Problem Analysis- Original

Problem: The Rejected Circuits

The simple question "Why?" is a poor substitute for the fourdimensional
questioning we use in Problem Analysis. Yet, whenever something goes wrong,
it is second nature to ask "Why?" and then review the flood of answers in hope
that one of them will instantly suggest the problem's actual cause. The usual
rationalization for the "Why?" approach is that people have been hired for
their expertise and experience. If they can't come up with answers for
problems that occur in the operation, those people don't belong in their jobs.
Concrete results arising from the combination of systematic techniques and
technical expertise are the only things that will convince a manager that
questions are as important as answers.
An electronics manufacturing company is involved in the demanding task of
producing miniaturized printed circuitry. One day, the production quality fell
off sharply and the number of rejected circuits skyrocketed. "Why?" demanded
the boss. "Why?" echoed his subordinates. "Temperature in the leaching bath
is too high," said one technician. So, temperatures were lowered. A week later,
when rejects climbed still higher, temperatures were raised, then lowered
again, then systematically varied up and down for days. Rejects remained
astronomical. "Cleanliness is not what it should be. That's what's causing the
trouble," someone offered. So, everything was scrubbed, polished, filtered,
and wiped. The rejects dropped, then rose again. Acid concentration was the
next idea. Same results. Water purity was checked out on Wednesday,
Thursday, and Friday. The possibility of oil transferred from the operator's
fingertips received full scrutiny on the following Monday and Tuesday. Rejects
still were high. They might have remained high had not one supervisor begun
to ask systematic questions. "What is wrong with the rejected pieces?" This
produced the information that the acid leaching step of the printed circuit
pattern was occurring unevenly-as if some waterborne contaminant in the
leaching solution was inhibiting the action. "When does it occur?" A check of
the records showed that rejects were at their highest on Monday mornings,
lower on Monday afternoons, and gone by noon on Tuesday. This cast a
different light on everything. Now nobody was asking "Why?" about the cause
of a general, ill-defined deviation. Instead, they focused on what was
distinctive about Monday rnornings compared with the rest of the week. They
focused on what might have been changed that bore a relationship to this
timing. An immediate distinction was recognized: "Monday morning is the first
work period following the non-work period of the weekend." And what
changed on Monday morning? On each Monday, as soon as the tap was
turned, water that had stood in the lines over the weekend came into the
printed circuit leaching laboratories. The water used in the process had to go
through intensive purification, since purity standards of a few parts per billion
are required. A quick search turned up the fact that some valves had been
changed several months before. These valves used a silicone packing material.
As water stood in the lines over the weekend, enough of this silicone packing
material had begun to diffuse into the water and degrade the leaching process.
The result? Many rejections on Monday morning, fewer in the afternoon, and
none after Tuesday noon. By then the contaminated water had been purged
from the system.
Use of Problem Analysis Techniques by the Management Team
Problem Analysis: Need for Internal Communication
Case Study 3
Problem: The Model A Sewing Machine A very large merchandising house sells
a private brand of sewing machines. The machines are manufactured for the
house in six different models. For over a year the sewing machine line has not
been selling as well as expected even though it was well received when it was
first introduced. The Sales Promotion Manager, the Merchandising Manager,
and the Sewing Machine Buyer finally met to analyze the problem. They began
to specify the deviation, each contributing his or her special knowledge. Their
deviation statement was simple enough: "Sewing Machines Selling Poorly."
They could not be more specific at the outset of the analysis, since the exact
nature of the deviation percentages of expected revenue achieved-was being
explored in detail for the first time. "What sewing machines?" was the first
question, in the dimension of the problem's identity. They were prepared to
say, "All six models." But when they began to review their data, they were
startled to see that only Model A was selling poorly. Models B and C were
selling somewhat better than A, and Models D, E, and F were selling quite well.
None of the three managers had realized there were such marked differences
in the rate of sales among the six models. Until then they had only been
dealing with average figures. Almost immediately after beginning the
specification, they went back to the deviation statement and revised it to read
"Model A Sewing Machine Selling Poorly, Models B and C Selling Fairly
Well. . . ." It did not look like a textbook deviation statement but it was more
precise and useful than the first one. They probed into the dimension of
location and found that low sales of Model A were nationwide. There were no
distinctions to be gleaned in terms of geographical factors such as "better in
the Midwest, worst in the Northeast." Coast to coast, Model A sales were poor
wherever Model A was available. They hit paydirt when they turned their
attention to the dimension of Timing. The records showed that the Model A
machine had sold as well as expected in the first couple of months but then
had fallen off sharply. Models B and C also had sold well and then had fallen
off-but not as sharply as Model A. Models D, E, and F had sold well when first
introduced and had continued to do so right up to the present time. The
discrepancies between the IS and IS NOT information were inescapable. They
couldn't be ignored and they wouldn't go away. There was something
distinctive about the Model A machine, and about the B and C models to a
lesser degree, when compared with the models that were selling well. The
Merchandising Manager suggested an explanation: "Model A has fewer
attachments and selling benefits. Perhaps it's simply a less desirable unit!" It
was a neat explanation but unfortunately not so. "The fact is," said the Sewing
Machine Buyer, "there isn't much difference among the attachments of the six
units." She thought it made more sense to look at differences in sales features.
"That won't wash," said the Sales Promotion Manager. "Sales features for the
six machines have been constant from the start. Anyway, how could that
possibly account for a falloff on Model A beginning a couple of months after its
introduction?" The Merchandising Manager then brought up the subject of
motivation. "Could there be some lack of motivation to sell Model A?" The
Sales Promotion Manager replied, "The salesperson makes the lowest
commission on Model A, the next lowest on Models B and C, and the best
commission on the other three. That's normal enough-and besides, it doesn't
represent any kind of change. It's always been that way." After a few moments
of silence, while everyone digested this information, the Sewing Machine
Buyer asked, "Wouldn't there be some lapse of time after a new item is
introduced before the salespeople begin to receive their commissions and
figure which sales bring in the greatest return? I mean, maybe the realization
that the commission rate on Model A is substantially lower than on Models D,
E, and F would take a little time to sink in. After a couple of months their
motivation to sell Model A might have changed." The team decided that this
explanation was the most likely cause of Model A's poor sales. To verify it the
Sales Promotion Manager raised the commission rate on Model A, leaving
price and all other variables the same. Within two months the Model A
machine began to sell better.