A Tree-Based Approach for Addressing Self-Selection in Impact Studies with Big Data

Outsourcing is fundamental to firm competitiveness. Firms are increasingly externalizing a variety of core business functions such as research and development, product development, and marketing to achieve diverse strategic objectives (Mani et al. 2012, 2013). Yet, despite their extended reach and impact, outsourcing initiatives involve high failure rates with adverse impacts on various critical performance metrics. Prior research in information systems attributes heterogeneity in outsourcing performance to efficacy of contract design and management (Mani et al. 2013; Susarla et al. 2011). The right contract aligns incentives between the client and the vendor to engender cooperative behavior that is necessary for effective execution of the outsourced task and adaptation to disturbances. Yet, what drives design of outsourcing contracts as well as how performance varies across these contracts are issues that have only recently attracted empirical attention.