Should We All Work in Sprints? How Agile Project Management Improves Performance

Abstract

Problem definition: Agile project management, in particular Scrum, is enjoying increased use in practice despite only scant scientific validation. This article explores how agile project management impacts project performance and execution. We compare the effects of agile sprints—short-term project phases characterized by time-boxed progression from one sprint to the next and self-imposed, phase-specific output goals—with those of traditional project management. Methodology/results: We decompose the two sprint elements of time-boxed progression and self-imposed, phase-specific output goals as factors in a 2 × 2 experimental design. We then conceptualize project execution as a simple real-effort task and conduct a controlled laboratory study. For a given duration, participants perform better with time-boxed progression as, without it, that is, with flexible progression, they spend too much time on early project phases at the expense of later ones. We refer to this effect as “progression fallacy” and show how it differs from well-known behavioral effects that cause project delays. Introducing self-imposed, phase-specific output goals in combination with time-boxed progression, as proposed by Scrum, does not significantly improve performance when compared with time-boxed progression alone. However, the combination of self-imposed, phase-specific output goals and flexible progression, as is common in traditional project management, amplifies the progression fallacy with the result that goal-setting has a negative performance effect. In two control treatments, we show that the progression fallacy is robust to planning and progression prompts despite some mitigation. Managerial implications: This study contributes evidence of higher project performance when working in agile sprints, which mitigate behavioral flaws present in traditional project management. Not only do these behavioral insights apply to project management; they are also relevant in the broader context of task completion.