Identifying factors affecting software development cost and productivity

  • Authors:

  • Robert Lagerström;Liv Marcks Würtemberg;Hannes Holm;Oscar Luczak

  • Affiliations:

  • Industrial Information and Control Systems, The Royal Institute of Technology, Stockholm, Sweden 100 44;Industrial Information and Control Systems, The Royal Institute of Technology, Stockholm, Sweden 100 44;Industrial Information and Control Systems, The Royal Institute of Technology, Stockholm, Sweden 100 44;Industrial Information and Control Systems, The Royal Institute of Technology, Stockholm, Sweden 100 44

  • Venue:
  • Software Quality Control
  • Year:
  • 2012

Quantified Score

Hi-index 0.00

Visualization

Abstract

Software systems of today are often complex, making development costs difficult to estimate. This paper uses data from 50 projects performed at one of the largest banks in Sweden to identify factors that have an impact on software development cost. Correlation analysis of the relationship between factor states and project costs was assessed using ANOVA and regression analysis. Ten out of the original 31 factors turned out to have an impact on software development project cost at the Swedish bank including the: number of function points, involved risk, number of budget revisions, primary platform, project priority, commissioning body's unit, commissioning body, number of project participants, project duration, and number of consultants. In order to be able to compare projects of different size and complexity, this study also considers the software development productivity defined as the amount of function points per working hour in a project. The study at the bank indicates that the productivity is affected by factors such as performance of estimation and prognosis efforts, project type, number of budget revisions, existence of testing conductor, presentation interface, and number of project participants. A discussion addressing how the productivity factors relate to cost estimation models and their factors is presented. Some of the factors found to have an impact on cost are already included in estimation models such as COCOMO II, TEAMATe, and SEER-SEM, for instance function points and software platform. Thus, this paper validates these well-known factors for cost estimation. However, several of the factors found in this study are not included in established models for software development cost estimation. Thus, this paper also provides indications for possible extensions of these models.