Testing has become an integral part of most software projects. It accounts for even as high share as 40% of the overall work effort. At the same time software projects are systematically exceeding their effort and schedule forecasts. Regardless of the overruns and big share of testing, there is very little advice for estimating testing activities. This case study research assesses the current practice of estimating testing activities, and the impact of these practices on estimation and project success. Based on the interviews with 11 stakeholders involved in two case projects and examination of project documentation, this study shows that companies easily deviate from their standard procedures, when estimating testing. This may lead to even severe estimation errors. The deviations can be explained by negative attitudes towards testing. Furthermore, this study shows that the extant literature has sparsely addressed estimation of software testing.
Jurka Rahikkala (Vaadin, University of Turku), Sami Hyrynsalmi, Ville Leppänen (University of Turku): Accounting Testing in Software Cost Estimation: A Case Study of the Current Practice and Impacts