Extending Online Travel Agency with Adaptive Reservations

Current online ticket booking systems either do not allow customers to reserve a ticket with a locked price, or grant a fixed reservation timespan, typically 24 hours. The former often leads to false availability: when a customer decides to purchase a ticket after a few queries, she finds that either the ticket is no longer available or the price has hiked up. The latter, on the other hand, may result in unnecessary holdback: a customer cannot purchase a ticket because someone else is holding it, who then cancels the reservation after an excessively long period of time. False availability and holdback routinely lead to loss of revenues, credibility and above all, customers. To rectify these problems, this paper introduces a transaction model for e-ticket systems to support a reservation functionality: customers can reserve tickets with a locked price, for a timespan that is determined by the demands on the tickets, rather than being fixed for all kinds of the tickets. We propose a method for implementing the model, based on hypothetical queries and triggers. We also show how to adjust the reservation timespan w.r.t. demands. We experimentally verify that our model and methods effectively reduce both false availability and holdback rates. These yield a practical approach to improving not only e-ticket systems but also other e-commerce systems.