Dynamic Pricing Effect on Hotel Revenue

In this paper, we study the dynamic pricing effect on hotel revenues using the example of the “Eva Hotel” located in Perm. We employ the panel data on bookings from August 2018 to January 2021, which consists of 8 000 observations. As predictors of demand, we used the hotel room characteristics, in particular, the room number and its real price, and the arrival date characteristics, such as the day of the week, month, possible holidays, events and some information about COVID-19. Estimates of the Cox proportional hazards model demonstrate a positive price effect, which contradicts the basic law of demand. The inclusion of an instrumental variable helps to fight against the price endogeneity, therefore, it qualitatively improves the demand model and its forecasts. Based on this model, the problem of price optimization is solved using a self-programmed algorithm implemented in the R language. We test two scenarios with narrower and wider price ranges, both of which demonstrate the positive dynamic pricing impact on hotel revenue and better performance compared to fixed pricing. Due to dynamic pricing, the potential revenue increase can be from 57% to 70%, respectively, relative to the revenue of 2020. Regardless of the scenario choice, the optimal strategy for the hotel under study is a significant decrease in prices relative to the base ones, provided that the room has not been booked yet, and then their increase closer to the check-in date, but still below the current price level.