Making-sense of the impact and importance of outliers in project management through the use of power laws

The academic literature and popular press has chronicled large IT project failures for the last 40 years. Two points of contention surround this debate. First, quantitative studies found mixed support of a wide-spread crisis, questioning the representativeness of failure cases. Second, organizational theories disagreed on underlying assumptions about the nature of uncertainty, in particular about stability, locus of control, and controllability of the causes of IT project disasters. To advance the understanding of these two gaps four hypotheses were tested with a sample of 4,227 IT projects. The findings showed that outliers are stable phenomena following power laws, occurrence and impact of outliers differs between public and private sector, benefits management is associated with thinner tails and lower risk, and agile delivery methods do not statistically significantly influence the thickness of the tails. In sum, outliers are stable and non-random phenomena. They matter more than medians or means when it comes to IT project risk. Second, the notion of outliers bridges the gap between qualitative and quantitative studies. The findings also show that causes of outliers are, at least to some extent, internal and controllable by organizations. Lastly, the paper draws implications for organizational decision-making, learning, and risk management.