Consider, for example, the washer thicknesses plotted in Figure 
.
The plot provides evidence that the washer thicknesses were produced by a stable process.
If management were happy with the output of the process, perhaps we would look no
further into it for the present. But management was displeased that their competitors
were producing washers that had thicknesses nearer on average to the specified thickness
of 2.16 mm. To make matters worse, the competition's washer thicknesses were also more
consistent (i.e. had lower variation).
What was wrong? It turned out that the company's managers and engineers violated an important rule of competitiveness:
They thought that the process was the end product: the washers produced. Therefore to monitor
the process they looked at the end product only. If this were the process, then, as seen in
Figure , the process is stable.
However, such a simplistic view can never lead to a proper understanding of the process. This is because it confuses the product with the process. Only by looking at the process, that is, how the product is created, can understanding be obtained. And only with understanding can steps be taken to improve the process and ultimately the product.
There are tools available to aid in understanding processes. Two of these, the
process flow diagram and the Ishikawa diagram are introduced in section .
For now, we will see how the most basic understanding of the process that
produced the washers revealed the main cause of the problems described above.
When they looked at the process, management realized that the washers were produced by
three different machines. Figure shows three plots of the same data plotted in
Figure , but the data produced by each machine are plotted separately.
These plots make the source of the problem immediately clear. Only machine 1 is producing
washers of acceptable thicknesses. Machine 2 is not stable, with a declining trend in
washer thickness evident. Machine 3 is producing washers of widely varying thicknesses.
When they investigated, repairmen found a worn cam responsible for the excessive variation in washers produced by machine 3. An overstretched spring on machine 2 was allowing a key control setting to drift, resulting in the trend in the thicknesses.