Forecasting h(m)otel guest nights in New Zealand

Abstract

The purpose of this paper is to highlight some time series models which hotel and motel industry practitioners could use to forecast guest nights. Given their considerable practicality, the lodging industry can easily benefit from using these models as forecasts can be obtained at low cost for effective management and planning. Monthly observations are used for estimating the model from 1997(1) to 2006(12). The Holt–Winters and Box–Jenkins ARMA models are able to forecast guest night demand accurately as 99% of the variations in the guest night forecast are associated with variations in actual guest nights in 2007.