A Current Assessment of Software Development Effort Estimation
Dirk Basten, Werner Mellis · 2011
In most cases, software projects exceed their estimated effort. It is often assumed that inaccurate estimates are the main reason for these overruns. Contrarily, our results show that this assumption is not reasonable in many cases. We conducted a survey to gather data on the current situation of software development effort estimation. Apart from objective criteria, we also asked the participants for their subjective perceptions of the estimates. The participants' perceptions indicate that they do not agree with the objective assessments as 82% perceive their estimate as 'good' despite large overruns and provide reasonable arguments for their perceptions. Furthermore, many projects do not re-estimate the effort due to change requests leading to meaningless comparisons of estimated and actual effort. As a consequence, research needs to find new measures to assess the actual estimation accuracy. Besides the actual estimation process, professionals need to intensify their effort to manage the estimation processes' surroundings.