вход по аккаунту



код для вставкиСкачать
Appl. Stochastic Models Bus. Ind. 2009; 25:29–32
Published online 16 January 2009 in Wiley InterScience ( DOI: 10.1002/asmb.757
Discussion: Fisher–Nair paper
It is an honor and a pleasure to add our comments to this interesting, and sometimes provocative,
paper by two eminent leaders in advancing the use of statistics for quality improvement. It is also
good to observe the globalization reflected by this collaboration of two authors whose home bases
are at different ends of the earth.
Fisher and Nair cover a great deal of territory and present many ideas. First, they trace the
historical development of quality management. We are pleased to see their acknowledgment of
the contributions of Myron Tribus and Homer Sarasohn, with whom today’s quality practitioners
might not be well acquainted. Then they provide us with their own thoughts on where things stand
today and how we should proceed.
We will comment on only a few of the many points that the authors bring out and, based upon
our own experience, raise or reiterate some tough, but important, issues. See Hahn [1] and Hahn
and Doganaksoy [2] for further discussion.
We would have predicted—at the start of the so-called ‘quality revolution’ (when Deming et al.
became prominent on the national scene) over a quarter of a century ago, and again upon the
popularization of Six Sigma in the mid-90s—that by 2008 we would be further ahead in achieving
high quality in our products and services than, in fact, we really are. We can all sadly point to
numerous examples from the news, such as minivan and computer battery recalls, of defective or
hazardous products that have been released in the market apparently without proper concern for
building quality into the product up front.
We especially concur with the authors’ contention that ‘the service sector tends to lag well
behind’. Here’s an example. Recently, one of us found that his cell phone had been disconnected.
Upon contacting my service provider, I learned that the reason was my alleged failure to make my
monthly payment. This seemed strange as I had authorized automatic monthly deductions from
my credit card three years ago, and these had been routinely applied ever since. On checking
further, I learned that my card’s expiration date had passed and that the provider had not taken
the trouble to verify its renewal or check with me. Upon my assurance that I had, indeed, updated
the card’s expiration date, I was told that my service would be restored within four to six hours.
One week and numerous phone calls later, this still had not happened. I was eventually referred
to a less-than-polite supervisor who informed me that all of this was my ‘fault’ (because I had
neglected to notify the company of my card’s renewal). The restoration of service, moreover,
was beyond his control and was being handled by ‘headquarters’, to whom I could write to if I
wished. (I finally got service restored upon approving an immediate $15 authorization). A week
later, I had a similar experience in trying to renew a matured bank account at an advertised
improved interest rate. Readers can, undoubtedly, add their own experiences—especially if they
have been flying commercially recently. The vast area of healthcare provides innumerable further
Copyright q
2009 John Wiley & Sons, Ltd.
Our personal frustrations might seem trivial, but unfortunately, they reflect the overall picture.
The imperative of high quality seems to have been drowned out by other imperatives. It is a clear
challenge to reinvigorate and revive the quality revolution and to adapt it to the 21st century. The
current period of economic slowdown—with business less hectic and customers more selective—
might provide a special opportunity for doing this and can have big payoffs when the economy turns
A key challenge to us all remains in developing, as Fisher and Nair suggest, holistic and lasting
systems that result in consistently high product and service quality. And perhaps—extending the
authors’ proposal to have us develop systems for ‘measurement-based quality management’ for
individual activities and companies—statisticians can contribute by helping develop national, or
even international, quality improvement metrics. These would quantify how we, as a society, are
progressing in delivering high quality and reliability products and services to our customers. But
we have no illusions about the difficulties of securing the needed data for such a system and
processing such data to develop useful and consistent metrics.
Most importantly, the resulting metrics—at whatever level—need also identify key areas for
improvement. One of us recently undertook a project whose stated goal was to develop metrics to
quantify and monitor illegal sharing of pirated movie files on the Internet. This rapidly expanded
into an effort to evaluate the impact of alternative strategies for reducing piracy.
We fully agree with the authors’ contention that, we, as statisticians, have much to offer to
companies, and even to society, in ensuring high product and service quality. All of this requires
a solid framework, as Fisher and Nair suggest, and our continued efforts to help others appreciate
the role of statistics. We cannot expect our customers to learn our language—instead we need to
become fluent in theirs. Also, we must help our colleagues and management gain an appreciation
of the basic concepts of statistical thinking and to advance proactive quality improvement.
To be effective, this often requires us to have a seat at the table. It is, therefore, imperative for
us to convince customers that our biggest contributions are typically made in problem definition
and in developing an appropriate system for data gathering. We need to roll up our sleeves and be
embedded, from the beginning, within the team addressing strategic endeavors. A consequence of
this is that we share in the credit for project success—and in the blame for possible failure. This is
in contrast to our more traditional role as consultants who spray gems of wisdom here and there,
often after the fact.
The continued interest by many companies in (lean) Six Sigma, and its emphasis on ‘In God we
trust, all others bring data’ has provided us with a foot in the door. This movement, and other
developments—sometimes referred to as ‘the democratization of statistics’—have led practitioners
Copyright q
2009 John Wiley & Sons, Ltd.
Appl. Stochastic Models Bus. Ind. 2009; 25:29–32
DOI: 10.1002/asmb
to use formal statistics more than ever before, starting with the design of the experiment. The
emphasis on Six Sigma in many companies has, moreover, resulted in many of today’s emerging
business leaders having been Black Belts or Master Black Belts. A consequence is often a common
goal (quality improvement), a common roadmap (e.g. DMAIC, DFSS) and a common language
(e.g. CTQ’s, gage variability), in addition to an appreciation of the importance of data. We applaud
these developments.
But, we regret that frequently the emphasis in Six Sigma training has been on the use of tools,
and insufficiently on the underlying statistical thinking. Thus, we need to help the leadership team
understand, value and support the effective use of statistical concepts throughout the organization.
To maintain the momentum and elevate statistical thinking to increasingly higher levels in the
business, we need to strive for the publication of convincing success stories in journals, such as
the Harvard Business Review, that are read by current and future executives.
The authors make a useful distinction between the broad management view provided by total
quality management (TQM) and the narrower application of specific tools.
At the same time, TQM does not seem to have a universally accepted definition and tends to
mean different things to different people. It appears to be closely associated in the minds of many
with global quality standards, such as ISO 9000; but unlike Six Sigma, TQM does not provide a
specific roadmap for moving forward.
A key barrier to up-front quality improvement is the fact that, while quality deficiencies often
take time to assert themselves, management is rewarded and promoted based upon short-term
accomplishments, such as the most recent quarters’ bottom line results. Thus, today’s apparently
high-performing managers are likely to be rapidly promoted in recognition of their ‘successes.’
By the time the quality/reliability problems that they neglected become evident, such managers
may well have moved on. Fixing the problem is left to their successors. This system surely does
not encourage up-front concern for high quality.
What further complicates matters is the difficulty in measuring the impact of quality, and
especially reliability, improvement. How can we quantify the effect of preventing a failure mode
that, if it had occurred, would have resulted in an expensive recall and loss of customer confidence?
Deming [3, p. 121] asserted that ‘the most important figures that one needs for management are
unknown or unknowable.’ We are not that pessimistic, but do feel that quantifying the impact of
building quality into a product or system up-front is one of our most important and most difficult
challenges. It calls for establishing metrics along the lines proposed by the authors. And, as they
astutely point out, measurement, indeed, drives behavior.
Fisher and Nair comment that ‘it is unlikely that statisticians will be able to make significant
progress on the higher-level performance measurement problems until they themselves start to fill
Copyright q
2009 John Wiley & Sons, Ltd.
Appl. Stochastic Models Bus. Ind. 2009; 25:29–32
DOI: 10.1002/asmb
leading positions in companies, to develop a proper appreciation of what is needed.’ We hope that
this is not the case.
We fully recognize the advantage of having clout. Myron Tribus’s positions as dean of Dartmouth
College’s Thayer School of Engineering, Assistant U.S. Secretary of Commerce for Science and
Technology (in the Johnson Administration) and Senior V.P. for Research and Engineering for the
Xerox Corporation helped him exert much influence. At the same time, we feel that statisticians
can be heavily involved as visionaries and effective communicators—and have important impact—
without necessarily taking on leading positions in a company. Society, moreover, cannot afford to
wait for this to happen.
Deming [3, p. 466] advocated the position of ‘statistical leader,’ urging that there ‘be a leader
of statistical methodology, responsible to top management. . . He will. . . be a regular participant
in any meeting of the president and staff’ (and ‘command a high salary’). However, attractive we
might find this concept, it seems fair to say that it has not materialized.
This, however, should in no way dilute the essential message that we need to act proactively
and exert leadership in what we think and advocate. Deming, indeed, was right on the mark when
he stated
• ‘Every appalling example in this book turned up because I was there on the line, on the job,
trying to be helpful by looking for sources of improvement and strong practices. If I had waited
for them to come for help, I’d still be waiting.’ [3, p. 469].
• ‘It is necessary to innovate, to predict the needs of the consumer, give him more. . . The secret
for reduction in time of development is to put more effort into the early stages.’ [4, p. 10 and
Of course, successful proactive endeavors require a very special type of person; Deming, indeed,
was such a person.
We congratulate Fisher and Nair on a thought-provoking paper. The fact that we have relatively
few comments about the specific points they raise—and, instead, have tried to reflect on the bigger
picture—is simply because we have no reason to disagree with most of what they say.
1. Hahn GJ. The business and industrial statistician: past, present and future. Quality and Reliability Engineering
International 2007; 23:643–650.
2. Hahn GJ, Doganaksoy N. The Role of Statistics in Business and Industry. Wiley: Hoboken, NJ, 2008.
3. Deming WE. Out of the Crisis. Center for Advanced Technology. MIT: Cambridge, MA, 1982.
4. Deming WE. The New Economics for Industry, Government and Education (2nd edn). MIT Press: Cambridge,
MA, 2000.
Retired Manager, Applied Statistics
Principal Technologist-Statistician
GE Global Research Center
Copyright q
2009 John Wiley & Sons, Ltd.
Appl. Stochastic Models Bus. Ind. 2009; 25:29–32
DOI: 10.1002/asmb
Без категории
Размер файла
45 Кб
asmb, 757
Пожаловаться на содержимое документа