Learning and Acting on Customer Information: A Simulation-Based Demonstration on Service Allocations with Offshore Centers
利用DSL服务客户通话数据，研究客户对国内和离岸服务的反应如何影响服务时长和保留率，通过动态规划学习客户偏好并优化分配，模拟显示学习能提升定制化分配和长期利润。
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 experiences affect 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 through simulations that learning enables a firm to make more customized allocations and that acting on long-term customer responses prompts the firm to make proactive decisions that prevent customers from leaving. As a result, the firm can improve customer retention and 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.