|Authors||K. J. Moløkken-Østvold, M. Jørgensen, S. S. Tanilkan, H. Gallis, A. C. Lien and S. E. Hove|
|Title||Project Estimation in the Norwegian Software Industry - a Summary|
|Afilliation||Software Engineering, Software Engineering|
|Publication Type||Technical reports|
|Year of Publication||2004|
|Publisher||Simula Research Laboratory|
This report provides an overview of the results obtained from a survey on project estimation in Norwegian software companies. The survey was conducted between February and November 2003. The main results are: • We observed that 76% of the projects used more effort than estimated, while 19% used less. The average effort overrun was 41%. • Average effort overrun was 67% in projects with a public client, compared to an average effort overrun of 21% for projects with a private client. • Projects that used an incremental or evolutionary development approach had an average effort overrun of 24%, as opposed to the average 55% overrun for projects that used a waterfall-based development approach. The frequency and magnitude of effort overruns found in this survey seems to be similar to results reported from surveys conducted in other countries in the past 20 years. The observed differences in effort overruns between private and public projects may be caused by differences in bidding procedures, level of client involvement or acceptance procedures. In order to reduce the risk of effort overruns, software companies should: • Focus on analyzing their own estimation performance, and invest in estimation improvement (for instance through experience databases or work breakdown structures). • Differentiate risk buffers based on the type of the customer, development approach and the size of the project. • Try to establish an “as close as possible” dialogue with the customers (e.g. through an incremental development approach). All companies should analyse completed projects, in order to benchmark their performance. This allows for improvement efforts to be identified.
Results from the BEST-Pro (Better Estimation of Software Tasks and Process Improvement) survey