Comparison of two models of success prediction in software development projects

Background: The size and complexity of software development projects are growing. At the same time, the proportion of successful projects is still quite low according to the previous research. One way to approach this problem is to develop and use methods that can predict project success beforehand and act accordingly. Aim: The objective of this study is to compare two existing models of success prediction (The Standish Group and McConnell models) and to determine their strengths and weaknesses. Method: The research was done as an empirical study. A survey with structured forms and theme-based interviews were used as the data collection methods. The comparison is made with observations from 48 projects in Russia, Belarus, and Ukraine. In addition, 19 interviews were conducted during the study. Conclusions: The results show that The Standish Group has a tendency to overestimate the problems in a project. McConnell predicts successful projects pretty well but underestimates the percentage of unsuccessful projects.