Like HowStuffWorks on Facebook!

How Six Sigma Works


DMAIC Model
The Six Sigma process doesn't end with the solution -- companies have to sustain progress over the long haul.
The Six Sigma process doesn't end with the solution -- companies have to sustain progress over the long haul.
­Digital Vision/Getty Images

Process is just as important as people. Most Six Sigma teams use what is known as the DMAIC model for process improvement. DMAIC stands for:

Define opportunity

Measure performance

Analyze opportunity

Improve performance

Control performance

Let's look at each of these steps in greater detail:

Define Opportunity: A Six Sigma project starts with a very specifically defined problem. Most people are used to defining problems broadly. For example, a business owner might say that accounts receivable is a concern. But such a definition won't work in Six Sigma. A better definition states the problem in quantitative terms. The business owner in the example above could revise his problem to get it to Six Sigma standards if he said, "Thirty percent of unpaid invoices are more than 45 days past due." With his problem specifically stated, he can now make meaningful measurements.

Measure Performance: Defining the problem is just the beginning. Next comes the most time-consuming part of the DMAIC methodology: determining the characteristics that influence the behavior of your process. This is accomplished by making measurements and collecting data.

Analyze Opportunity: It's not enough, however, simply to collect data. You must analyze the data using powerful mathematics and statistical tools. When you employ those tools properly, you get a clear picture of the variation in your process -- and how to limit it. Analysis reveals whether or not a problem is real or just a random event. If the event is random, then there is no solution within the Six Sigma framework.

Improve Performance: Once it's been determined that the problem is real and not a random event, Six Sigma teams look to identify possible solutions. Solutions must be tested to see how they interact with other input variables. Ultimately, the team chooses the best solutions for implementation.

Control Performance: It might seem that the application of a solution ends the Six Sigma process, but it doesn't. To make sure a solution can be sustained over the long term, control planning is required. Control planning involves collecting quality control data and verifying measurements according to a regular schedule. This ensures that processes continue to run efficiently and deliver peak performance.

Six Sigma experts use specific methods to carry out the DMAIC model -- find out about them on the next page.­ ­