Performance Evaluation of Software Development Teams: A Practical Case Study

Software development projects have become a challenge for both industry and academia regarding the performance evaluation of teams. Recently, a Stochastic Automata Networks (SAN) model was proposed as theoretical representation for performance prediction of software development teams. In this paper, we present an exercise of such SAN analytical modeling for a practical case study from an Information Technology company that has multiple sites and different participants' roles and expertises. We present the matching of our model predictions with the actual project observations. Then, we focus our attention on the central entity varying its availability and the level of provided support in order to observe the impact on the participants' performance. We summarize our study with further discussions of numerical results and possible model extensions.