The Average Method adds the defects for all processes together, and divides that number by the sum of the samples for all processes. This method is better for start-up situations. As more processes are brought on line, the DPM level will not continue to climb just because you are collecting more data. This method is also useful for situations where you run a high-defect process for a short period of time. If you use the Additive Method, this process will show up as a spike on the DPU chart; but if you use the Average Method, the spike will be somewhat smoothed.
Here is an example of the same data plotted using the Averaged Method:
Average Method: ( DPM = Sum of defects for all operations / sum of all Units * 1,000,000)
|
Defects |
Sample Size |
DPM |
Process A |
7 |
50 |
|
Process B |
5 |
50 |
|
Process C |
3 |
50 |
|
Total DPM (Average Method) |
15 |
150 |
100,000 |
The Average Method is less sensitive and is, therefore, attractive to customers whose processes change dramatically or who need to add or drop processes from the database frequently. This method should be used if you want to track the number of opportunities in all operations.