Airline ancillary revenue continues to grow strongly (in fact reaching about USD $80 billion worldwide in 2017). To further improve on this fast-paced growth in ancillary sales, airlines are becoming more interested in a data-driven, science-based approach to airline offer management. Many airlines have asked for the ability to deliver data-driven dynamic offers with 1:1 personalization on their websites; furthermore, they want to extend these capabilities to global distribution channels using IATA new distribution capability standards. Unfortunately, many of the commercially available vendor products in airline offer management incorporate rudimentary, rules-based technology that is often deemed insufficient.
Building on recent AI research work on trip-purpose based customer segmentation models for air travel, Sabre designed additional models and built a working prototype that generates dynamic, bundled ancillary discount offers for the airline direct channel. To maximize sales conversion rates and revenue, both the ancillary bundle content and pricing decisions are dynamically adapted to the customer trip-purpose segment in our framework.
A unique feature of our prototype is that it generates dynamic bundles for anonymous shoppers (an important feature given that most air shoppers are not known) using customer segmentation based on trip-purpose type. Such segmentation allows market basket analysis to determine the best set of ancillary bundle items which are the most relevant to the segment considered. It can also handle intelligent, customer-specific recommendations which automatically blend together the trip-purpose segment and customer-specific history in determining ancillary preferences. These blended preferences are essential because the travel purpose for a specific future trip may not necessarily match past history.
The bundled offers are based on data-driven, ancillary preferences by trip-purpose segment (and customer history if known). Both the content and pricing are dynamic and take into consideration branded fare purchases, loyalty programs and credit cards affiliations when constructing the bundled offers. The bundled offer generation process uses a unique, adaptive algorithm which automatically learns how to improve bundles over time to maximize conversion rates.