Measuring risks in software development projects

Abstract

There is a consensus that poorly managed adverse factors lead to project failures. Risk is an inherent part of any project. In other words risks are always present and may cause problems that lead to projects failure or poor performance. In software development projects, this scenario is not different. However, the practice of risk management in software development projects environments involves a high level of subjectivity, making it difficult to perceive the real impacts not only of risk management practices but also of the influence of risk on projects. A possible alternative for reducing this gap is through measurement-based approaches. In this sense, indicators are a relevant instrument for project’s assessment and decision-making, since they are used to represent information in a clear and objective way. In this context, the objective of this thesis is to develop studies about software risk measurement and proposing and evaluating an indicator, called Project Risk Index (PRI), whose goal is to measure the risk level of a software development project at a given moment. This indicator starts from the premise that, in addition to the identified risk factors, some project’s characteristics also contribute to raise the project risk levels. To achieve the research objectives, we started with an exploratory case study, seeking to resolve relevant issues in refining the proposed sources of information. Next, we performed a mapping of the most relevant risk factors in software development projects through a systematic literature review. Finally, two case studies were conducted with the proposed indicator, aiming to evaluate it and identify improvement points. The results show the existence of an indicator aiming to measure risk level of a software development project is crucial for better perception and clarity of the most critical project items. Regarding the proposed indicator, there are indications that its application may be useful and effective in risk management. Information about risk factors, combined with project’s characteristics may be relevant for decision-making by the managers at the tactical and even strategic levels, enabling continuous and systematic monitoring of software development projects risk levels.