Content area
Full Text
Keywords Information networks, Case studies, Technology led strategy, Management failures
Abstract This study investigates the potential of actor-network theory (ANT) for theory development on information technology project escalation, a pervasive problem in contemporary organizations. In so doing, the study aims to contribute to the current dialogue on the potential of ANT in the information systems field. While escalation theory has been used to study "runaway" IT projects, two distinct limitations suggest a potential of using ANT: First, there is a need for research that builds process theory on escalation of IT projects. Second, the role of technology as an important factor (or actor) in the shaping of escalation has not been examined. This paper examines a well-known case study of an IT project disaster, the computerized baggage handling system at Denver International Airport, using both escalation theory and ANT. A theory-comparative analysis then shows how each analysis contributes differently to our knowledge about dysfunctional IT projects and how the differences between the analyses mirror characteristics of the two theories. ANT is found to offer a fruitful theoretical addition to escalation research and several conceptual extensions of ANT in the context of IT project escalation are proposed: embedded actor-networks, host actor-networks, swift translation and Trojan actor-networks.
Introduction
Escalation is a phenomenon in which an organization or other acting entity persists in pouring resources into a failing course of action (Staw, 1976). In the field of information systems, the problem of project escalation is an important issue given that escalation in software projects is quite common (Keil et al., 2000), that it is often a precursor of failure (Ewusi-Mensah and Przasnyski, 1991; Lyytinen and Hirschheim, 1987) and that the frequency and costs of IS development failures are considerable (KPMG, 1995; Johnson, 1995). The problem of IT project escalation remains highly relevant, while the remedies are far from well known (Keil and Robey, 1999, 2001).
Notwithstanding their important insights in understanding runaway IT projects, existing escalation studies present two distinct limitations that suggest an opportunity to apply ANT. First, there is a need for research that builds process theory on the escalation of IT projects (Montealegre and Keil, 2000), a shortcoming shared with escalation theory (Ross and Staw, 1993). Second, previous escalation studies have not investigated whether...