Title: Spreading of performance fluctuations on real-world project networks

Abstract: Understanding the role of individual nodes is a key challenge in the study of spreading processes on networks. In this work we propose a novel metric, the reachability-heterogeneity (RH), to quantify the vulnerability of each node with respect to a spreading process on a network. We then introduce a dataset consisting of four large engineering projects described by their activity networks, including records of the performance of each activity; such data, describing the spreading of performance fluctuations across activities, can be used as a reliable ground truth for the study of spreading phenomena on networks. We test the validity of the RH metric on these project networks, and discover that nodes scoring low in RH tend to consistently perform better. We also compare RH and seven other node metrics, showing that the former is highly interdependent with activity performance. Given the context agnostic nature of RH, our results, based on real-world data, signify the role that network structure plays with respect to overall project performance.