Demand forecasting in hospitality using smoothed demand curves

Demand forecasting is one of the fundamental components of a successful revenue management system. This paper provides a new model, which is inspired by cubic smoothing splines, resulting in smooth demand curves per rate class over time until the check-in date.This model makes a trade-off between the forecasting error and the smoothness of the fit, and is therefore able to capture natural guest behavior. The model is tested on hospitality data. We also implemented an optimization module, and computed the expected improvement using our forecast and the optimal pricing policy. Using data of four properties from a major hotel chain, between 2.9 and 10.2% more revenue is obtained than using the heuristic pricing done by the hotels.