The hotel discovery layer: Booking channels have stopped doing the work

The decision has moved upstream — and most distribution strategy hasn't followed

Destination choice has always happened off-channel. The decision to go to Lisbon was never made on Booking.com — it formed earlier, in a friend's photo, a film, a magazine. That part of hotel distribution has always sat upstream of any booking channel.

What used to happen on the booking channel was the comparison. Travelers arrived knowing the destination and used Booking.com or Expedia or a brand site to do the work of choosing — scrolling listings, reading reviews, comparing rates, narrowing a shortlist. The channel was the comparison engine. The transaction was its byproduct.

That is the part that has changed.

Hotel distribution has always had two layers. A discovery layer above, where destination choice formed. A booking layer below, where comparison, selection, and transaction happened. For two decades the discovery layer was small and the booking layer was large. Most of the meaningful decision sat inside the channel.

In 2026, the proportions have inverted. The discovery layer has absorbed the comparison work. By the time a traveler reaches a booking channel, the property is largely chosen, the shortlist already formed, the rate expectation already set. The booking channel still records the transaction. It is doing far less of the deciding.

Most commercial strategy is still pointed at the layer that has shrunk.

What used to live on the booking channel

For two decades, channel-centric distribution thinking worked because the channel was the comparison engine. A traveler doing serious research on a Lisbon hotel was almost certainly doing it on Booking.com, on a metasearch site, on a brand website, or on Expedia. That is where the listings were, where the reviews lived, where the rates were comparable side by side, where photos and amenities and cancellation terms could be weighed against each other.

This is what made channel-mix optimization the right craft. Channel reports were the right instruments. Rate parity was the discipline that mattered, because parity errors were visible exactly at the moment of comparison — on the channel itself. Conversion was the metric, because most arrivals on the channel were undecided when they arrived.

None of that is wrong about the booking layer. What has changed is how much of the decision the booking layer still touches.

What moved upstream

Three forces have taken functions that used to live on the booking channel and pushed them above it.

The first is social platforms. Property shortlisting — not just destination choice — is now routinely formed inside TikTok, Instagram, and YouTube. A traveler watching a creator walk through a hotel is doing on the platform what they used to do scrolling Booking.com listings: deciding which properties are worth considering. By the time they reach a channel, two or three names are already in mind.

The second is AI search. Review reading was the booking channel's unglamorous workhorse — the friction that determined which property in a shortlist actually got chosen. AI summaries have compressed that function into a single answer. A recent hotelrank.ai study of 4,000 Google AI Mode hotel queries found that OTAs receive only 3.6% of clicks, despite providing 46% of the sources the model cites. The model is learning from OTAs while routing traffic elsewhere.

The third is the absorption of metasearch itself. Standalone metasearch — the original upstream comparison surface — is being compressed into AI-native answers. Booking Holdings took a $457 million writedown on Kayak in October 2025, attributed by its CEO to Google's AI Overviews squeezing the category. The comparison function has not disappeared. It has migrated to surfaces hotels do not control.

Naming the layer

The booking layer is the one commercial teams know well. Booking.com, Expedia, the GDS networks, the brand booking engines, the metasearch entry points still in operation. Its intermediaries are named, contractually engaged, and operationally familiar. Its metrics are mature: channel mix, conversion rate, cost of acquisition, contribution margin.

The discovery layer used to be small. It was where someone decided to go to Lisbon, and that was about it. In 2026 it is most of the funnel. It is where shortlisting happens, where reviews get summarized, where rate expectations form, where comparison takes place. Its intermediaries are different — Google, Meta, ByteDance, OpenAI, the creator economy that operates inside their platforms — and most hotels have no direct commercial relationship with any of them.

The two layers operate by different rules. The booking layer responds to commercial levers hotels are used to pulling: rate, availability, parity, contract terms, channel-specific promotions. The discovery layer responds to a different set of inputs entirely — content presence, citation visibility, model training data, creator partnerships, platform algorithm shifts. A property can have a flawless booking layer setup and remain invisible at the discovery layer. A property can have unmanaged rate parity and still get chosen at the discovery layer because of where it shows up in an AI summary.

The work commercial teams do on the booking channel matters only for travelers who already decided to consider that hotel. If the property never made the shortlist upstream, there is nothing on the channel left to optimize.

What this means

Two observations follow.

The first is that the booking layer has not become unimportant. It has become narrower. The craft of optimizing it is still real, but the share of the eventual outcome it can influence has shrunk. Commercial teams that have spent two decades getting good at the lower layer are operating with sharper tools on a smaller part of the problem.

The second is that hotels can measure the booking layer well and the discovery layer barely at all. Channel reports, parity scans, conversion analytics, rate shopper tools — these are mature instruments pointed at the booking layer. There is no equivalent instrumentation for the discovery layer at most properties. Hotels rarely know which AI summaries they appear in, which creator videos drive their consideration, or how their visibility in any of these surfaces is changing month over month. The blind spot is the layer that increasingly matters most.

What follows

The four pieces that follow examine each force shaping the discovery layer in turn — social platforms, AI search, the absorption of metasearch, and the early shape of agentic discovery. They share a single argument with this one: the comparison work has moved upstream, and the booking channel is no longer the layer where most of the decision is made.

The series does not end with a prescription. It ends with the harder question — what kind of commercial function is fit for a layer hotels do not own, cannot fully see, and increasingly depend on.

Next in this series: The shortlist forms before the search

by Markus Busch, Editor/Publisher Hospitality.today

Enjoying this analysis? Hospitality.today delivers daily insights on hotel distribution, AI trends, and travel commerce — straight to your inbox. Subscribe for free at Hospitality.today →