Beyond Waterfall and Agile Methods: Towards a New Contingency Model for IT Project Management

Abstract:     
IT projects are notorious for not performing as planned, costing organizations vast amounts of time, effort, and money. Attempts to find a silver bullet, a methodology that can be used for all projects types, have failed. Some methodologies, for example the waterfall and the agile methodologies, have proven variably successful and have endured in practice despite not necessarily being optimal for a given project. We contend that more extreme methodologies are needed than what is currently practiced, but that these should apply to particular project types alone. Optimizing the methodology to the given project, we suggest, will yield better outcomes. Building on existing methodologies, we develop the notion of Degree of Linearity of the methodology as a continuum along which to rank order the methodologies, with waterfall being the most linear, agile methodologies being of medium linearity, and exploratory methodologies as the least linear. As organizations increasingly rely on innovation to compete, the more extreme project types will become increasingly important. Some of the most valuable projects must rely on methodologies that are more exploratory than IT projects traditionally have dared do. The proposed contingency model helps sort out these promising projects and helps ensure that their development will be as successful as the more traditional projects.

Keywords: Software engineering methodologies, management, organizational management and coordination

working papers series