Modeling strategic customers using simulations - with examples from airline revenue management

Abstract

A condition for airline revenue management is the possibility of identifying and differentiating customer segments (refer to Chiang et al. (2007) for a state of the art). Traditionally, customer differentiation has been realized by the time of request in days before departure as well as by restrictions connected to the tickets sold. Customer segments have been regarded to be fixed over time, based on myopic customer behavior. With the market transparency increased through the Internet as well as the rise of no-frills offers and flat-rates, customer behavior has changed during the last decades. Strategic customer behavior describes a tendency to remember previous buying experiences, adapt expectations and observe the market over longer periods of time before deciding on what (and whether) to buy. The empirical consequences of strategic customer behavior for traditional as well as state-of-the-art revenue management have been little examined. A major reason for this is that measuring the degree of strategic versus myopic tendencies of demand in real customers is difficult and expensive. In this paper, we formulate a mathematical model of strategic customer behavior including parameters defining the propensity to delay buying as well as learning and communicating. We test the empirical consequences of our model using a stochastic simulation, in which customers act as agents deciding whether and when to buy. Thereby, we provide first results on how different ways of strategic behavior affect the success of methods of revenue management, highlighting the possible weaknesses and strengths of approaches when confronted with strategic customers. Of course, strategic customer behavior is not limited to airline revenue management. Useful models of strategic behavior as implemented in the simulation can be applied to analyze a wide range of situations: conditions for this transfer as well as examples are provided as part of the outlook.