Overbooking practices in the hotel industry and their impact on hotels' financial performance

Abstract

Hotel overbooking occurs when the number of rooms available for reservation exceeds the capacity. Hotels overbook with the goal of maximizing their revenue and improving their profitability. Despite its prevalence, many aspects of the hotel overbooking have never been researched. This study provides a clear picture of the current state of overbooking in the US hotel industry and explores the relationship between overbooking practices, cancellation policies, data availability and financial performance. Two data sets were used to answer the research questions. For the first data set, a group of data collectors recorded the cancellation policies of nearly 600 US hotels by manually checking their websites and going through the reservation process. For the second data set, a survey was distributed among a random sample of 10,000 US hotels asking them about different aspects of their overbooking policies. A survey response rate of 3.77% was achieved. Following data cleaning, the overbooking data set contained 365 hotels while the cancellation policies data set contained 492 hotels. After anonymizing the hotels, their performance indicators were added to the data sets. Analysis of Variance (ANOVA), independent samples T Test, Kruskal-Wallis test, Mann-Whitney test, Spearman correlation, stepwise multiple linear regression and multivariate multiple regression were the statistical methods used in this study. Results indicated that overbooking (vs. not overbooking) results in better hotel performance. Among the four major overbooking approaches (i.e., deterministic, risk-based, service-level and hybrid), findings indicated that risk-based overbooking results in the highest RevPAR index values. It was also found that keeping overbooking limit at minimum (i.e., less than 5% of capacity) and overbooking frequency at moderate levels (i.e., 6-10 days in a month) results in the best performance, while excessive overbooking (i.e., more than 10% of capacity and/or more than 20 days in a month) could result in lower RevPAR index values. Data analysis revealed that neither data availability nor cancellation policy can moderate the relationship between the four major overbooking approaches (i.e., deterministic, risk-based, service-level and hybrid) and the RevPAR index. Finally, analysis of the cancellation policies data indicated that moderate cancellation policies are associated with better performance.