Dynamic offer creation for airline ancillaries using a Markov chain choice model

Abstract

Abstract Customers have become accustomed to a highly streamlined and personalized experience when shopping online. While tech giants such as Apple, Amazon, and Netflix are experts in using customer information and shopping context to deliver relevant offers, airlines are falling behind in this regard with their static content and one-size-fits-all retailing approach. To meet the growing expectations of their customers, the airline industry has expressed a vision for dynamic offer creation, which will allow airlines to dynamically bundle and price a set of offers that is customized to the context of the shopping request. Realizing this vision requires significant advancements in both distribution and science. On the distribution side, these advancements will come with the adoption of the New Distribution Capability. On the science side, which is the focus of this paper, little progress has been made despite years of research. In particular, airlines still lack a tractable scientific model to dynamically create and price offer sets at scale. In this paper, we present a novel approach to solve the airline dynamic offer creation problem using a Markov chain choice model. Our model displays attractive qualitative properties—the resulting offers and prices are chosen in such a way as to discourage purchases of unprofitable offers and nudge customers towards more profitable ones. Our model naturally proposes offers that are relevant to the customer, as including irrelevant offers in the offer set leads to a reduction in revenue and ancillary purchase rates. In a simulation study with two customer segments, we find that our model significantly increases ancillary revenue over a naïve, unsegmented pricing model that mimics current state-of-the-art practice. While our studies are conducted under several idealized assumptions, they demonstrate a substantial revenue potential from dynamic offer creation in both unsegmented and segmented applications.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 10 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.