1.1. AVERAGE AND VARIATION

First, no one knows how to make anything "perfect." If you order 50 1.000" diameter ball bearings and then measure the bearings once you get them, you will find that they are not exactly 1.000" in diameter. They may be extremely close to 1.000", but if you measure them carefully, with a very good calibrated measuring device, you will find that the bearings are not exactly 1.000".

The bearings will vary from the 1.000" target in two ways. First, the average diameter of these 50 bearings will not be exactly 1.000". Whatever amount the average deviates from the target 1.000" is due to the bearing manufacturing process being off-center. Second, there will be a spread of measurements around the average bearing diameter. This spread of dimensions may be extremely small, but there will be a spread. This is due to the bearing process variation.

If the combination of the off-center bearing process and the bearing process variation is small compared with your needs, then you will be satisfied with the bearings. If, however, the combination of the off-center and variation is large compared with your needs, then you will not be happy. The Six Sigma methodology strives to make the total effect of an off-center process and process variation small compared with the need (tolerance). This is illustrated below in Figure 1-1.

Figure 1-1. Off-center and variation

If you understand the general concepts I just discussed, what follows is just terminology and detail.