Under Pressure: The Effects of Iteration Lengths on Agile Software Development Performance

摘要

The iterative cycles used in agile software development create a series of deadlines for project teams. We investigate the optimal iteration length under behavioral conditions for agile software projects. Based on system dynamics simulation, we argue that both very long (as in sequential development) and very short (monthly) iterative cycles create huge and disruptive fluctuations in schedule pressure experienced by the team. These lead to poor performance in terms of quality, as well as costs and time compared with optimal iterative cycles of two to three months. The monthly cycles used in many agile practices might be too short to ensure optimal performance.