A Network Perspective on the Dynamics of Market Structure for Outsourced IT Services: A Bayesian Inference Approach

Abstract

While considerable evidence exists on contract design for outsourced information technology (IT) services, limited work has systematically examined the partner selection strategies of firms seeking to outsource their IT services in tandem with overall dynamics in the market for IT outsourcing. Conceptualizing the set of vendor firms as an organizational field, we consider several mechanisms that influence the dynamics of vendor selection by a client firm in the market for IT services. First, we consider preferential attachment by which firms tend to prefer to contract with central vendors. Second, we consider prior interactions, which serve as an important source of information for both clients and vendors about the reliability and capabilities of their partners. Third, we consider the experience and diversity-based exchange logic, whereby client firms’ benefit through the depth and breadth of experience of vendors. Finally, we consider multi-connectivity based logic of exchange wherein a client can benefit from the heterogeneity of information sources. Analyzing the dynamics of vendor selection poses a number of statistical and computational challenges in examining partner selection for a large set of vendor-client pairs. We employ big data methods to develop a new method based on discrete choice models with reduced choice sets, the Randomly Reduced Choice set with Bayesian Inference (RRCBI) to analyze the dynamics of vendor selection in the market for IT services. We employ a dataset that consists of 22039 unique contracts between 8072 client firms and 2569 vendors over a 20-year period. We find that clients seek deep relationships through continuing interactions with vendors as well as vendors with diversity and depth of experience in terms of their knowledge of market conditions, customer-facing interactions, and organizational processes. We also find evidence for preferential attachment and multi-connectivity based logic of exchange. Our study is the first to investigate the impact of network structure on the client’s vendor selection for IT outsourcing using a longitudinal dataset. Our results strongly suggest that a network-based logic of exchange provides insight into the systematic pattern in which vendors build their competitive position and market segmentation strategies and the manner in which clients build their outsourcing portfolios. Implications for researchers and practitioners are discussed.