Learning and Acting Upon Customer Information: An Empirical Application to Service Allocations with Offshore Centers Bao-Hong Sun Shibo Li 10.1184/R1/6706493.v1 https://kilthub.cmu.edu/articles/journal_contribution/Learning_and_Acting_Upon_Customer_Information_An_Empirical_Application_to_Service_Allocations_with_Offshore_Centers/6706493 As service centers become crucial corporate assets for increasing customer relationships and profits, it is imperative to understand customer reactions to service allocations. Using customer call history from a DSL service, the authors empirically investigate how customers’ onshore and offshore experience affects service duration and customer retention. They formulate service channel allocation decisions as solutions to a dynamic programming problem in which the firm learns about heterogeneous customer preferences, balances short-term service costs with long-term customer retention, and optimally matches customers with their preferred centers to maximize long-term profit. They demonstrate that learning enables a firm to make more customized allocations, and acting on long-term customer responses prompts the firm to make proactive decisions that prevent customers from leaving. As a result, the derived allocation decisions (1) reduce service costs, (2) improve customer retention, and (3) enhance profit. The proposed framework also mirrors the recent trend of companies seeking solutions that transform customer information into customized and dynamic marketing decisions to improve long-term profit. 2008-05-01 00:00:00 service outsourcing; channel allocation; customer retention; long-term customer value; customer information management; decision support system; customer-centric marketing; stochastic dynamic optimization