Date of Award
Doctor of Philosophy (PhD)
The rapid development of information technology has changed how firms interact with their customers. On one hand, firms are better capable of collecting customer data, and equip themselves with more powerful analytical tools. On the other hand, customers are becoming more sophisticated in their purchase decision making and other non-purchase interactions, which create higher demand uncertainty for the firm. To survive in this complex and dynamic environment, firms need to manage their customer relationships with an integrated and proactive approach. Recent studies in adaptive learning helped the firm to answer the question of How to learn about customers so they can be proactive in their CRM practice. In this study, we introduce the concept of Double Loop Learning, where we added a strategic learning loop to the adaptive learning loop. With this double loop structure, we also answer the questions of Why to learn and What to learn and Who should be learn simultaneously in an integrated framework. We use a Partially Observable Markov Decision Process (POMDP) approach to 1). Generate optimal marketing contact policy which balances exploration (learning how various modes of marketing contacts affect the transition of customer relationship state) and exploitation (maximizing short-term profit), and 2). Assess the Value of Learning (VOL) at individual customer level to give a feedback to the strategic learning loop where we can answer the questions of Why, What to learn at individual customer level. Theoretically, we introduced the concept of Double Loop Learning to marketing literature which is fundamental in that it achieves both effectiveness and efficiency in the marketing strategy development. Methodologically, we adopted a POMDP approach which enables us to access the value of information for connecting two loops in an integrated framework. In the first essay, we did extensive review on the CRM and Adaptive Learning literature, based on which we developed the conceptual framework for Double Loop Learning model. We also developed an analytical model to demonstrate the relationship between the VOL and Dynamic Customer Value (DCV) of the customers. In the second essay, we apply the proposed framework to an IT B2B firm. We show that the firm can achieve value gains by managing VOL and DCV simultaneously.
Fan, Jia, "A Double Loop Learning Model For Integrated and Proactive Customer Relationship Management." Dissertation, Georgia State University, 2015.