Time for testing at an intermediate Dutch SME

Recent development projects as well as their worried project managers show an urge for improving software quality by proper testing. High post-release defect levels, uncertain software quality and lack of testing methodology underlie this urge. A case study was performed on two of the six SME business units to first assess current testing workings and thereafter to identify a desired future testing state. The various applied research methods – literature studies, semi-open interviews, and a survey – together provide a multi-perspective overview of SME’ testing methodology. The research shows testing at SME to reside in a troubling state, probably in a worst-case scenario. This is because various aspects of testing are considered weak: responsibilities are too informal, there is a lack of testing knowledge amongst employees, testing holds a low priority, there’s a lack of testing resources, quality management lacks objectivism and finally customers are offered little guidance in proper testing. Furthermore, all types of testing – except for acceptance testing – are hardly applied and testing is performed without a methodology. To improve these weaknesses, three best-practices have been identified out of structured literature studies on agile testing methods. These are: Test-Driven Development, which holds a new paradigm on testing that shortens and increases amounts of test cycles; Continuous Integrated Testing, forcing regression and integration testing via automated test runs; and finally the use of Metrics to provide real-time insights in software quality and performed tests, enabling planning, steering and control. Together these practices will solve current testing methodology issues, and improve software quality significantly by reducing defect injection rates and by finding and fixing defects closer to their origin.