Abstract
A major share of software project investment is assigned to activities concerning the detection and removal of defects. Software project managers tend to apply the most efficient QA techniques to assure low defect density within their software project. However, the criteria of selecting a QA technique based on its efficiency is not always safe and cost effective. Software defects vary in their severity in terms of the magnitude of their negative impact on both the testing process and the whole project. Some defects would make intense implications if it passed to the operational use as it belongs to significant functional components of the software. The consideration of potential risk with some defects types should be taken into account when selecting a QA technique to avoid future failure. In this paper, we build on previous work of software quality optimization by proposing a model whereby QA decisions are normalized by the risk associated with their defects detection and removal activities.