When history fails you, borrow from geography

Summary

When historical travel data became uninformative during the asynchronous post-COVID recovery, Airbnb redesigned its demand forecasting to borrow signals across geographies rather than waiting for local data. Using a hierarchical Bayesian framework, the team propagated updated posterior estimates from early-recovering corridors as informative priors for structurally similar, later-recovering corridors. This allowed accurate corridor-level demand forecasting in near real time during periods of high disruption and uneven market reopening.

Takeaways

1. 1

   Posterior demand parameters from early-affected geographic corridors can serve as informative priors for structurally similar regions before local shock data accumulates.

2. 2

   Information transfer between market corridors should be similarity-weighted based on structural attributes like traveler composition, domestic-versus-international mix, and accommodation types.

3. 3

   The geographic prior propagation framework generalizes beyond crisis recovery to sequential product rollouts, regional policy changes, and macroeconomic shocks.