The Benefits and Challenges of Planning Poker in Software Development: Comparison Between Theory and Practice
Planning Poker is a collaborative software effort estimation technique widely used in agile software development methodologies such as Scrum and Extreme Programming (XP). Proponents of Planning Poker claim that the method has the benefits of multi-perspective effort forecasting, promoting group interaction and participation, while avoiding the risk of the “anchoring” cognitive bias, leading to accurate effort forecasting without unrealistic precision. While there is body of empirical research about Planning Poker that investigates the estimation accuracy of the technique in practice, there is not any research that looks at the detail of the implementation of the process of Planning Poker estimation in practice and whether these claimed benefits are realized in practice. This thesis focuses on analyzing the practical execution of Planning Poker in a detailed case study, aiming to identify how the practical process of Planning Poker differs from the theory, and whether the expected benefits are seen or not, and why.
The aim of this investigation is to provide a comprehensive conceptualization of Planning Poker estimation in practice, as well as understanding of the difference between practice and theory. In addition, this study also expects to deepen an understanding of the activities of the Planning Poker technique that contribute to the observed benefits and challenges. This will be helpful for software estimators to optimize execution of the Planning Poker technique to maximize the claimed and observed benefits it can provide.
The approach to this investigation is based on analyzing video and audio recordings of two Planning Poker sessions undertaken by a team in a case organization. This is supplemented by field notes and photographs related to the Planning Poker meetings. The study starts with an in-depth literature review to explore the theory related to some special behaviors in practical Planning Poker estimation meetings. Both Planning Poker meetings are held by the same development team for the same software project, and conducted in two different development iterations as part of Sprint planning using Scrum as their software development approach. This thesis explores the differences between the process inferred from analyzing the meetings and theory pf Planning Poker. The differences between both meetings also is explored. This thesis also explores whether those claimed benefits in literatures are achieved in practical Planning Poker estimation meetings, and discusses some of the challenges observed in both meetings.
The findings of the research includes identification of several different estimation process patterns, such as a gradual changing of the benefit of promoting group interaction is achieved well in practice, whereas avoidance of the “anchoring” effect and promoting individual participation, are largely not achieved. In addition, this thesis identifies some of the practical challenges in implementing Planning Poker, for example domination of the discussion by a single practitioner.