Ancillary revenues in the airline industry : impacts on revenue management and distribution systems

Abstract

Airlines have increasingly depended on ancillary revenue in response to rising fuel costs, de- creased yields, and an increasingly competitive environment. Estimates indicate that U.S. airlines collected over $8 billion in ancillary revenue in 2012. Ancillary revenue poses challenges for airlines, including revenue management (RM) and distribution since total revenue maximization requires consideration of ancillary revenue and ticket revenue. In this thesis, we: (1) describe trends contributing to the movement towards ancillary revenue; (2) present three methods for incorporating ancillary revenue into revenue management and distribution; (3) evaluate the revenue performance of these methods using the Passenger Origin Destination Simulator (PODS), a competitive airline simulator. One method of including ancillary revenue into RM is RM Input Adjustment with Class Level Estimates, which involves modifying input fares to the optimizer. Because fare values to the optimizer are aggregated by market and class, the airline uses class level estimates of ancillary revenue potential to augment fares. Another method involves modifying the fare value at the time of availability control, or Availability Fare Adjustment. In network optimization, the availability fare refers to the fare used to compare an itinerary-class to the control mechanism, like displacement adjusted virtual nesting (DAVN) or additive bid price (ProBP). Availability Fare Adjustment with Class Level Estimates also involves using class level estimates of ancillary revenue. Alternatively, we test scenarios where the airline estimates ancillary revenue for individual passengers in Customized Availability Fare Adjustment with Passenger Specific Estimates. Although this type of estimation is not feasible yet, results from Customized Availability Adjustment give a theoretical bound to revenue gain. We nd that incorporating ancillary revenue opens availability for lower yield passengers. Revenue increases occur from extra bookings in these classes because more bookings are taken. Revenue losses occur from higher class passengers buying down to cheaper seats. Without willingness to pay (WTP) forecasting, net revenue losses of up to {2.6% are observed. In advanced RM systems with WTP forecasting, revenue gains of +0.6% are observed for Class Level RM Input Adjustment, +0.9% for Class Level Availability Fare Adjustment, and +2.6% for Passenger Specific Customized Availability Adjustment.