Page 46 - How To Implement Lean Manufacturing

Lean Manufacturing and the Toyota Pr oduction System 27

concerns with them, we could determine how to supply quality and we could provide it
to a very high level. We used a long list of tools to achieve these communications skills, but
the most important two were the simple customer quality questionnaire—which we sta-
tistically analyzed, of course—and we also became very proficient at Quality Function
Deployment (QFD). Long ago, we ceased using inspection, especially visual inspections
by humans, as a means of achieving quality. Instead, we moved to process control as a
means to make the process more robust. To do this, we first became proficient in data
gathering and analysis using such techniques as those that Ishikawa outlines in his writ-
ings. Now the vast majority of our data is used for process improvement rather than prod-
uct evaluation. In addition, we became very proficient in a wide range of statistical
techniques so we could analyze and make better decisions with our data. The four funda-
mental statistical techniques of Measurement System Analysis (MSA), Statistical Process
Control (SPC), Designs of Experiments (DOE), plus Correlation and Regression are widely
understood at even the supervisory level in our plants. We also made all levels of person-
nel responsible for root cause problem resolution, which means we trained them in vari-
ous levels of problem solving—the “5 Whys” being the cornerstone technique. Another
significant effort was the transition in our quality, process, and product data. Initially, the
majority of our data was attribute data on the product. We moved from a high percentage
of attribute quality characteristics of the product to variables characteristics of the process.
We recognized early on that high levels of quality could not be achieved if we used attri-
bute data, so this meant we needed to correlate the attribute defects to process parameters,
and so we became skilled at this very early in our quality efforts. We had been very com-
mitted to providing quality products and had been working very hard on quality. With the
help of Deming and a unified effort pushed by JUSE, we could supply excellent quality
and our costs and losses associated with quality were very low. We had what most West-
erners would call a very mature manufacturing system that could consistently produce
high-quality goods and deliver them on time for a reasonable price. Quality was no longer
a production problem. Now we needed to look at the losses caused by producing the
wrong quantities—especially the wrong quantities produced and delivered to the wrong
places at the wrong times.”

If I were Ohno, that is what I would have written, because that is where they were
as a manufacturing company. So what Ohno built, the TPS, had a foundation of quality,
but his TPS is not a quality system. Yes, it had jidoka, but we will learn that jidoka is
there to support JIT. In addition, and as an example, Ohno makes nearly no mention of
Cp and Cpk which are the two accepted measures of process capability or “process
goodness.” Nearly every book you read on manufacturing and process quality reduces
the concept of quality to Cp and Cpk for all measure-
ment data, yet Ohno hardly mentions it. One has to
ask why? Well, it is because of what he said—they Point of Clarity The TPS
simply could supply high-level quality, and quality was built on a foundation of
improvements were not what they needed to focus quality and the focus is to
on to reach a higher level of manufacturing excel- control quantities.
lence. Their focus, as he says, was on quantity. Make —(see House of Lean
sure you spell that: q-u-a-n-t-i-t-y. in Chap. 20).
The implication of this information, to some
company that wishes to embark on the journey into Lean, is usually quite sobering.
Ohno says they just spent seven years—seven very focused years—learning how to
deliver quality, after which they embarked on the journey of quantity control. I can say