobile phones on its Six Sigma program in
1998, MAIC was the initial model used, but Six Sigma Academy soon transitioned to the
new model and added the “define” step as a prelude to MAIC.) The purpose of “define”
was to establish the boundaries of the problem, define the performance indicator,
establish the potential entitlement gain available to improve the current state,
and to initiate a charter from management to commission a DMAIC improvement
project (Harry, 2000b). However, within three years, Harry had introduced three other
steps to further expand DMAIC. “Recognize” was defined as a precursor to DMAIC – to
align with strategic change and problem-solving activity for the project selection and
coordinate projects across functional areas and focus improvement on the full business
system. Harry also recommended “standardize” and “integrate” follow the DMAIC
process to assure that project recommendations were put into practice, standardized in
the work experience, and leveraged to all applicable applications of the learning. These
steps consolidate improvements into standard work and integrate lessons learned and
new knowledge into all potential applications within a business system and paralleled
the extension of the Japanese PDCA model to link with an standardize-do-check-act
(SDCA) model as promoted by Kano in the methodology he named Task Achieving QC
Story (Kano, 2005; Harry, 2000b; Harry and Schroeder, 2000).
The final evolution of the DMAIC model was initiated by George (2001) to formally
recognize integration of the lean production tools and methods in the DMAIC logic
whereby he proposed changing the name of the Six Sigma initiative to Lean Six Sigma
(LSS). However, this extension was in reality a moot point, because even the earliest
methods of Six Sigma proposed by Motorola had included work process simplification,
cycle time reduction, mistake proofing, and all the other tools that were part of the Design for Six
Toyota Production System (Motorola Corporation, 1990; Harry, 1993). So the shift in Sigma: caveat
label from “Six Sigma” to “LSS” was a marketing invention and another example of the
“not invented here” syndrome which did not add any substantial improvement to emptor
the way that DMAIC analysis had been taught or conducted before that time. However,
the power of market-based advertising; the desire for “new” ideas; and a “backlash”
against “Cowboy quality” all combined to make this new LSS label broadly adopted 73
(Maguire, 1999).

Understanding the new DMAIC tradition

Thus, for most American companies, DMAIC became a popular solution to problem
solving and it was perceived as an advance over the standard American style PDCA
methodology found in companies applying the early versions of TQM. Why did
American management welcome DMAIC over PDCA, or at least over PDCA as
practiced American style? We can postulate a number of reasons:
.
Most-importantly, American companies did not want to look like they were using
the methods of Japanese competitors to turn back the threat. We must remember
that during this time there was a strong patriotic backlash on both sides of the
Pacific (Morita and Ishihara, 1991). Many improvements made by Japanese
companies were indeed incorporated without attribution (although in fairness, it
should be noted that there has not been much in the way of attribution for many
of the sources of process tools that were incorporated into DMAIC).
.
Many American applications of TQM and PDCA focused on the extensive use of
structured opinion analysis tools (e.g. brainstorming, affinity diagrams,
multi-voting) rather than data analysis tools. Thus, often the result of a TQM
project was just agreement on the common team wisdom rather than creation
of profound knowledge of process performance based on statistical analysis of
the problem.
.
There was a lack of connectivity among the various statistical tools that were
included in the basic TQM toolkit. Indeed, many quality presentations in the mid
to late 1980s would picture a “TQM Umbrella” covering a group of statistical tools
and methods as a means to illustrate what was included. But, this graphic picture
failed to communicate what should be the logical order or sequence of application
for these tools. This lack of structure made it seem that all tools were equally
valuable and could be applicable whenever desired by a team. This led many
teams to select their “favorite tool” for analysis and permitted them to ignore tools
or methods that they did not understand or found more difficult to apply.
.
There was an unclear pathway for analysis from step to step – while Japanese
PDCA was closely linked to the set of “seven basic QC tools” but PDCA
improvements reflected in the QC story processes developed under the auspices
of JUSE were not widely exposed to Americans and the momentum to correct
the same problem took over as many companies collaborated in the work of the
Six Sigma Research Institute so DMAIC took on a life of its own that was
independent from the Japanese refinements of PDCA.
.
Previous implementations of statistical process control (e.g. based on the author’s
personal experience in the early to mid-1980s at Hewlett-Packard) did not link
IJLSS “scientific measurement” principles to the use of support process control methods
1,1 – or define what would be a good follow-up process to investigate concerns about
special cause variation (this was addressed early in the development of MAIC and
was seen as a major breakthrough in thinking (Perez-Wilson, 1989; Harry, 1993).
.
There is also an American arrogance that dictates “Made in America” is a better
label for just about everything! Thus, industry had a prejudice toward accepting
74 “news” about this method and shifting to what it perceived as a different “quality
product” even though perhaps the largest difference was in the labeling.

What is the next phase of DMAIC evolution?

The biggest problem in DMAIC was its weak front end that aligns with strategy and
identifies which projects should be defined. Harry recommended adding a recognize
step to link DMAIC projects to strategic change management. In addition, DMAIC
fell apart in its concluding steps as control does not aptly describe how to close a project
and effect a change. Perhaps this is an issue with DMAIC that could have been healed by
using the updated PDCA logic. In Japan, PDCA is linked to SDCA standardization cycle.
The combination of these two cycles accounts for the activity to change and standardize
(in other words to manage effectively) an organization so that change projects are
integrated into daily management. While PDCA provides a front-end analysis for
problem solving, once the system passes a “check” for stability, then it transitions
into the SDCA cycle. Instead of taking the Japanese approach, Harry proposed adding
the Standardize and Integration steps to DMAIC (Harry, 2000a). Despite all its
problems, DMAIC was actually much better understood and developed than was the
approach to DFSS – this is the final focus of this paper. When the DMAIC approach can
provide no further improvement and more capability is required then DFSS is invoked.
So, how did the DFSS logical models originate?

The evolution of design quality to concurrent engineering

Throughout much of the first half of the twentieth century, the process of design applied
the craftsman model to identify “new ideas” and through iterative testing prepare them
for the market. The initial structured model for design was typically called
test-analyze-fix (TAF), a 1970s model for the iterative steps of the engineering process.
The engineering management process evolved by the US Department of Defense (DOD)
in the 1980s in an effort to improve management of the acquisition process for weapon
systems which had significant cost over runs (Watson, 1985). It is instructive to note
that the development of the military’s approach to concurrent engineering (CE)
overlapped in timing with the development of DFSS. In Juran’s (1992) seminal work on
design quality, there is no mention of CE, product line management (PLM) or DFSS.
Robert G. Cooper summarized the current state of the art in product development in
his 1980s era “stage-gate” model for the design and implementation of new product
development: ideation (Gate 1 initial screen); Stage 1 preliminary screen (Gate 2
secondary screen); Stage 2 detailed investigation (build the business case) (Gate 3
decision on business case); Stage 3 development (Gate 4 post-development review);
Stage 4 testing and validation (Gate 5 pre-commercialization business analysis); Stage 5
full production and market launch (post-implementation review) (Cooper, 1985).
One of the major works on product development of the early 1990s was the study by
Steven C. Wheelwright and Kim B. Clark of Harvard Graduate School of Business.
The initial state of the new product development process logic is described as a Design for Six
development strategy that relies on cross-functional integration of a design-build-test Sigma: caveat
cycle though a sequence of development prototype testing: initial concept testing,
design verification testing, design maturity testing, production verification testing, and emptor
volume production (Wheelwright and Clark, 1992). At this time a cross-industry study
by consulting firm Pitaglio, Todd, Raburn, and McGrath identified structured product
development as a best practice (naming the steps: concept evolution, planning and 75
specification, development, evaluation, and product release) and also they
recommended cross-project management for effective resource scheduling across a
portfolio of projects that are managed at functional interfaces (McGrath et al., 1992).
Cooper had dictated the state of the art in design engineering until the military decided
to move forward under the leadership of the Defense Advanced Research Projects
Agency in the early 1990s and it introduced the term “CE” in its Defense Initiative on
Concurrent Engineering. It is noteworthy that at about the same time the DOD
adopted TQM as a methodology – causing these two streams of thinking to merge
(Rosenblatt and Watson, 1991; Cousins, 1991a, b; AitShalia et al., 1995; Prasad, 1996, 1997).
Karl T. Ulrich and Steven Eppinger proposed a generic six phase product
development process: planning, concept development, system-level design, detail
design, testing and refinement, and production ramp-up which served to simplify
the CE process and transition it toward PLM (Ulrich and Eppinger, 1995, 2000).
C. Merle Crawford and C. Anthony Di Benedetto promoted a “new products process”
that included strategic planning, concept generation, pre-technical evaluation,
technical development, commercialization, and launch as a logical structure for PLM
(Crawford and Di Benedetto, 1997, 2008) which signaled the final stage in
transformation of the CE process into PLM.
In 2001, Brian Semkiw, CEO of Rand Worldwide, proposed using three-phased PLM
process consisting of the phases of product planning, product development, and product
management to summarize the second generation in development of PLM (following the
TAF first generation). Gregory H. Watson suggested that the integration of DFSS
methodology into PLM would create a third generation of PLM (Watson, 2005).

Evolution of the concept of DFSS

DFSS began its life as an ill-defined method in search of a logical framework. To answer
the question what is DFSS, consider some of the various definitions from literature:
(1) “While Six Sigma helps fix what is broken [. . .] Design for Six Sigma helps to
design things that don’t break in the first place, things that do more and cost
less” (Chowdhury, 2002a, b).
(2) “The ultimate goal of DFSS is to:
.
Do the right things.
.
Do things right all the time” (Yang and El-Haik, 2003).
(3) “The term ‘Six Sigma’ in the context of DFSS can be defined as the level at
which design vulnerabilities are not effective or minimal. Generally two major
design vulnerabilities may affect the quality of a design:
.
Conceptual vulnerabilities that are established because of the violation of
design axioms and principles.
IJLSS .
Operational vulnerabilities due to the lack of robustness in the use
1,1 environment. Elimination or reduction of operational vulnerabilities is the
objective of [. . .] Six Sigma” (Yang and El-Haik, 2003).
(4) “DFSS adds another dimension to product development, called critical
parameter management (CPM). CPM is the disciplined and focused attention to
the design’s functions, parameters, and responses that are critical to fulfilling
76 the customer’s needs [. . .] DFSS is about preventing problems and doing the
right things at the right time during product development. From a management
perspective, it is about designing the right cycle-time for product development
of new products. It helps in the process of inventing, developing, optimizing,
and transferring new technology into product design programs. It also enables
the subsequent conceptual development, design, optimization, and verification
of new products prior to their launch into their respective markets” (Creveling
et al., 2003).
(5) A broader definition was offered in Watson (2005): “DFSS has three major
components: product line management, design and new product development
project management, and the Six Sigma toolkit (define-measure-analyze-
design-verify (DMADV)) that is applied in the product creation process. The
definition of DFSS that we will use is: Design for Six Sigma is a process to define,
design and deliver innovative products that provide competitively attractive value
to customers in a manner that achieves the critical-to-quality characteristics for all
the significant functions.”

Alternative approaches to the process of DFSS

To understand logical structures of alternative DFSS models, a literature search was
performed to identify the proposals that have been made. A summary of this search is
reported below.
Initially, Harry (1997) made no mention of a unique logical model for DFSS.
As Harry’s ideas crystallized by 2000, he referenced DFSS, but incorporates it as part of
the “improve” phase of DMAIC. “The closer a company comes to achieving Six Sigma,
the more demanding the improvements become. At 4.8 sigma companies hit a ‘wall’ that
require redesigning of processes, known as ‘Design for Six Sigma’”. They later stated
that “the improve phase encompasses the process known as DFSS, as well. Using DFSS,
the processes that create the products or services are designed from the beginning or
reconfigured in such a way that they produce six sigma” (Harry and Schroeder, 2000).
Harry also cites GE Medical Systems (GEMS) as the initiator of DFSS as reported in
the 1997 GE Annual Report where GE described that after 1998 every new product will
be the result of DFSS application (Harry and Schroeder, 2000). However, until this time,
DFSS was contained in the DMAIC framework and was invoked following the DMA
(from DMAIC) whenever new capability was required and then the team would shift
to the final steps of design and verify to complete the sequence DMADV. At least
one handbook written in this period as an attempt to develop a structured body of
knowledge for Six Sigma did not mention either DMAIC or DFSS, while in its
2003 edition the DMADV model is recognized (Pyzdek, 2001, 2003).
In 2001, the American Society for Quality (ASQ) introduced a new journal dedicated
to the Six Sigma movement. The first two articles it published on this subject did not
refer to a specific model, but used the general DFSS framework to describe the approach
(Berryman, 2002; Humber and Launsby, 2002). ASQ also publishes handbooks for black Design for Six
belts and green belts to study Six Sigma. It is interesting to note that while the Black Sigma: caveat
Belt Handbook defines DFSS as the DMADV model (Kubiak and Benbow, 2009), the
Green Belt Handbook offers both the DMADV and identify-design-optimize-verify emptor
(IDOV) models (Munro et al., 2003).
It must be noted that until this time most DFSS work was being developed by
consultants who needed to differentiate their methodology from others as they sought 77
an advantage in promoting their own consulting practices. This conclusion becomes
clear when one considers the degree of logical chaos that prevailed in the period
2000-2005 in promotions of DFSS:
.
Tennant (2002) (an independent management consultant) recognized that DFSS
has a DMADV model (based on his past experience in GE) but proposes a
different model, define-customer-concept-design-implement (DCCDI) based on
self-analyzed linguistic clarity that is not very strongly convincing and not
related to practice by any rigorous analysis. Perhaps, it was the litigious nature of
the Six Sigma community during this period of time that lead to this development
as some firms legally challenged the rights of other firms to use the same models
and terms in promoting related practices of management consulting in the
Six Sigma methods (despite the fact that the trademark for Six Sigma was held by
Motorola who had generously put it into the public domain for use by all
organizations).
.
Chowdhury (2002a, b) of American Supplier Institute declares there is no
consistency among practitioners about the terms that define the process.
As a result, the acronyms range from DMADV to define-measure-explore-
develop-implement (DMEDI) to identify-define-develop-optimize-verify (IDDOV).
He states that it “really does not matter what you call it. The DFSS methodology is
still a straightforward five-step process, just as is Six Sigma’s DMAIC.” (p. xvi).
He then uses the process IDDOV to refer to DFSS. Chowdhury cites the GEMS
history from 1998 to 2001, but does not identify their methodology as DMADV
(p. 12). He does not describe the origins of IDDOV.
.
Creveling et al. (2003) ignored DMADV completely. Creveling promotes two other
methods: invent-innovate-develop-optimize-verify for new designs and
sub-systems and concept design, design development, optimization and verify
certification which are derived from the consulting of Steve Zinkgraf of Sigma
Breakthrough Technologies, Inc. Creveling further acknowledges that CDOV is
“largely based on Walter Shewhart’s famous contributions”.
.
Yang and El-Haik (2003) also ignore the DMADV model completely and do not
address it – proposing an alternative that they named identify-
characterize-optimize-verify (ICOV) (identify requirements-characterize the
design-optimize the design-verify the design). There is no reference given to
the origin of the ICOV model, so it is considered a unique logical model that is
outside the frame of this historical development.
.
Brue and Launsby (2003) (Six Sigma Consultants, Inc.) propose a unique logical
model for DFSS – the plan-identify-design-optimize-verify process; however,
they do reference other models but state that since there is no standard that all
models are equal: DMADV, define-measure-analyze-design-optimize-verify
IJLSS (DMADOV), DCCDI, DMEDI, define-measure-analyze-design-improve-control
1,1 (DMADIC), and RCI (however, the origins of these models are not described by
them).
.
Refreshingly, Breyfogle (2003) (principal of the Smarter Solutions consultancy)
cites the DMADV model and does not reference any other methodology.
.
When Evans and Lindsay (2005) initially refer to DFSS they reference DMADV
78 and then cite the CDOV approach proposed by Creveling et al. (2003) and then
they postulate the logical equivalence of the models by stating that the DMA
phases of DMADV are equivalent to the concept development phase in CDOV
and that design and optimization are equal to the design phase while verification
is equal in both models. Based on these choices they elect to describe the CDOV
model in their book.
.
In a book focused on management methods related to DFSS, Watson (2005)
(consultant with Business Excellence Solutions, Ltd) described the DFSS process
as relative to PLM methods and suggested a different logical meaning for the
steps of the DMADV model. He proposed wholly distinct actions in the DMA
phases of the DMADV model compared with DMA phases of the DMAIC model
(Watson, 2003). However, this approach creates confusion as now two logical
models have different actions related to the same named steps. Indeed, clarity
would have been better achieved by renaming the steps in the DMADV model.
.
When ASQ hired consultant Scripps (Scripps & Associates) to develop a DFSS
webinar (delivered April 14, 2005), he identified several competing models (based
on a citation from www.isigma.com): the first DFSS model identified was the
DMADV model which is attributed to use at GE Plastics in 1996; the second
model is identified as IDOV with no source attached; the third model identified
was ICOV which Scripps credits to the consulting company SigmaPro. The fourth
model is the revision of the GE DMADV to add a step and become DMADOV.
The next model cited is Tennant’s model DCCDI. The sixth model cited was
created by Price Waterhouse: DMEDI. Because he identified no prevailing model,
Scripps took the unorthodox approach to create a totally new model which
he proposed to serve as the industry standard: identify-design-evaluate-
affirm-scale-up (Scripps, 2005).
.
Jiang et al. (2007), wrote in Quality Progress Magazine (published by the ASQ) on
the approach for integrating design for excellence with DFSS and in their article
they only referenced DMADV as a DFSS process.
.
When Park and Anthony (2008) describe DFSS, they identify competing models that
exist DMAD(O)V (cited as a GE upgrade to DMADV), IDOV (no origin citied) and
define-initiate-design-execute-sustain was proposed by the Qualtec Consulting
Company, and then propose the use of another modification of DMADV: DMA plus
the extra steps of redesign-implement-control for transactional Six Sigma.

Assessment of the current state of DFSS model logic

This development of models for DFSS logic is reminiscent of the comment by Box et al.
(2005): “all models are wrong, but some models are useful.” The DFSS model requires
stiff academic discipline to set it aright. Consider the following unresolved questions
that have been raised as a result of this review of the ten-year history in the Design for Six
development of DFSS logic: Sigma: caveat
.
Should the entry into DFSS be linked to the DMAIC model or should the logical emptor
models for DMAIC and DFSS be independent?
.
Should there be one model for DFSS (e.g. DMADV or its equivalent or three
(a program management model, project management model and a logical toolkit
for organizing the use of design tools (e.g. a model like DMADV)?
79
.
Does “Identify” recognize all of the front-end issues that must be considered in a
design or should three steps describe the front end (the logical equivalent of
Watson’s DMA)?
.
Does the design phase include optimization or are these two distinct steps?

Clearly from the profusion of models offered, there is no general agreement on these
questions. However, what we can see that two issues should be challenged from an
engineering perspective:
(1) Should the quality community dictate to the engineering community the process
to use for design? Note that the initial steps used in DMAIC do not relate to the
preliminary processes in the CE or PLM models. An assessment of a current
process or product may lead to the conclusion that there is not enough process
capability to achieve its performance targets and conclude that DFSS is
required. But, this is not equivalent to following the first three logical steps of
DMAIC in all applications of DFSS.
(2) Engineers would find it strange to divide design into two phases. Note that in
the DMADV process that the output of DMA is the high-level conceptual
design, which is ready for detailed design. Thus, the “design” that occurs in
the Design phase should refer to detailed engineering design which is an
iterative process that concludes the concepts of both design and optimization.
A design should not be considered to be complete unless it has been optimized.
These are not separate tasks, but closely related tasks which require an iterative
sequence of work to achieve optimization.

Some models collapse the DMA activities of DMADV into a single step (focusing on
either identification or conceptualization of the design outcome). However, the PLM
approach has three distinct steps that typically involve different focus groups within an
organization: define spotlights on creating the high level conceptual design based on
product line dynamics, technology available and business needs. Measure focuses on
the market, customer and competitive research required to sharpen the technical design
into a marketable product. Analyze creates a business case (using risk-benefit analysis)
based on a high-level design and a financial assessment of the opportunity (Watson,
2005). Since each of these phases involves different groups, it is a good practice to
separate these individual model steps. Continuing to examine the last two steps of the
DMADV model we see again that different parts of the organization are involved –
design is an engineering task that will involve the iterative technical design,
development testing, and redesign to the point of optimization for the intended purpose.
Verification is an external focus on testing the design in its intended market and
“polishing the design” in preparation for market launch. This last phase should involve
IJLSS both dedicated user tests (so-called a and b-tests) as well as studying market price
1,1 tolerance to validate the commercial expectations.
Perhaps the largest question of all from this proliferation of models is the confusion
that it causes to potential users. There are three unique applications of DFSS: product
(although it could be argued that hardware and software are unique applications),
service, and business process improvement. Does each of these applications need a
80 unique model or could they all follow the same logical model? It would be more elegant
to apply a single logical model that represents all applications of the DFSS toolkit, much
like DMAIC can be applied for both engineering and business applications. DMAIC is
used for both product and service improvements with the key distinctions at the level
below the logical model based on the types of data and processes that are being
analyzed.
One additional expectation for a new DFSS model would be that its acceptance
should not generate an undue influence for any consulting venture. Since most of the
models proposed come from consulting firms it is not proper to choose one of the
current models over the others and thereby endorse a single approach. So, how to
proceed? Let’s turn to the earlier lessons from the development of PDCA in Japan to see
what this history suggests. It would be conceptually tidy if the approach to Six Sigma
were bundled as tightly as the logic used by the Japanese quality movement to
simultaneously separate and link PDCA and SDCA. Would this be a possibility, and if
so, then who should be asked to make this contribution?

Concluding comments
Let us begin with an observation from this intellectual history of DFSS that the authors
believe is still most appropriate:
The strength of DFSS is not in its stand-alone performance, but it comes from integration of
methods and concepts that have been independently developed into a system of thinking and
working that result in product designs that serve customers better while generating
attractive profit. DFSS is not strategic planning but it builds upon the “recognize” component
in strategic planning where business improvement needs are unveiled. DFSS is not program
management, but it provides a philosophy and methodology for more effectively coordinating
multiple project programs. DFSS is not project management, but it supports project
managers with an analytical process that facilitates a “right-the-first-time” approach to
product creation. Finally, DFSS is much more than the DMADV process, but DMADV is one
way to summarize how DFSS fits into an overall business system. While DFSS is somewhat
amorphous in this format, it becomes defined within the context of an organization’s specific
business system. Thus, there is no one-size-fits-all definition of DFSS, rather DFSS concepts
and methods must be adapted and customized to support a particular business purpose,
cultural style and design technology in order for it to have real substance. DFSS supports
management of design programs in the product creation process (Watson, 2005).
Certainly, this observation is open to challenge. However, the point is clear that
distinctions in the way that DFSS is applied will vary by organizational business model
and the application of a specific logical model needs to comprehend the distinctions in
product design requirements for various types of businesses. This will be challenging.
However, perhaps there is a way to move forward, if the global quality community can
learn a lesson from way that the JUSE Research Committees developed the Japanese
approach to TQM and use a similar approach to developing the standard logical models Design for Six
for Six Sigma. Sigma: caveat
In May, ASQ (2009) announced the formulation of a quality body of knowledge
(QBOK) which would organize the intellectual property of quality that has been emptor
generated over the past 60 years and provide stewardship to assure that quality
information contained in the QBOK is important, accessible, dependable, accurate and
authentic (ASQ, 2009). One of the objectives of the QBOK is to evaluate gaps in the 81
recognized quality knowledge base and to work with teams of experts to close such
gaps. It should be clear from the observations in this paper that the western quality
movement is facing an intellectual challenge regarding the logical approach to DFSS.
Observing the vacillation in understanding the proposed logical models for DFSS, we
conclude that tasking the QBOK team to develop a standard approach to this
methodology is a good application of this team’s energy. In establishing this task, ASQ
should engage the engineering community to help improve the quality and
acceptability of the final design.
Until such an approach matures and reaches consensus, then the best advice that
can be given to companies that are interested in implementing DFSS is: caveat emptor.

References
Ackoff, R. and Emery, F. (1972), On Purposeful Systems: An Interdisciplinary Analysis
of Individual and Social Behavior as a System of Purposeful Events, Aldine-Atheron,
Chicago, IL.
AitShalia, R., Johnson, E. and Will, P. (1995), “Is concurrent engineering always a sensible
proposition?”, IEEE Transactions on Engineering Management, Vol. 42 No. 2, pp. 166-70.
Akao, Y. (1991), Hoshin Kanri, Productivity Press, Cambridge, MA.
ASQ (2009), Guide to the Quality Body of Knowledge (QBOK), American Society for Quality,
Milwaukee, WI.
Bacon, F. (1620), Novum Organum, Oxford University Press, Oxford (translated by J. Spedding
(1855)).
Berryman, M. (2002), “DFSS and big payoffs”, Six Sigma Forum Magazine, Vol. 2 No. 1, pp. 23-8.
Box, G., Hunter, S. and Hunter, W. (2005), Statistics for Experimenters, 2nd ed., Wiley,
New York, NY.
Breyfogle, F. (2003), Implementing Six Sigma, 2nd ed., Wiley, New York, NY.
Brue, G. and Launsby, R. (2003), Design for Six Sigma, McGraw-Hill, New York, NY.
Chowdhury, S. (2002a), Design for Six Sigma, Dearborn Trade, Dearborn, MI.
Chowdhury, S. (2002b), The Power of Design for Six Sigma, Dearborn Trade, Dearborn, MI.
Cooper, R. (1985), Wining at New Products, Perseus Books, Cambridge, MA.
Cousins, R.E. (1991a), “Rules for concurrent engineering – Part I”, Computer (IEEE), Vol. 24
No. 11.
Cousins, R.E. (1991b), “Rules for concurrent engineering – Part II”, Computer (IEEE), Vol. 24
No. 12.
Crawford, C. and Di Benedetto, A. (1997, 2008), New Products Management, 9th ed., Irwin
McGraw-Hill, New York, NY.
Creveling, C., Slutsky, J. and Antis, D. (2003), Design for Six Sigma in Technology and Product
Development, Prentice-Hall, Upper Saddle River, NJ.
IJLSS Deming, W. (1986), Out of the Crisis, MIT Press, Cambridge, MA.
1,1 Deming, W. (1993), The New Economics, MIT Press, Cambridge, MA.
Evans, J. and Lindsay, W. (2005), An Introduction to Six Sigma and Process Improvement,
Southwestern, Mason, OH.
Forrester, J. (1961), Industrial Dynamics, Pegasus Communications, Waltham, MA.
Forrester, J. (1968), The Principles of Systems, 2nd ed., Productivity Press, Portland, OR.
82
George, M. (2001), Lean Six Sigma, McGraw-Hill, New York, NY.
Harry, M. (1985), Practical Experimental Design, Research Dynamics, Tempe, AZ.
Harry, M. (1993), The Vision of Six Sigma, 3rd ed., Six Sigma Academy, Scottsdale, AZ.
Harry, M. (1997), The Vision of Six Sigma, 5th ed., Six Sigma Academy, Scottsdale, AZ.
Harry, M. (2000a), “Abatement of business risk is key to Six Sigma”, Quality Progress, July.
Harry, M. (2000b), “Framework for business leadership”, Quality Progress, April.
Harry, M. and Schroeder, R. (2000), Six Sigma: The Breakthrough Management Strategy
Revolutionizing the World’s Top Corporations, Doubleday, New York, NY.
Hosotani, K. (1989), The QC Problem-solving Approach: Solving Workplace Problems the Japanese
Way, 3A Publications, Tokyo.
Humber, C. and Launsby, R. (2002), “Straight talk on DFSS”, Six Sigma Forum Magazine, Vol. 1
No. 4.
Ikezawa, T. (1970), How to Make TQC More than Just a Slogan, PHP Press, Tokyo.
Ishikawa, K. (1985), What is Total Quality Control? The Japanese Way, Prentice-Hall, Englewood
Cliffs, NJ.
Jiang, J., Shiu, M. and Tu, M. (2007), “DFX and DFSS: how QFD integrates them”, Quality
Progress, American Society for Quality, Milwaukee, WI, October.
Juran, J. (1968), Managerial Breakthrough, McGraw-Hill, New York, NY.
Juran, J. (1973), “The Taylor system and quality control”, Quality Progress, Vol. 6, May.
Juran, J. (1974), Quality Control Handbook, 3rd ed., McGraw-Hill, New York, NY.
Juran, J. (1992), Juran on Quality by Design, The Free Press, New York, NY.
Kano, N. (1993), Task Achieving QC Story for QC Circles, JUSE Press, Tokyo.
Kano, N. (2005), “Causal relationship model and a comprehensive procedure for quality
management”, Proceedings of the 3rd Annual Asian Network for Quality, Taipei.
Kano, N., Ando, Y. and Eiga, T. (1997), Task Achieving QC Story for Management, JUSE Press,
Tokyo.
Kolesar, P. (2008), “Juran’s lectures to Japanese executives in 1954: a perspective and some
contemporary lessons”, Quality Management Journal, Vol. 15 No. 3.
Kubiak, T. and Benbow, D. (2009), The Certified Six Sigma Black Belt Handbook, 2nd ed.,
American Society for Quality, Milwaukee, WI.
Kume, H. (1985), Statistical Methods for Quality Improvement, 3A Corporation, Tokyo.
McGrath, M., Anthony, M. and Shapiro, A. (1992), Product Development: Success through Product
and Cycle-time Excellence, Butterworth-Heinemann, Stoneham, MA.
Maguire, M. (1999), “Cowboy quality”, Quality Progress, Vol. 32 No. 10, pp. 27-34.
Morita, A. and Ishihara, S. (1991), The Japan that Can Say No, Simon and Schuster,
New York, NY.
Motorola Corporation (1990), Design for Manufacturability, Motorola University, Schaumburg, IL.
Munro, R., Maio, M., Nawaz, M., Goindarajan, R. and Zrymiak, D. (2003), The Certified Six Sigma Design for Six
Green Belt Handbook, American Society for Quality, Milwaukee, WI.
Park, S. and Anthony, J. (2008), Robust Design for Quality Engineering and Six Sigma, World
Sigma: caveat
Scientific, New York, NY. emptor
Perez-Wilson, M. (1989), Machine/Process Capability Study: A Five-stage Methodology for
Optimizing Manufacturing Processes, Advanced Systems Consultants, Phoenix, AZ.
Prasad, B. (1996), Concurrent Engineering Fundamentals: Integrated Product and Process 83
Organization, Vol. 1, Prentice-Hall, Upper Saddle River, NJ.
Prasad, B. (1997), Concurrent Engineering Fundamentals: Integrated Product Development, Vol. 2,
Prentice-Hall, Upper Saddle River, NJ.
Pyzdek, T. (2001, 2003), The Six Sigma Handbook, McGraw-Hill, New York, NY.
Rosenblatt, A. and Watson, G.F. (1991), “Special report: concurrent engineering”, IEEE Spectrum,
Vol. 28 No. 7, pp. 22-39.
Scripps, T. (2005), “Increase your profitability: how to use DFSS to optimize your product
development process”, American Society for Quality Webinar, April 14, 2008.
Senge, P. (1990), The Fifth Discipline: The Art and Practice of the Learning Organization,
Doubleday, New York, NY.
Shewhart, W. (1939), Statistical Method from the Viewpoint of Quality Control, The Graduate
School, US Department of Agriculture, Washington, DC.
Smith, A. (1776, 2000), An Inquiry into the Nature and Causes of the Wealth of Nations, Princeton
Review, Princeton, NJ.
Taylor, F. (1911, 1990), The Principles of Scientific Management, Dover Publications, Mineola,
NY.
Tennant, G. (2002), Design for Six Sigma, Gower House, Aldershot.
Ulrich, K. and Eppinger, S. (1995, 2000), Product Design and Development, 2nd ed., McGraw-Hill,
New York, NY.
von Bertalanffy, L. (1968), General System Theory: Foundations, Development, Applications,
George Braziller, New York, NY.
Watson, G. (1985), Marketing R&D Concepts to the Navy, Continental Press, Washington, DC.
Watson, G. (1994), Business Systems Engineering, Wiley, New York, NY.
Watson, G. (2003), Six Sigma for Business Leaders, GOAL/QPC, Salem, NH.
Watson, G. (2005), Design for Six Sigma, GOAL/QPC, Salem, NH.
Wheelwright, S. and Clark, K. (1992), Revolutionizing Product Development: Quantum Leaps in
Speed, Efficiency, and Quality, The Free Press, New York, NY.
Yang, K. and El-Haik, B. (2003), Design for Six Sigma: A Roadmap for Product Development,
McGraw-Hill, New York, NY.

Further reading
Ginn, D. and Streibel, B. (2004), The Design for Six Sigma Memory Jogger, GOAL/QPC,
Salem, NH.

About the authors

Gregory H. Watson is a doctoral student in Industrial Engineering at Oklahoma State University
and an adjunct faculty member in Engineering and Technology management. He is the President
of the International Academy for Quality, past-President and Fellow of the ASQ, and senior
IJLSS member of the Institute of Industrial Engineers. His books include: Strategic Benchmarking
(1993), Six Sigma for Business Leaders (2004), Design for Six Sigma (2005), and Strategic
1,1 Benchmarking Reloaded with Six Sigma (2007). In 2009, Mr Watson was the first non-Japanese to
be awarded a Deming Medal by the Union of JUSE and has been elected to membership in the
International Statistical Institute.
Camille F. DeYong is an Associate Professor of Industrial Engineering and Management at
Oklahoma State University. Her research interests are economic analysis, quality management,
84 customer service, and women in engineering. She teaches courses in service quality, quality
management, benchmarking, and Six Sigma. Camille F. DeYong consults in industrial
engineering applications related to performance measurement, strategic planning, and life cycle
costing. She has served as a Judge of the Oklahoma State Quality Foundation and as a senior
examiner for the Malcolm Baldrige National Quality Award. Camille F. DeYong holds
membership in the Institute of Industrial Engineers, ASQ, American Society for Engineering
Education, and Sigma Xi.

To purchase reprints of this article please e-mail: [email protected]

Or visit our web site for further details: www.emeraldinsight.com/reprints