Pricing Strategies in the Presence of Interventions

Primary author: Sadegh Kazemi
Faculty sponsor: Stergios B. Fotopoulos

Primary college/unit: Carson College of Business
Campus: Pullman


Motivated by real-life pricing practices, we consider a pricing problem under uncertain conditions where the customer’s willingness-to-pay (WtP) changes at an unknown point over the selling horizon. An important feature of our model is that the seller only observes the sales outcomes and has limited knowledge of the underlying WtP distribution before and after the intervention. Given the uncertainty associated with the seller’s estimate of the time of change, we obtain the probability distribution of the seller’s maximum likelihood estimate (MLE) of the intervention time. Furthermore, we characterize the seller’s expected revenue loss due to the under- or overestimation of the intervention time and propose an easily implementable procedure to approximate the seller’s revenue under-performance. Our study reveals two important findings. First, the seller tends to underestimate the intervention time in the face of negative events that lower the customer’s reservation price. Conversely, the seller is prone to overestimation when the intervention inflates the customer’s reservation price. Second, we show that the seller’s revenue under-performance is minimal both when the shift in WtP distribution parameter(s) is either very small or considerably large. While our analytical results significantly contribute to the revenue management literature on their own, we also provide an accurate numerical method to easily obtain and interpret the results in a meaningful way for managerial use.