Ensuring High Quality Software Metrics
by WatchIT
Doug Brindley, President of Software Productivity Research, identifies the software metrics that best-in-class organizations collect, and explains how even well-intentioned metrics programs can become dysfunctional. He also presents a straightforward technique for ensuring that your organization both collects and uses high-quality software metrics. Joining Brindley in an interview is Stan Rifkin, founder of Master Systems Incorporated and a noted process improvement and change management expert. Brindley begins the program by making the case for software measurement, and describes the metrics most commonly collected and analyzed. Then, he discusses what he calls “evil-twin” concepts: “Measurement Noise” and “Dysfunctional Metrics.” Next, from a classic case of dysfunctional metrics, Brindley outlines how the efforts of an employment office measurement program worked against its goals by “penalizing the truth” and reinforcing subversive behavior and record keeping. Finally, he examines the goal/question/metric (GQM) paradigm. This is a powerful technique that helps ensure metrics are functional because the paradigm is based on two important assumptions: that measurement programs should be goal-based, and that goals and metrics are defined and customized to the individual organization's needs.
...