A goal question metric based approach for efficient measurement framework definition

Abstract

In software engineering, measurements can be used to monitor, understand and improve software processes as well as products and resource utilization. Commonly, measurement frameworks are ambitious undertakings that require large data collection and analysis efforts. Unfortunately, such frameworks usually become too extensive, resulting in graveyards with a lot of data never being analyzed and used. One well-known way for defining measurements that tries to focus on the most important ones is the Goal Question Metric (GQM) approach. GQM focuses on eliciting goals and questions as drivers for finding metrics necessary to collect. This means that there is always a purpose with the defined measurements. GQM solves part of the problem mentioned, but it is still possible that the measurements defined become too many. In this paper, an extended GQM approach is presented that facilitates identification of and focuses on the most important measurements for an organization. The approach makes it possible to be more efficient when it comes to the application of measurement frameworks. Besides the presentation of the approach, two case studies are presented where the suitability of the approach is investigated in an industrial context. In the case studies, the suggested approach has been applied on two different processes within software development: change management and requirements engineering. From these case studies, a number of experiences are highlighted and discussed as issues to consider when applying the approach.