Last-minute bookings are up, compressing lead times and making traditional forecasting less reliable. Revenue management teams must pivot to real-time, multi-source demand and pricing data – including OTA, brand.com, GDS, and metasearch feeds – while factoring in event-driven surges.
NB: This is an article from Aggregate Intelligence, one of our Expert Partners
Subscribe to our weekly newsletter and stay up to date
The Rise of Last-Minute Bookings
Travel behavior is changing – and not in a way that leaves revenue managers much breathing room. According to recent data highlighted by Skift, last-minute bookings have surged, surpassing the long-term average by roughly 5% in March 2025. By June, the U.S. market saw vacation rentals jump 10% year-over-year, while hotels scraped by with a 1% increase. The shift is clear: travelers are holding off until the final days – or even hours – before departure.
For revenue management teams, this isn’t just a quirky consumer trend. It’s a structural shift in demand patterns that changes how pricing data must be collected, interpreted, and acted upon.
1. Shorter Booking Windows = Less Reliable Forecasts
Traditional forecasting models rely heavily on historical booking curves – steady demand ramps leading up to check-in. When those curves compress, forecasting accuracy suffers. Instead of weeks of gradual demand buildup, teams may now see a flood of bookings inside a 72-hour window, forcing faster, more aggressive pricing decisions.
The data challenge here is granularity: to understand and respond to sudden spikes, revenue systems need high-frequency, real-time inputs – not just daily snapshots.
2. Events as Demand Catalysts
Major events – concerts, sports championships, festivals, conventions – are becoming even more impactful in a last-minute booking world. When travelers decide on impulse to attend, bookings can spike sharply just days before arrival.
Revenue managers must integrate event calendars and predictive event impact models into their pricing logic. Detecting an upcoming high-demand event five days out could mean the difference between selling out at optimal rates and leaving revenue on the table.
When paired with historical event data, these models can predict both the timing and magnitude of a booking surge, enabling more precise rate adjustments.
3. Price Sensitivity Is Shifting
In yield management theory, booking lead time is a key segmentation variable – guests who book late are often less price-sensitive, making higher rates viable. But as late booking becomes mainstream, this assumption weakens. More price-sensitive guests are now part of the last-minute crowd, meaning old elasticity models may misprice inventory.
Revenue managers will need to re-test these assumptions, using streaming booking data from multiple sources and competitor rate intelligence to pinpoint the real willingness-to-pay for each segment.
