Antidotes for high complexity and ambiguity in software development

Abstract

In a longitudinal study of 47 software development teams, we investigate interactions between team and technology factors and the degree of complexity and ambiguity of the projects themselves. From the literature, we propose a theoretical model that identifies a characteristic of the technology (modularity) and a characteristic of the team process (conflict resolution) used during system development, as effective for minimizing the adverse effects of high task-based complexity and ambiguity (those tasks for which multiple acceptable solutions exist). We hypothesize that modularity and conflict-resolution techniques will account for a significant amount of the variance in user satisfaction for highly complex and ambiguous projects, but that this will not be the case for simple and unambiguous projects. Our findings confirm this hypothesis, indicating that effective conflict resolution and modularity are associated with significantly higher client satisfaction six months after implementation for all projects. An explanation for these findings is offered, followed by implications for theorists and practitioners.