As we consider the DMAIC process, it is important to bear in mind the fundamental principle of determinism: Y = f (X). Of course, Y is the outcome variable and X is the many variables that induce Y to exist. For any given problem-solving situation, we are behooved to recognize how DMAIC relates to Y = f (X):

1) Define Y and its performance expectation (ought-to-be-condition).

2) Measure Y in order to baseline its performance (way-it-is-condition).

3) Analyze the gap (the difference between “ought” and “is”).

4) Improve the gap by discovering the critical Xs and their best settings.

5) Control the optimal settings of each critical X so they don’t vary over time.

The question now becomes: “Should this process [DMAIC] be made different in form or substance as we change application scenarios?” The answer is resoundingly: “No.”

As we look across all possible application circumstances, it is expected that only a few of the related details will change. For example, what we call Y and X will change between situations. Of course, the types of data that reports on the behavior of Y and X will likely be variant (discrete versus continuous). In addition, the nature and extent of sampling will prove to be different (but extending from the same basic theory). Finally, the statistical tools that are employed to “crunch-the-numbers” will probably change in both type and order of execution. Thus, we recognize that the DMAIC methodology is quite generic and robust.