Data-driven class closure method for airline revenue management

Abstract

This article describes a practical data-driven approach to increase revenue, identifying inelastic and elastic demand segments used to decide when to close classes. The method proposed defines class closure policies that will close the lower fares a specific number of days before the departure. By identifying the remaining inelastic and elastic demand, choice can be made when to close lower classes. This approach maximizes revenue from inelastic demand, when customers are forced to book at a higher fare, to exceed the loss of revenue from the rejected elastic demand. These class closure polices are defined at the detailed level of: (i) Origin/Destination, (ii) Point-of-sale, (iii) Date, (iv) Departure time, (v) Number of days before departure. This process reflects the function analysts perform manually, or via rules as determined in the revenue management system. Closer to departure, and as the share of inelastic demand increases, lower classes will close to create sell-up by inelastic demand. Generally analysts work by rule of thumb or based on ad hoc data analysis. The purpose of this article is to use data to define the class closure settings at a more detailed level than could be expected from a manual approach.