Evidence-Based Evaluation of Effort Estimation Methods

In the selection of appropriate effort estimation methods, there are different questions, depending on the situation in the concrete IT environment. The main goal is of course the best and most accurate estimate of the efforts and costs for their own IT department in conjunction with least cost to the estimate itself. On the other hand, questions of comparability within the company or (internationally distributed) business areas seems to be relevant in a global marketplace of software production. Furthermore another target is to compare themselves with competitors and make the effort estimation transparent outside. This paper deals with these issues, examining the various cost estimation methods for their importance or "evidence" and the resulting selection criteria for individual applications. Our paper is based on our theoretical and industrial experience in effort estimation over the last twenty years in our software measurement community.