The dashboards look the same. The line items look the same. But the person who used to decide where a hotel’s ad budget went now shares that decision with a system that doesn’t explain itself — and three deadlines already on the calendar are about to test how much that matters.
Hotel advertising in 2026 runs on the same platforms it did five years ago — Google Ads, Meta, Google Hotel Ads — but the marketer buying that advertising has less direct say over where the money lands than at any point since paid search existed for hotels. This is not a piece about whether AI will eventually change travel planning, or another explainer of what Performance Max is. It is about five things that are already happening, on dates that are already fixed or have already passed: Google reversing a five-year privacy commitment the industry built around; a September 30 cutoff that will silently drop some hotels out of Google’s price-comparison box; automated campaign types that can put a franchisee’s own money against their own brand’s search terms; a free listing quietly outperforming paid placements on AI-driven search; and, as of this week, Google’s search bot booking hotel rooms directly. For a GM, DOSM, or revenue manager approving Q4 and 2027 budgets, each of these touches a specific cost line — and none of them wait for the next planning cycle.
Table of Contents
1. Google Kept the Cookies. The Reason Hotels Collected First-Party Data Anyway Still Holds
For five years, Google told the advertising industry to prepare for a Chrome without third-party cookies, first targeting 2022, then repeatedly pushing the deadline back. On April 22, 2025, Google’s VP of Privacy Sandbox, Anthony Chavez, announced that Chrome would keep supporting third-party cookies indefinitely and would not roll out the “user choice” prompt Google had proposed as a replacement. Six months later, on October 17, 2025, Google went further and retired most of the Privacy Sandbox APIs it had spent years building as cookie substitutes — Topics, Protected Audience, Attribution Reporting, and others — keeping only a handful of narrower tools. The infrastructure hotels were told to prepare for never fully arrived, and the alternative Google was building to replace it has now been largely shelved too.
For a hotel that spent the past few years investing in first-party data collection — CRM matching, loyalty capture, hashed email uploads to ad platforms — the reversal can look like wasted effort. It isn’t, for a reason that has nothing to do with Chrome’s policy and everything to do with how the automated campaign products actually work. Google’s own guidance on AI Max states plainly that manual bidding limits what the system can do: AI Max’s search-term matching and creative optimization depend on signals from automated bidding strategies, and a campaign run on manual cost-per-click gets a degraded version of the product. Performance Max works the same way. These systems are not targeting engines that a marketer configures — they are optimization engines that a marketer feeds, and the quality of what they’re fed now determines the cost of every booking they generate more directly than any targeting decision would have five years ago. A hotel with clean, matched, first-party conversion data — real guest value, not just “booking completed” — will see a materially different cost per acquisition than a competitor spending the same budget with poor CRM-to-ad-platform matching, holding everything else equal. The commercial consequence sits on the acquisition cost line, and it has moved from being a targeting question to a data-plumbing question.
Third-party cookies also remain a patchwork regardless of what Chrome does: Safari and Firefox have blocked them by default for years, accounting for a meaningful share of global browsing that was never going to be reachable through cookie-based retargeting no matter what Google decided. What is worth monitoring is less the cookie headline than the plumbing underneath it — since June 2026, Google Ads has consolidated cookie and identifier controls for its own tags into a single Consent Mode setting, and a misconfigured Consent Mode is now a quiet, common cause of degraded automated-bidding performance that shows up as an unexplained CPA increase rather than an obvious error.
2. A September 30 Deadline Most Hotel Teams Have Not Tested Yet
Google is retiring a feature it calls Google third-party rates for hotel ads. Some hotels advertising through Google Hotel Ads never had to send Google their own pricing — Google sourced the rate itself from a third-party provider on the advertiser’s behalf. After September 30, 2026, that sourcing stops. Campaigns depending on it will stop serving across Hotel Ads inventory entirely, including the booking module and Travel Promotion Ads, according to Google’s own Hotel Center documentation. Performance Max for Travel Goals campaigns will keep running on Search, Display, and YouTube, but will lose access to Google’s hotel-specific inventory. To keep appearing in Hotel Ads after that date, a property needs a direct price feed connected through its own Hotel Center account or an approved connectivity partner — and existing affected campaigns cannot simply be switched over; new ones have to be built and pass Google’s price-accuracy review before they serve.
The commercial consequence is direct booking share and, by extension, commission cost. A hotel that drops out of Google’s price-comparison surfaces does not disappear from Google search — but Booking.com, Expedia, and other OTAs run their own rate feeds and are unaffected by this change, so a hotel that loses its Hotel Ads presence loses visibility next to its own OTA listings at the exact moment a traveler is comparing prices. Rebuilding a compliant feed and campaign is not a same-day settings change; it involves technical setup through a channel manager, booking engine, or Hotel Center account, a price-accuracy review cycle, and campaign recreation with new bidding history. Teams that start this in September will be racing a review process, not flipping a switch.
What is worth checking now, plainly: whether a property’s current Google metasearch presence already runs through a connected Hotel Center feed — in which case this changes nothing — or whether it has been running on Google-sourced third-party rates without the team realizing it, which is common where metasearch has been managed by a generalist agency rather than a hotel-technology specialist. That audit takes an afternoon. Waiting until October to run it does not.
3. When the Algorithm Bids on Your Own Franchise’s Brand Name
Performance Max and AI Max both work by matching a hotel’s ads to a much wider set of search queries than manual keyword lists ever covered, using broad and keywordless matching plus automated asset generation across the property’s domain — with far less query-level visibility into what triggered which ad than a standard Search campaign provides. Google has added tools to rein this in, including brand lists that can exclude or restrict a hotel’s own branded terms inside Performance Max, but these have to be configured deliberately; they are not switched on by default. For an independently marketed property, an unconfigured brand list mostly wastes a modest amount of spend against its own name. For a flag-affiliated hotel, it is a different problem: a franchisee running its own Performance Max or AI Max campaign in parallel with the brand’s centrally managed co-op or brand.com program can end up bidding against the flag’s own branded search terms — or against a neighboring franchisee of the same flag — inflating cost-per-click on traffic that should have converted for close to nothing.
The commercial consequence extends beyond wasted media spend. Meta’s move toward fully automated creative generation in Advantage+ has already produced ad variations, in other retail categories, that diverge from a brand’s approved messaging and imagery without human sign-off. For a hotel operating under a franchise agreement, an automatically generated ad that misstates rate parity, uses off-brand imagery, or breaches a co-op advertising program’s terms is not simply an underperforming ad — it is a brand-standards exposure that sits alongside the property’s other compliance obligations to its flag, with a different owner and a different consequence than a marketing team missing its own CPA target.
What is worth watching: whether a property’s franchise and co-op marketing agreements — many drafted before Performance Max or AI Max existed — say anything at all about automated or AI-generated campaign activity, and whether local marketing teams have actually applied brand exclusion settings to keep their own automated campaigns from cannibalizing brand.com or co-op spend. In several hospitality-marketing circles this has started to be framed explicitly as a governance question — which layers of an account get full automation and which stay on tightly controlled exact-and-phrase-match search — rather than an all-or-nothing choice between manual control and Google’s default settings.
4. The Free Listing Is Now Outperforming the Ad Budget on Its Own Turf
Google’s AI Overviews and AI Mode are changing where clicks go before a hotel search even becomes a hotel search. Pew Research Center tracked the browsing behavior of 900 U.S. adults across 68,879 real Google searches in March 2025 and found that when an AI-generated summary appeared on the results page, users clicked a traditional link only 8% of the time, against 15% when no summary appeared — and only 1% of visits resulted in a click on a source cited inside the summary itself. That is not a hotel-specific finding, but it describes the pool of clicks that any paid or organic hotel listing is now competing to survive within.
Within hotel search specifically, an independent study by researcher Nicolas Sitter, analyzing 4,000 Google AI Mode queries across eight cities in February 2026, found that roughly 79% of clickable hotel links inside AI Mode responses route first to a Google Business Profile, about 17% go directly to the hotel’s own website, and only around 4% go to an OTA — despite OTAs supplying close to half of the underlying source citations the AI response draws on. This is a single, independently conducted study with a defined but moderate sample, not an industry-standard benchmark, and it should be read as a directional signal rather than a precise global figure. But the direction is consistent with what Google itself has been building toward since November 2025, and with what it just shipped: a search experience that keeps hotel-related clicks inside Google’s own ecosystem rather than sending them outward.
| Where AI Mode sends hotel-related clicks | Share of clickable links |
| Google Business Profile | ~79% |
| Hotel’s own website | ~17% |
| OTA | ~4% |
| Metasearch or review site | <1% |
Source: Nicolas Sitter, “Google AI Mode Hotel Study 2026,” February 2026 — an independent analysis of 4,000 AI Mode queries across eight cities, 6,249 hotel mentions, and 84,329 reference citations, collected via a third-party search API. Independent, single-study data; treat as directional rather than a settled industry benchmark.
The commercial consequence is where marginal effort should go. Paid search spend is increasingly competing for a shrinking pool of clicks that survive an AI Overview or AI Mode response at all, while an unpaid Google Business Profile — provided it carries a live, accurate rate and availability feed through Hotel Center — is disproportionately where the clicks that do occur end up. That makes Google Business Profile completeness a channel with a more direct line to captured demand than an incremental increase in paid budget, on the queries where AI-generated responses dominate the page — though it depends on the same Hotel Center connectivity discussed above, which is exactly what the September 30 cutoff is testing. Worth watching is whether GBP click and call volume is being reported to a hotel’s revenue team with anything like the rigor applied to paid-channel reporting; at most properties, currently, it isn’t.
5. Conversational Booking Went From Hypothetical to Live in One Week
On October 6, 2025, OpenAI launched an Apps SDK inside ChatGPT with Booking.com and Expedia as its first travel partners, giving both OTAs a presence inside ChatGPT’s then-roughly 800-million-weekly-user base and letting users search, compare, and move toward a booking without leaving the conversation. That was, until recently, the most concrete evidence that conversational booking was becoming real rather than theoretical. On August 27, 2026 — two days before this piece was filed — Google went further, rolling out in-chat hotel booking directly inside AI Mode for U.S. users: travelers can find, compare, select a room, review cancellation terms, and pay with Google Pay without leaving the chat, while the hotel or OTA remains the merchant of record handling the charge, confirmation, and guest support. Launch partners include Booking.com, Choice Hotels, Expedia, Hilton, Hotels.com, IHG, Marriott, Priceline, Trip.com, and Wyndham, according to Skift’s reporting on the rollout. Google has said it passes back only the minimum data needed to complete a booking — name, payment, occupancy, property — and withholds the broader chat context that produced the booking decision.
The revenue impact of this is not yet quantifiable, and it is worth being precise about why: the product is days old, U.S.-only at launch, excludes flight booking for now, and represents the first real test of whether travelers will let an AI chat interface — rather than the OTA’s or hotel’s own site — be the place a hotel booking actually gets completed. What is quantifiable already is an attribution gap: because Google is deliberately withholding chat context from its booking partners, a hotel’s existing analytics and PMS setup may not correctly capture where an AI Mode-originated booking came from unless referral and UTM handling is updated specifically for the new surface — a technical task, not a strategic one, but one that will otherwise show up as a mysterious gap in a hotel’s channel-attribution reporting.
Whether this deserves a DOSM’s attention now or in twelve months is the harder question, and the honest answer is that it is moving faster than the industry’s usual multi-year adoption curve for new distribution channels. Ten months passed between ChatGPT’s first travel app integrations and Google shipping actual in-chat payment; that is not the pace hotels are used to planning around. Treating agentic booking as too early to be a budget line item is defensible for most properties today. Treating it as something to revisit next year, rather than something to keep a standing watch on, is a different and harder-to-defend position given how much changed in this space in the ten months between those two product launches.
None of these five developments resolve into a single instruction, because the evidence does not point to one. What connects them is that in each case, control that used to sit with the person setting the hotel’s ad budget has moved — to an algorithm that needs better data to perform, to a technical dependency with a hard deadline, to a franchise relationship that predates the tools now spending its money, to a free listing doing more work than the paid campaign next to it, and to a chat interface that completed its first hotel booking this week. The tools look the same as they did in 2020. What each of them actually does with a hotel’s money does not.
Data Source
- Google Privacy Sandbox, “Update on Plans for Privacy Sandbox Technologies,” October 17, 2025 — Google’s own announcement retiring most Privacy Sandbox APIs (Topics, Protected Audience, Attribution Reporting, and others), six months after confirming it would not deprecate third-party cookies in Chrome.
- Google Hotel Center Help, “About Google third-party rates for hotel ads” — Google’s official documentation confirming the September 30, 2026 cutoff for Google-sourced hotel rate feeds and the migration requirements for affected advertisers.
- Google Ads Help, “Frequently asked questions about AI Max for Search campaigns” — official Google documentation on how AI Max’s performance depends on automated-bidding signals rather than manual CPC input.
- Google, “Google’s Dynamic Search Ads are upgrading to AI Max,” June 2026 — Google’s own blog post confirming the phased, largely automatic migration of Search campaigns to AI Max, concluding by February 2027.
- Pew Research Center, “Do people click on links in Google AI summaries?,” July 22, 2025 — tracked browsing-behavior study of 900 U.S. adults across 68,879 Google searches conducted in March 2025.
- Nicolas Sitter, “Google AI Mode Hotel Study 2026,” February 2026 — independent analysis of 4,000 Google AI Mode hotel queries across eight cities, examining where clickable hotel links route.
- SiteMinder, “Changing Traveller Report 2026,” January 2026 — survey-based research on where and how travelers begin and complete hotel research and booking, including the growing share starting on OTAs and the OTA-to-direct conversion pathway.
- SiteMinder, “Hotel Booking Trends 2026,” July 2026 — booking-value and channel-mix data drawn from bookings processed through SiteMinder’s platform, including average booking value by channel for 2025.
- Skift, “Google’s Agentic Hotel Booking Tool Comes to AI Mode,” August 27, 2026 — reporting on Google’s in-chat hotel booking launch inside AI Mode, including launch partners and data-sharing terms.
- PhocusWire, “ChatGPT brings apps into chat, starting with Expedia and Booking.com,” October 7, 2025 — trade reporting on OpenAI’s Apps SDK launch and its initial travel partners.
