The Challenge of Introducing a New Software Cost Estimation Technology into a Small Software Organisation.
Abstract
Fostering innovation is the key to survival in today's IT business and is exemplified by introducing new technologies and methods to improve the development processes. This paper presents a follow-up case study of technology transfer in a small software organisation. A new software estimation technique, Web-CoBRA was introduced to a small software company to improve their software estimation process. Web-CoBRA was considerably more accurate than the company's current estimation process. However, despite management being aware of this improvement, the company has not fully adopted the new method. We used interviews and the Technology Acceptance Model (TAM) questionnaire to assess the extent to which Web-CoBRA was used by the company. We found take-up of part of the Web-CoBRA technology but the full technology and the support tools were not used. We identify the reasons for the failure to adopt the Web-CoBRA technology and identify several areas for improving technology transfer activities. Keywords: Software Cost Estimation, Software Engineering, Technology Transition, Web Development 1. Introduction The primary motivation for this research is to investigate the impact of introducing a new software cost estimation technology into a small software development company. Software engineering researchers have achieved considerable success in improving software estimation accuracy in the last two decades[1], however software companies continue to find it difficult to adopt these methods[2]. As a result, software cost estimation has barely penetrated the mainstream of software engineering [2, 3]. Nevertheless, accurately estimating a new software project early in the development life cycle is still a goal of every project manager. In late 2001, the Fraunhofer Institute for Experimental Software Engineering (IESE) and the Centre for Advanced Empirical Software Research group (CAESAR) at the University of New South Wales collaboratively investigated the appropriateness of the CoBRA [4] method to predict development effort for web applications in the context of a small software development company. The research team conducted onsite research for six months, and developed