Customer behavior in revenue management

This thesis examines several topics incorporating customer choice behavior, customer strategic behavior and switching costs effects into revenue management.; Motivated by Gallego et al. [32]'s choice-based deterministic linear programming model proposed recently, we extend the analysis of this model in the first essay. Using the notion of "efficiency" developed by Talluri and van Ryzin [65], we show that only efficient sets are used in an optimal policy asymptotically. We develop a practical decomposition heuristic to dynamically generate control policies. Numerical examples show this heuristic significantly improves the performance of the static linear programming.; Two more essays address strategic purchase behavior when customers face different prices over time. We develop a stylized capacity rationing model, in which a firm can deliberately create shortages by under-stocking to induce early purchases at high prices. We show that a large number of high-value customers, risk aversion, large price differences over time and a low level of competition all tend to make rationing an optimal strategy.; Yet customers may not have fully rational expectations; instead, they may learn only through repeated experience. The second essay on this topic considers how a firm should use its capacity decision to profitably influence the customer learning process. We develop an adaptive learning model in which customers update expectations adaptively, and the firm's decision problem is imbedded in a dynamic program of capacity choices over time. We show that the market converges to either a rationing equilibrium or a low-price-only equilibrium. Indeed, a critical value of customer expectation determines which equilibrium is the long-run optimal outcome. We also show this equilibrium is the same as the rational expectation outcome when future profits are not discounted.; Finally, we examine switching costs effects on tactical pricing decisions, which are especially important in subscription and account-based services. There is a tradeoff between increasing revenue by charging current customers a high price versus pricing low to build for the future market share. We formulate an optimal pricing control model and show the optimal strategy is a simple "target market share" policy. We also extend the analysis to a duopoly market.