The following measurement necessities are the same whether you’re measuring business, service, process or laboratory variables. Together, they constitute a measurement plan.
Definition of purpose: Why is a measurement being made? What process or variable is being measured? What will the resulting data be used for?
Statement of the required measurement performance indicators (accuracy, precision, resolution): These may be determined by company policy, by adherence to a published standard or to an analysis of the requirements based on use, ability to measure, etc.
The unit or variable being measured and a statement as to why measuring that particular variable supports the purpose of the measurement.
An operational definition: A detailed description of the measurement process, given in such a way that it can be easily understood.
Example: An operational definition for the measurement of a sales–fulfillment cycle time might be: “The time interval to be measured begins when the sales department places a validated order form in the sales order out box, and ends when the completed, boxed order is delivered to the loading dock for pickup.”
An analysis plan: A typical example is a monthly report that makes comparisons to the previous month, same month last year and year–to–date. Besides the problems that automatically come from being retrospective, these data are hard to interpret for two reasons—they are not presented graphically, and there’s not enough context.
The simplest analysis plan here is a control chart, which would provide a graphical context that shows the continuity of changes over time, plus some analysis (control limits) that enables the viewer to differentiate among common causes, special causes and random variation.
Excerpted from Philip Stein’s “Effective Measurement of Business Performance,” ASQ Annual Quality Congress proceedings, May 1998.