Booking Holdings is spending more on marketing per dollar of travel booked lately — what's driving that extra cost?

This explores why Booking might be paying more and more to acquire each booking, possibly because AI agents are changing how travelers find hotels and flights. One caveat first: the corpus has no note that reports or explains Booking's marketing expense ratio. It can't confirm that the gap exists or how big it is. What it does offer is a set of mechanisms that would push customer-acquisition costs up for any online travel platform right now, and those are worth knowing.

The first mechanism is that the old way of buying attention is losing value. In the corpus's argument, shopping and travel agents now do the comparing and ranking themselves, based on what the traveler says they want rather than on paid placement. That makes attention-based ads less effective for aggregators Are AI agents making attention-based ads irrelevant for aggregators?. Tripadvisor's Hotels revenue fell 23% in one quarter, and management blamed structural shifts in search. If organic and SEO-driven traffic is eroding, a platform like Booking has to buy back that demand through paid channels at higher prices, or pay for placement inside the agents themselves. Either way, marketing spend grows faster than the bookings it brings in.

The second mechanism is a new middleman taking a cut. One model has agents moving from tiny payment commissions toward affiliate fees and eventually aggregator-style take rates of around 12% in some categories How could an AI agent's revenue per transaction grow over time?. That 12% would be revenue for the agent and a cost for whoever supplies the inventory. A related formal model shows how a platform that controls discovery can price access so that surplus shifts away from sellers toward itself Does platform pricing of inspection tokens encourage excess search?. Infrastructure is shifting too: Amadeus now caches precomputed fares because agent searches are approaching 200,000 per ticket sold. That gives whoever controls the cache influence over which offers agents see Does caching fares give infrastructure providers pricing power?. So the gatekeepers to demand are multiplying, and each one can charge rent.

The other side of the story is that this doesn't mean OTAs are dying. Their stocks rose in 2026 despite warnings that agents would cut them out, which suggests agents make discovery cheap but leave payment, supply, and trust with the OTAs Are AI agents actually disintermediating online travel platforms?. Independent hotels' reliance on OTAs actually rose to 63.4%, and Expedia's business-to-business segment, which sells its booking infrastructure to other companies, grew 24% Are AI agents actually removing online travel agencies from bookings?. Put these notes together and a likely reading emerges: a rising marketing ratio may be the cost of a transition, not a sign of decline. OTAs pay more to stay visible at the discovery layer while their real defensible value moves to the settlement layer behind the agent.

The idea you may not have expected is that a rising marketing ratio and a strong stock price may not contradict each other. They can be two sides of the same shift: discovery is getting more expensive to buy while fulfillment is becoming harder to replace. To check whether that holds for Booking specifically, you would need its earnings disclosures, which this corpus doesn't contain.