The When-Who-How analysis of defects for improving the quality control process

Most large software products have elaborate quality control processes involving many tasks performed by different groups using a variety of techniques. The defects found are generally recorded in a database which is used for tracking and prioritizing defects. However, this defect data also provides a wealth of information which can be analyzed for improving the process. In this paper, we describe the when-who-how approach for analyzing defect data to gain a better understanding of the quality control process and identify improvement opportunities. At the component level, the analysis provides the capability to assess strength of dependency between components, and new ways to study correlation between early and late defects. We also discuss the experience of applying this approach to defect data from an earlier version of Windows, and the improvement opportunities it revealed.