Investigating intentional distortions in software cost estimation – An exploratory study

Abstract

Cost estimation of software projects is an important activity that continues to be a source of problems for practitioners despite improvement efforts. Most of the research on estimation has focused on methodological issues while the research focused on human factors primarily has targeted cognitive biases or perceived inhibitors. This paper focuses on the complex organizational context of estimation and investigates whether estimates may be distorted, i.e. intentionally changed for reasons beyond legitimate changes due to changing prerequisites such as requirements or scope. An exploratory study was conducted with 15 interviewees at six large companies that develop software-intensive products. The interviewees represent five stakeholder roles in estimation, with a majority being project or line managers. Document analysis was used to complement the interviews and provided additional context. The results show that both estimate increase and estimate decrease exist and that some of these changes can be explained as intentional distortions. The direction of the distortion depends on the context and the stakeholders involved. The paper underlines that it is critical to consider also human and organizational factors when addressing estimation problems and that intentional estimate distortions should be given more and direct attention.