In this chapter, we will discuss the calculation of defects per million opportunities. We begin with the number of defects found in a test sample, divided by the total number of defect opportunities for that sample. Then we multiply that number by 1,000,000. For example, if we found 50 defects after inspecting a sample of 1,000 units, we would divide that 50 by 1,000, and get to a factor of 0.05. We would then multiply that by 1,000,000. Our defects per million opportunities would be 50,000. There is an alternative DPMO formula that’s also very important to understand. It is different because we have the same number of defects in the sample. But we get to the DPMO by taking the sample sizes times the defect opportunities per unit. For example, let’s pretend that we are making steel pipes. In this case, our testing method is a destructive method. Meaning that in order to actually get the measurements, we have to cut a piece out of the pipe. And measure it in four different places to determine if it’s going to meet our quality requirements, and to capture the data. Since we are taking four measurements per pipe, there are four defect opportunities per pipe that we test.
Destructive testing is a case where we may want to take a specific sample size, say 10 units every hour for an 8 hour shift. So we would have a sample size of 80 and 4 defect opportunities per unit. This would give us a numerator of 320. And then we can count the number of defects from that sample, let’s say that there were 10. By dividing 10 by 320, and multiplying that number by 1,000,000, we find out that the defect per million opportunities, without having to destroy every piece of pipe, is 31,250. Let’s examine this in detail. So we want to find the DPMO for a process of completing application forms in a financial institution. So here, we’re going to want to count the number of defects we found in the number of application forms. So let’s just say that we found, initially, 45 mistakes in our forms. Now we go on to say, okay, so let’s take a look at the sample size. We’re going to divide that by our sample size, which would be 300 different forms that we looked at. But we’re not done yet. We now have to think about how many defect opportunities are there in each form? Well, in this case, we have 10 different fields that our operators or technicians need to fill out as they’re doing the work. And of course, we’re going to have to multiple that by 1,000,000 opportunities to get to the DPMO calculation.
So if we work out the numbers here, we would have 0.015. Which would then be the calculation of 45, divided by the product of 300×10. Multiplied by that 1,000,000 opportunities we talked about, and what do we get? We have a DPMO value of 15,000. So even though we didn’t find that many, 45 out of 300. Because we extrapolated to where these opportunities are, we can see that our DPMO is a pretty high number. Let’s move on. Now we can use DPMO to understand the Sigma level. This table is populated with the DPMO numbers and the corresponding Sigma values. In the previous example, we had the DPMO calculation of 15,000. So according to this table, this would suggest that we have a Sigma level between 3.0 and 4.0 sigma. If we imagine the normal distribution of samples in a bell-shaped curve, one standard deviation to the left or to the right of center. The opportunity for defects is pretty high, or about 690,000 defects out of 1 million opportunities. If our DPMO is calculated as 308,000, we are probably operating at plus or minus 2.0 standard deviations from the mean. Defects per million opportunities decreases until we get to a 3.4 defects per million opportunities, or 6.0 Sigma.