How and Why Software Outsourcing Projects Drift—An Actor-Network-Theoretic Investigation of Control Processes

This study seeks to explain the perplexing phenomenon that many software outsourcing projects drift, i.e., they enter into a creeping process of targeting emergent goals often at the expense of losing sight of initial goals. Such drift is difficult to reconcile with the traditional logic of control found in the literature. According to this logic, clients should be able to ensure goal achievement through close monitoring. If drift occurs despite rigid control, this suggests that within the control process forces are at work that divert controls from their initial objectives. To better understand these forces in the control process and how they relate to drift, we contrast the logic of control with concepts and assumptions from actor-network theory (ANT). ANT allows us to understand the process of designing, enacting, and adapting controls as one of creating and changing actor-networks. Our longitudinal case study of four software outsourcing projects reveals that drift processes differ depending on three interconnected changes in the actor-networks, i.e., changes in who partakes in the (re-) negotiation of control mechanisms, what specific control mechanisms are (re-) defined, and how they are inscribed in the software artifact and the software task.