When working on a Six Sigma project, accurate methods to measure process performance metrics are critical for understanding the current state of a process and the value of changes made. Four of the most common measurements are Defects Per Unit (DPU), Defects per Million Opportunities (DPMO), Parts per Million Defective (PPM), and the Rolled Throughput Yield (RTY).

The following provides how each is used. However, it’s important to first understand the difference between two terms commonly used in connection with these performance measurement tools. The first is “defect.” The second is “defective.”

  • Defect: This refers to a flaw or discrepancy in an operation or on an item where more than one flaw (defect) can be found. For example, a car is one finished unit in a process. A car also contains many different areas that are assembled to create a finished vehicle. Any of these areas – the seats, the dashboard, the engine, the exhaust system, etc. – could have defects. Given that, 10 finished cars could have more than 10 defects.
  • Defective: This refers to a decision made that an item is unacceptable, typically based on an accumulation of multiple defects. Again, using the car scenario, this means that 10 cars can have a maximum 10 defective units, because each car represents one unit.

Another way to look at this is opportunity vs. units. A unit is the final product delivered to a customer. It can contain many defects and be found to be defective. Opportunities represent everything that goes into making a unit – materials, labor, delivery, etc. Each of these opportunities has the potential of having a defect.

Defects Per Unit (DPU)

DPU measures the average number of defects per every product unit. It’s found by dividing the total number of defects found by the number of units.

For example, if 30 units are produced and a total of 60 defects have been found, the DPU equals 2.

Defects Per Million Opportunities (DPMO)

This represents a ratio of the number of defects in one million opportunities. In other words, how many times did you have a flaw or mistake (defect) for every opportunity there was to have a flaw or mistake.

The formula for calculating DPMO is as follows.

For example, consider a form that contains 15 fields of information. If 10 forms are sampled and 26 defects are found in the sample, the DPMO is:

It’s also possible to translate DPMO to a Six Sigma level. The goal is to reach 3.4 defects per 1 million opportunities.

Parts Per Million Defective (PPM)

The PPM represents the number of defective units per 1 million units. Again, using the car scenario, the PPM would include the total number of defective cars – cars determined to be too flawed to be sold – per every 1 million cars manufactured.

PPM is arrived at by simply taking the number of defective units in a same size, dividing that number by the total sample size, and multiplying by 1 million.

For example, a sample of 50 cards found that three are defective. The PPM defective is then:

Rolled Throughput Yield (RTY)

RTY (also known as the First Pass Yield) measures the probability (or percentage of time) that a manufacturing or service process will produce a defect-free unit. This requires mapping out a process to determine how many steps it involves.

The reliability formula for a system in series with n process steps is: Rs = (R1) (R2) (R3) (R4) … (Rn)

Since the reliability of a process step is the yield of that process step when quality is the performance metric, this formula then becomes: RTY= (Y1) (Y2) (Y3) (Y4) … (Yn) where Y is the yield (proportion good) for each step

For example, a four-step process has a yield of 0.98 in step 1, 0.95 in step 2, 0.90 in step 3, and 0.80 in step 4.

RTY = (0.98)(0.95)(0.90)(0.80) = 0.67032

This means that only 67.032% of the units completed on this process will make it through all four steps without needing any rework or repair.

Once an organization understands the different performance measurement tools and how to use them, the important next step is to determine how to calculate baseline sigma and determine what other metrics to measure.