Better unconstraining of airline demand data in revenue management systems for improved forecast accuracy and greater revenues

Abstract

Accurate forecasts of passenger demand are the heart of a successful revenue management system. The forecasts are usually based on historical booking data. These bookings do not reflect historical demand in all cases because booking requests can be rejected due to capacity constraints or booking control limits. This paper examines six different methods of unconstraining bookings to demand. Simulation analysis of many different scenarios of historical booking data are used with different percentages of the data being constrained, using simulated data, to show that the expectation maximisation and projection detruncation methods are the most robust and that, as the percentage of data constrained increases to 60–80 per cent, their estimate of the unconstrained mean increases by 20–80 per cent over the naïve unconstraining methods, which leads to less bias and more accurate forecasts. Finally, by means of actual booking data from a major US airline, it is shown that upgrading the unconstraining process can lead to revenue gains of 2–12 per cent.