Airlines have traditionally segmented digital audiences with business rules. If clients fulfill certain criteria or act in a given way, they would be targeted as part of a segment, offered vouchers or an incentive to book. When this occurs at a segment level there is inevitable wastage. The benefits of any uplift are therefore diminished if the number of vouchers results in unnecessary dilution. But measuring this is hard!
Segments are made up of a wide variety of individuals with different motivations. There is an understandable fear among Revenue Management Executives that
if discounts are made available to adhere to a set of rules, these can be learned and “gamed” by customers to trigger unnecessary incentives.
So the challenge was set: is there a way to use individual profiling and machine learning to target incentives more efficiently, achieving the goals of e-commerce (higher conversion) while avoiding revenue dilution.