DPM is an estimate based on the number of defects found and the number of samples tested. Another way of estimating DPM is called the "Best Estimate Method" which uses the following equation:
DPM = X2 (.50,2r) * 106 / 2N
Where r is the number of defects found and N is the number of samples tested.
The Best Estimate method is turned on in the Administration module.
For a complete discussion of the best estimate method, see Six Sigma Producibility Analysis and Process Characterization, Mikel J. Harry and J. Ronald Lawson. Addison-Wesley, 1992, pages 6-19 to 6-21 and D-1.
Customers who have small lot sizes may find that the standard DPM calculations described above are not sensitive enough for their processes. For example, if you only produce a few items each day, it is entirely possible that you would have zero defects for several days in a row, followed by one day when you have one defect out of a few items. Your PPM levels might run at zero for several days, followed by a jump to 250,000 PPM (one in four). These wild swings do not give a realistic presentation of the process.
We provide two ways of making DPM values more conservative:
The Conservative Calculation Option always adds one defect to the total found.
The Never Report Zero Option always calculates DPM based on at least one defect found, even if no defects are found.
These two ways of calculating more conservative DPM values will be affected more when the number of samples is small. The change in DPM becomes negligible as the number of samples gets very large. You can configure GainSeeker to use one or both of these methods in the Administration module.