Requirements Prioritisation and Retrospective Analysis for Release Planning Process Improvement

Abstract: The quality of a product can be defined by its ability to satisfy the needs and expectations of its customers. Achieving quality is especially difficult in market-driven situations since the product is released on an open market with numerous potential customers and users with various wishes. The quality of the software product is to a large extent determined by the quality of the requirements engineering (RE) and release planning decisions regarding which requirements that are selected for a product. The goal of this thesis is to enhance software product quality and increase the competitive edge of software organisations by improving release planning decision-making. The thesis is based on empirical research, including both qualitative and quantitative research approaches. The research contains a qualitative survey of RE challenges in market-driven organisations based on interviews with practitioners. The survey provided increased understanding of RE challenges in the software industry and gave input to the continued research. Among the challenging issues, one was selected for further investigation due to its high relevance to the practitioners: requirements prioritisation and release planning decision-making. Requirements prioritisation techniques were evaluated through experiments, suggesting that ordinal scale techniques based on grouping and ranking may be valuable to practitioners. Finally, a retrospective method called PARSEQ (Post-release Analysis of Requirements SElection Quality) is introduced and tested in three case studies. The method aims at evaluating prior releases and finding improvement proposals for release planning decision-making in future release projects. The method was found valuable by all participants and relevant improvement proposals were discovered in all cases.