Avoiding Paralysis by Analysis
A. Blanton Godfrey
A few weeks ago, a friend called
with a question. Part of a critical process within his company
involved bonding fibers to a backing sheet. Something was
wrong with the adhesion, and the fibers were pulling loose
in the middle of the sheets. At least eight critical variables
had been identified as sole causes.
My friend, a well-trained Six Sigma Black Belt, would
have no trouble solving this problem if he and the company
decided to make it a Black Belt project. But his question
was far more direct: Was it OK to use one of the Six Sigma
tools without going through the entire five-step process?
My answer, of course, was, “Of course!” He
was faced with a classic design of experiments problem.
This simplest solution would probably identify the critical
variables and the right levels at which to control them
to eliminate the problem.
Later, I started thinking about how we’re training
people in Six Sigma and other quality initiatives. Perhaps
we’re overreacting to the days when “tool merchants”
would oversell a particular method as the solution to all
problems. Now, most quality training courses--especially
in sessions such as Six Sigma Black Belt training--emphasize
a thorough process for quality improvement. We carefully
define, measure and test before we plan our experiments
and establish our controls. It’s easy to construct
examples in which skipping a step leads us down the wrong
path. But by helping people do everything exactly right,
perhaps we’re also helping to create a new version
of “paralysis by analysis.”
Often, when I review Black Belt projects, I hear apologies
because all the tools haven’t been used. Not long
ago I reviewed a project in a contract research organization
that designs and manages clinical trials for drug companies.
One of the most difficult steps in an extremely complex
process involved designing the study’s data-collection
plan. If this were not done correctly, the entire study
could be useless. It might generate enough data to provide
conclusive results, but these might not be accepted by the
Federal Drug Administration, and, consequently, the product
wouldn’t be approved. In that case, the study would
have to be extended to include groups not adequately represented
in the sample.
Due to the importance of this step, designing the data-collection
process took many months and numerous reviews and discussions.
Until the design was done, almost nothing else could proceed.
This was the company’s major bottleneck in increasing
its capacity and growing its business. During the project
review, the Black Belt apologized for the limited number
of tools and methods used. Basically, the team had discovered
breakthroughs by using a series of bar charts and other
simple graphical methods. They clearly saw the changes they
could make to both improve designs and radically reduce
cycle times.
In another review, a Black Belt hadn’t gone much
beyond the first stages of defining the process. She’d
done a thorough process flow diagram--actually a value-stream
analysis--and discovered just how fouled up the process
was. Her company had expanded rapidly during the past 10
years and added new models, steps and tests in the manufacturing
process. With each addition, they also added workstations
wherever space was available. After a while, the process
made no sense at all. Stepping back and looking at it as
a complete process, she found many areas for quick improvements.
Some were simple changes to prevent damage that occurred
along the line but was discovered late in the process and
repaired at significant cost. Other changes involved providing
proper lighting at workstations so employees could more
easily detect problems as they occurred and eliminate the
causes.
In case after case, I’ve found examples in which
the thorough application of simple methods solved a number
of important problems in a company. We should ensure that
people feel comfortable using the right tool in the right
place. For years, many of us in the quality world have been
quite critical of those who, knowing only a few tools, try
to apply the same ones to every problem. But perhaps we
should stop and think: If a company is full of nails that
stick up, maybe it should just be using hammers.
A. Blanton Godfrey, Ph.D., is dean and Joseph D. Moore
Distinguished University Professor at North Carolina State
University’s College of Textiles. Prior to his current
assignment, he was chairman CEO of Juran Institute Inc.
During the past 15 years, Godfrey has worked with
companies in more than 60 countries. Letters to the editor
about this column can be e-mailed to letters@qualitydigest.com.
|
