Since the entire premise of having a Six Sigma process is based on measurement, comparison of measurements and setting corrective action plans, it is crucial that the variables (metrics) being measured are appropriate. Wrong measurements lead to wrong decisions which are expensive and a wastage of time and resources.

As explained in the Define module, there must at least be two types of variables being measured i.e. the primary metric and the secondary metric. The primary metric is the variable whose value must be checked for improvement, whereas the secondary variable is the variable whose value must be checked to ensure that it does not get deteriorated.

Apart from that one also needs to ensure that both inputs and outputs are being measured. Since the process has the form Y=f(X) both, the Y and the X must be carefully monitored real time to ensure that f() i.e., the process is behaving as expected.

Here are a few characteristics that the variables being used to create a Six Sigma control plan must possess:

Characteristics of Variables

Must Display the State of Affairs: There is always a difference between the measurements that variables show and what the desired state of affairs is. This usually happens when a single metric is used to understand the state of affairs.

For instance, if in a call center, the only metrics used is call handling time, agents are likely to start hanging up calls without giving appropriate service to meet their numbers. It is therefore essential that management must brainstorm ways in which the system can be broken. This means that they must try to find out ways in means in which all the variables are showing desired values, however the process is still out of control. If such a scenario exists,the variables must be corrected to include this scenario.

Critical: In a Six Sigma process, there are likely to be many variables. However, as Pareto analysis has shown, all these variables may not be equally important. Hence, the metrics chosen must have a material impact on the process. It is impossible to monitor all the variables without building in mind-boggling complexity in the system.

Easy to Measure: The variables being chosen to measure must not require installation of new equipment or machinery. If this is the case, a cost benefit analysis must be done to justify the added costs. Also the variables to be recorded must not require considerable intelligence or expertise on the part of the person recording them.

Produce Correct Measurements: Needless to say that the variables must be put through the Measurement Systems Analysis tests described in the Measure Module. The success of the entire project depends upon the reliability of this data. Therefore one must ensure that the data is in fact reliable!

Measurable over a Period of Time: Also, the metrics must facilitate easy collection of data over a period of time. This is because control charts are time series plots and data needs to be collected over a period of time and plotted on these graphs.

The correct variables are the fundamentals of a good control plan. The procedure may look deceptively simple. However, wrong variables can have significant impact on your project, inclusing making your project a failure.