The key to achieving the specified target figures lies in process variation control which depends on statistical tools. Decreasing process variation involves multitasking of various parameters and aspects of a process.

How Can We Devise A  Six Sigma Metric?

We can begin by considering a simplified example where there are just 3 tasks, with each having one performance acceptability parameter. It is quite possible for all of them to have a correlation or not at all. When they are all correlated or if at least 2 of them are, you must look into how they behave upon varying the other/s. For example, if heating elongates some other part you will also be interested in knowing whether contracting the second reduces the temperature of the first. Emphasis must be placed on measuring every aspect of the task.

Considerations Before You….

Use Composite Metrics: You can choose to have a composite metric, which in specific areas, like a multivariate environment, might be exclusively helpful. You can consider using the composite metric for calculating overall deviations of a huge number of data over a considerably long period like a month. But you need to be careful to having any one parameter altering the others.

Break Up The Project: If time permits, work strategically on using all three tasks independently by breaking the project into multiple steps and go about working on the metrics one step at a time. In this case, you need to have a thorough understanding of the interaction of process flow and the statistical viewpoints.

Use A DFSS Scorecard: Some experts advocate using a DFSS scorecard. A DFSS scorecard will be extremely helpful when each of the parameters is a variable and when the standard deviations of the parameters are known along with DPMOs.

Some Guidelines

As a general rule, you must pay attention to each metric/parameter irrespective of whether or not you are using it. Analyze all metrics one at a time without overlooking the overall process. A company-wide metric for COPQ may not work very well, as many companies have learned from their costly mistakes. However, if you can, get the values like baseline definitions at the sector levels right. It takes an analytical approach to have an enterprise-wide metric. Some companies are practicing the measurements in percentage points for effectiveness or using the number of people trained.