Hotel Discovery Has Moved Upstream.
The 2026 Data Confirms It.
For more than a decade, a major share of hotel direct-booking strategy rested on a simple assumption: if hotels could intercept travelers through search engines before OTAs did, they could reduce intermediary dependence and reclaim the guest relationship.
The 2026 data suggests that assumption needs to be revisited.
In November 2025, SiteMinder published its Changing Traveller Report 2026, based on surveys of 12,000 travelers across 14 countries. The report’s core finding on discovery behavior shows a sharp year-over-year shift in how travelers say they begin hotel research.
For the first time in the report’s history, OTAs have overtaken search engines as the primary starting point for hotel discovery. 26% of travelers now begin their hotel research on an OTA, up from 18% the year before. Those beginning with a search engine have fallen to 21%, down from 36%.
That is a 15-point year-over-year shift in reported starting points. Based on self-reported discovery behavior, not observed booking-path data, it is directional rather than absolute. But the direction is unambiguous.
Where the 15 Points Went
The shift away from search engines did not consolidate into a single replacement. The 15 points fragmented across several channels.
OTAs captured the largest share, rising 8 points to 26%. Word-of-mouth recommendations from friends and family doubled, from 7% to 14%. Familiar hotel brands grew from 3% to 7%. AI use as a declared discovery starting point reached 4%, up from 1% the year before.
A luxury hotel CMO reading those numbers might conclude that word-of-mouth and brand recognition are the real winners here. Eleven of the 15 points that left search engines went to channels that appear to favor hotels directly: personal recommendations and brand engagement. That reading is understandable. It is also misleading for independent luxury operators specifically.
The growth in word-of-mouth and brand direct accrues disproportionately to hotel companies with global loyalty infrastructure, the brands whose names travelers already carry in memory before they start planning. An independent luxury property without a loyalty ecosystem does not capture the brand direct gain. It competes for the word-of-mouth gain against organizations with billions in marketing investment behind them.
What appears as channel diversification at the aggregate level functions as competitive concentration at the property level. The independents are not gaining ground in the channels that grew. They are losing ground in the one channel, search, that had previously given them equal footing with larger competitors.
The Declared Starting Point Understates AI’s Role
Four percent as a declared starting point for AI looks like a rounding error. It is not.
AI as a declared starting point quadrupled year-over-year in SiteMinder’s data. It is growing faster among Gen Z and Millennials than in any other demographic. And the declared starting point significantly understates AI’s actual influence on which hotels enter a traveler’s consideration before any channel is opened.
This gap between declared starting point and actual influence is what makes the aggregate channel data incomplete. Two additional datasets published in the same window illustrate the scale of the discrepancy.
Operto’s 2026 hospitality data, reported by Hotel Dive, found that 40% of travelers now use AI tools at some stage of trip planning and booking. Similarweb’s 2026 Generative AI Brand Visibility Index, measuring US consumer behavior broadly, found that 35% of consumers now use AI at the product discovery stage, compared to 13.6% who use traditional search as their discovery mechanism.
These figures are not in conflict with SiteMinder’s 4%. They also are not directly comparable on a one-to-one basis, because each dataset defines discovery behavior differently and measures different populations. Taken together, they indicate direction and sequence rather than precise channel share: AI is operating upstream of the channels travelers identify as their starting points, shaping consideration before formal research begins.
This is the analytical construct the data points toward: a Pre-Search Consideration Layer, the stage at which AI systems shape which hotels enter a traveler’s mental shortlist before the traveler opens any declared channel. This is not an established industry term; it is an interpretation of what the combined data describes. But the phenomenon it names is measurable in the gap between SiteMinder’s 4% declared AI use and Operto’s 40% AI-influenced journey figure.
A hotel excluded from that layer is not losing to a competitor who ranked higher on the same results page. It is significantly less likely to encounter those travelers downstream at all.
In the search era, hotels competed to be found. In the AI era, they must compete to be included before search begins.
The Pre-Search Consideration Layer
The Pre-Search Consideration Layer is the first of three stages in the current discovery funnel:
Stage 1: Pre-Search: AI systems, recommendation engines, brand memory, and social proof shape the traveler’s consideration set. No search bar has been opened. No OTA has been loaded. The shortlist is already forming. For luxury properties, brand memory and trusted advisor networks form part of this layer, but must now compete with algorithmic recommendation engines for the same mental real estate.
Stage 2: Declared Channel: The traveler opens a search engine, OTA, or brand site to begin what they experience as their research. This is the layer most hotel marketing investment has historically targeted.
Stage 3: Transaction: Booking completes, either on the declared channel or via a direct switch to the hotel’s own site.
The strategic problem for independent luxury hotels is that the competition for Stage 2 and Stage 3 has been intensifying for a decade while Stage 1 was not yet commercially visible. It is now. Similarweb’s US consumer data indicates that AI tools are now used at the product discovery stage at a higher share than traditional search, a signal of Stage 1’s scale even though the measurement is not hospitality-specific.
The Search Channel Was Shrinking From a Second Direction
While the declared starting point data shows search engines losing ground to OTAs, search was simultaneously contracting as a referral mechanism from a different angle.
Similarweb’s research documents that zero-click searches on Google grew from 56% to 69% in the 12 months following Google’s AI Overviews rollout. Nearly seven in ten Google searches now end without a single click to any website.
Separately, Similarweb’s data shows AI platform visits grew 28.6% between January 2025 and January 2026. Referrals from AI platforms to external sites over the same period remained essentially flat despite the growth in usage. The platforms are oriented toward answering queries within the interface rather than routing users to other destinations.
Gartner projected a 25% decline in traditional search engine traffic across industries by 2026 as generative AI absorbed user intent. SiteMinder’s hospitality-specific data, a 15-point absolute decline in search as a declared starting point, suggests hospitality is running ahead of the broader macro trend. Travel is a high-consideration, multi-variable research category, well-suited to AI-assisted synthesis because travelers often need to compare destination, property type, price, availability, amenities, reviews, and timing before choosing where to book.
The OTA Position in the New Funnel
In late 2025, Booking.com was included among ChatGPT’s initial app integrations, placing OTA hotel-search functionality inside the conversational AI environment. The implication is not that every reservation now completes inside the chat interface. It is that OTA inventory, pricing, and transaction infrastructure are being positioned at Stage 1 of the new discovery funnel.
The intermediary that already governs a large share of independent hotel booking economics has moved into the AI layer before many hotels have fully recognized that layer as a competitive arena. Hotels that were downstream of OTAs in the search era are downstream of OTAs in the AI era. The structural position is largely unchanged. The environment governing that position has shifted entirely.
What the Generational Data Means for the Trajectory
Deloitte’s 2026 Travel Industry Outlook identifies generative AI adoption as one of four primary forces reshaping how guests discover, evaluate, and book hotels, with Gen Z and Millennial travelers leading the behavioral shift.
SiteMinder’s data confirms the generational concentration. AI as a declared starting point is rising fastest among Gen Z and Millennials, the traveler cohorts with the longest remaining booking horizon.
TakeUp AI’s April 2026 survey of 300 US leisure travelers found that 90% are now aware AI can help plan or book travel, while only 38% have actively used it for trip planning. The 52-point gap between awareness and active adoption is characteristic of early-stage behavior change, not mature penetration. Whether adoption accelerates evenly across traveler segments and income levels remains uncertain. What is clear is that the cohorts leading current adoption are the same cohorts with the most forward booking volume.
What This Data Does Not Say
This data does not say search engines are dead or that direct booking strategies built around search have failed entirely.
SiteMinder’s 21% of travelers who still begin with search engines represent a commercially meaningful segment. The data shows significant variation by market: Italian travelers maintain the strongest search engine orientation globally in SiteMinder’s data, and Australian guests begin with search at 31%. For certain property types, geographies, and traveler demographics, search engine visibility remains relevant. Luxury segments likely exhibit stronger brand and word-of-mouth resilience, but the data suggests Stage 1 AI shaping determines whether that brand strength is activated or bypassed.
This data also does not say the shift is complete or permanent in its current form. Discovery behavior will continue to evolve as AI platforms change their architectures and booking integrations expand.
What the 2026 data establishes is narrower than the headlines suggest and more consequential than the channel percentages imply. The primary starting point for hotel discovery has shifted away from search engines for the first time in SiteMinder’s longitudinal dataset. AI is shaping consideration sets at a stage that precedes the declared starting point. The major OTA platforms have positioned themselves inside the AI layer. And the traveler cohorts with the longest booking horizon are leading the adoption.
The Implication for Independent Luxury Hotels
Much of the direct booking playbook still used by independent luxury hotels was designed for a discovery environment in which search engines were primary and OTAs were the competition at Stage 2. That environment is no longer current.
The practical consequence of the Stage 1 shift is not that hotels should abandon Stage 2 and Stage 3 investment. It is that the outcome of that investment is increasingly determined by what happens at Stage 1, before the traveler has opened a search bar, loaded an OTA, or made any declared move. A property that is not present in the AI consideration set forming at Stage 1 is significantly less likely to encounter those travelers further downstream.
The strategic priority the 2026 data points toward is not a new marketing tactic. It is a question of where in the discovery funnel independent luxury hotels are allocating their attention, and whether that allocation is aligned with where the decision is now being made.
