Travel & Hotel Marketing Statistics 2026: OTA Spend, Direct Booking Share and AI Planning

Travel is the most expensive category in digital marketing to compete in, and the annual reports now say so out loud. Two online travel agencies moved more than $300bn of bookings between them in 2025, and the marketing line in their accounts runs to well over $15bn combined.

That spend is not abstract. It flows almost entirely to Google, meta-search, affiliates and social platforms, and it sets the floor for what an independent hotel or travel brand pays to be visible in the same auctions.

What has changed for 2026 is that the same filings now name generative AI as a live threat to the booking funnel. Expedia's own risk factors warn of "reduced direct-booking rates, and increased marketing expenses as we compete for visibility".

Key travel and hotel marketing statistics for 2026

• Booking Holdings recorded $186.1bn of gross bookings in 2025, up 12.4% year on year, on 1,235 million room nights [1]
• Booking Holdings spent $8,186m on marketing in 2025, equal to 4.4% of gross bookings and 30.4% of total revenues [1]
• Expedia Group spent $7,349m on direct selling and marketing in 2025, or 49.9% of its total revenue of $14,733m [2]
• Expedia Group B2B gross bookings grew 20% to $35.7bn in 2025 whilst B2C gross bookings grew just 3% [2]
• Airbnb spent $2,588m on sales and marketing in 2025, 21.1% of revenue and 2.8% of its $91.3bn gross booking value [3]
• trivago spent €418.2m on advertising in 2025, equal to 76.2% of its total revenue, with Booking Holdings and Expedia together supplying 72% of that revenue [4]
• Tripadvisor spent $791.4m on marketing in 2025, 41.8% of revenue, rising to 58.2% of revenue inside its Experiences segment [5]
• IHG Hotels & Resorts says its direct digital booking channels accounted for over 26% of total room revenue in 2025, and that its 160 million-plus loyalty members booked 66% of room nights globally [8]
• Marriott International reports that approximately 68% of global hotel room nights and 75% of US room nights were booked by loyalty programme members in 2025 [6]
• Hilton ended 2025 with 243 million Hilton Honors members, up 15% year on year [7]
• Deloitte found that nearly a quarter of travellers used generative AI tools for trip planning in late 2025, roughly three times the 2022 share [12]
• The Office for National Statistics estimates UK residents made 94.6 million visits abroad in 2024 and spent £78.6bn [10]

How much of a travel booking does the platform actually keep?

The blended take rate across the big platforms sits between 12% and 14.5% of gross bookings. That is the real "commission economics" number, and it is lower than the headline hotel commission rates people quote, because it blends flights, cars and agency-model bookings.

Booking Holdings reported total revenues at 14.5% of gross bookings in 2025, up from 14.3% in 2024 [1]. Expedia Group came in at 12.3%, with a B2C revenue margin of 11.3% and a B2B margin of 13.6% [2]. Airbnb sat at 13.4% on $91.3bn of gross booking value [3].

The gap between those blended rates and the 15% to 25% commission a small independent hotel is quoted matters. Flights carry a fraction of the margin of accommodation, so a platform with a fast-growing air business shows a falling take rate even as its hotel commissions hold firm. Booking Holdings sold 68 million airline tickets in 2025, up 36.6%, and flags explicitly that flight gross bookings carry lower revenues as a percentage of gross bookings [1].

Commission is also not the only cost of the channel. Rate parity, loyalty discounting and the traffic an OTA diverts from a hotel's own site all sit outside the headline percentage.

How much do online travel agencies spend on marketing?

Enormous amounts, and the ratios differ wildly by business model. Marketing runs at 30.4% of revenue at Booking Holdings, 49.9% at Expedia, 21.1% at Airbnb, 41.8% at Tripadvisor and 76.2% at trivago.

Booking Holdings describes its marketing as "substantially variable in nature" and holds it at a steady 4.4% of gross bookings across both 2025 and 2024 [1]. Expedia's direct selling and marketing costs are 6.1% of its gross bookings on the same calculation.

That difference is the most useful comparison in travel marketing. Booking spends 1.7 percentage points less of every booking on marketing than Expedia does, on a book of business more than 50% larger.

Airbnb is the outlier in the other direction at 2.8% of GBV. Its brand does much of the acquisition work its rivals pay search engines for, which is the position every hotel group says it wants and almost none achieve.

What share of hotel bookings actually come direct?

Direct share depends entirely on who is measuring and what they count. On the platform side, Booking.com says a mid-fifties percentage of room nights came from consumers arriving directly in 2025. On the hotel side, IHG puts its own direct digital channels at just over a quarter of total room revenue.

Booking Holdings reports that the share of room nights booked by consumers coming directly to its platforms was a mid-fifties percentage in 2025, and higher still if affiliate bookings are excluded [1]. Its mobile app mix also reached a mid-fifties percentage, up from a low-fifties percentage in 2024, and the filing notes the significant majority of app room nights are direct.

IHG Hotels & Resorts states that its direct digital booking channels accounted for over 26% of total room revenue in 2025, with overall enterprise contribution up 2 percentage points year on year [8].

Read those two together and the picture is uncomfortable for hotels. Booking.com's "direct" is direct to Booking.com, not direct to the property, and nearly half of its room nights still arrive through paid search, meta-search, affiliates and social. That is where the $8.19bn goes.

For a hotel marketer, the OTA's direct share is a measure of brand strength the hotel does not benefit from. Winning share back means competing with a company that treats roughly a third of revenue as an acquisition budget.

Do loyalty programmes actually drive direct bookings?

They are the most reliable direct channel the hotel groups have. Two of the three largest report that roughly two thirds of global room nights now come from loyalty members, with US shares in the seventies.

Marriott International states that in 2025 approximately 75% of its US hotel room nights and approximately 68% of global room nights were booked by Marriott Bonvoy members [6]. IHG grew One Rewards enrolments 25% in 2025 and puts member-booked room nights at 66% globally [8]. Hilton reported 243 million Honors members at 31 December 2025, up 15% [7].

The economics show up on the balance sheet as well. Marriott's deferred revenue liability for the loyalty programme stood at $7,992m at the end of 2025, up $473m in the year, primarily reflecting points earned by members [6].

A caution. "Booked by a loyalty member" is not the same as "booked direct". A Bonvoy member who books through Expedia and adds their number at check-in still counts in that 68%. The figure measures programme reach, not channel share, and the two are routinely conflated in hospitality marketing decks.

Does meta-search still work, and who is paying for it?

Meta-search remains a functioning channel, but its economics are brutal and its customer base is dangerously narrow. trivago spent 76.2% of its total revenue on advertising in 2025 and took 72% of its revenue from just two customers.

trivago reported total revenue of €548.9m in 2025, up 19%, against advertising spend of €418.2m and total selling and marketing expense of €445.6m, or 81% of revenue [4]. Booking Holdings and its affiliates represented 39% of trivago's revenue and Expedia Group 33%.

That is a business that buys traffic from Google, sells it to two buyers, and keeps a sliver. It works only whilst both buyers keep bidding.

The lesson for hotel marketers is about the shape of the auction rather than trivago itself. On meta-search you are bidding against companies whose cost structure is built around paying more than you can for the same click, because they monetise the traveller across several future trips.

How are travellers researching trips in 2026?

Research is fragmenting away from search towards social video and AI assistants, and the split is generational rather than universal. Older travellers still behave broadly as they did five years ago.

Deloitte found that more than half of Gen Z respondents use short-form social video for travel research, against 34% of millennials and 14% of Gen X and baby boomers combined [12]. It also reports that as millennials use generative AI more, they report less use of every other research source.

Booking Holdings adds a note from the transaction side, saying it saw the booking window expand in 2025 compared with 2024, which benefited room night growth [1].

The commercial implication is awkward. A longer booking window plus more research surfaces means more touchpoints to fund, at a moment when attribution across social video, AI answers and search is at its weakest in a decade.

How many people use AI to plan and book travel?

Around a quarter of travellers, on the best-sourced estimate available, and the growth curve is steep. Actual AI-completed bookings remain far rarer than AI-assisted planning.

Deloitte's 2026 travel industry outlook reports that nearly a quarter of travellers used generative AI tools for trip planning in late 2025, roughly three times the 2022 figure, with use tripling between 2023 and 2025 [12]. Millennials lead adoption for itinerary building, hotel selection and finding things to do at the destination.

The filings are more nervous than the survey. Booking Holdings warns that generative AI "lowers barriers to entry", and that "AI agents may further evolve into full-service booking platforms, increasing competitive pressure and disintermediating OTCs" [1]. Expedia goes further, warning of "reduced direct-booking rates, and increased marketing expenses as we compete for visibility" if consumer behaviour shifts toward AI platforms [2].

Whitbread, which owns Premier Inn, has already acted on it, stating that during the year it "continued to update our online collateral so as to ensure that the online experience for consumers choosing to search using AI tools is optimised" [9]. When a FTSE 100 budget hotel operator is writing generative engine optimisation into its annual report, the practice has stopped being experimental.

How much do reviews and user-generated content still matter?

They remain the largest single content asset in travel, and the scale is difficult to replicate. Tripadvisor now hosts more than one billion reviews and opinions covering over nine million destinations, points of interest, accommodations, restaurants, airlines and cruises [5].

Tripadvisor attracts hundreds of millions of visitors annually across more than 40 countries and over 20 languages, and describes that content as a "core strategic asset" underpinning its AI plans [5]. Its Hotels and Other segment revenue nonetheless fell 8% in 2025 to $750.1m, whilst Experiences revenue grew 10% to $924.4m.

That divergence is the story. Review scale did not protect the hotel meta business from being squeezed, and Tripadvisor is now spending 58.2% of Experiences segment revenue on marketing to buy demand for the part that still grows [5].

For an individual property, review volume is table stakes rather than an advantage. The differentiator is recency and response, not the raw count on a profile page.

How seasonal is UK outbound travel?

Sharply, and more so for holidays than for any other trip purpose. UK residents made 94.6 million visits abroad in 2024 and spent £78.6bn, according to the Office for National Statistics [10].

Quarterly data for Great Britain residents shows 28.9 million visits abroad in Q3 2024 against 16.8 million in Q1, a ratio of 1.72 to 1 [11]. Holiday visits specifically swing harder: 18.7 million in Q3 against 9.2 million in Q1, a ratio of 2.02 to 1.

Spain took 17.8 million GB visits in 2024, ahead of France on 9.3 million and Italy on 4.8 million [10]. Package holidays accounted for 46.1% of all holiday visits abroad on the same dataset.

Budget planning that treats travel demand as flat across the year will misallocate badly. Demand peaks in Q3, but the buying curve sits well ahead of it, and the Q1 search spike for summer holidays is a marketing event with almost no revenue recognition attached.

The domestic picture is different again. Whitbread puts the UK hotel market at around 198 million room nights a year across approximately 735,000 rooms, with independent supply down 11% since 2019 and total supply not expected back at 2019 levels until at least 2028 [9].

How to read these numbers

Company filings and consumer surveys are not the same kind of evidence, and mixing them carelessly produces nonsense.

The platform figures here are audited and server-side, straight from 10-K and 20-F filings, so they are as reliable as travel data gets. They are also company-defined: "gross bookings" and "direct" mean whatever each company says they mean, and the definitions differ. Booking Holdings discloses direct share as "a mid-fifties percentage" rather than a precise number, so treat it as a band.

The ONS travel statistics are labelled official statistics in development. The 2024 annual estimates combine two survey designs, with Q1 and Q2 collected under the previous International Passenger Survey design and Q3 and Q4 under a new one, and the ONS has flagged known issues including incorrect airport codes and underestimated overseas spending. Quote the direction of travel, not the third significant figure.

Deloitte's generative AI figure is self-reported survey data drawn from three separate 2025 surveys, with sample sizes not disclosed on the outlook page. Self-reported AI use tends to run ahead of observed behaviour, so a quarter of travellers "using" AI for planning is not a quarter of trips being planned by AI.

Finally, the marketing ratios compare businesses with different revenue recognition. Booking's marketing is expensed when a booking is made whilst revenue is recognised at check-in, so the ratio is not a clean return on spend in any single quarter.

What this means going into 2027

The acquisition floor is set by companies spending a third of revenue on marketing. Booking at 30.4% and Expedia at 49.9% of revenue define what a competitive travel keyword costs. Independent operators should assume they cannot win the head terms and build around intent that platforms undervalue.

Loyalty is now the only durable direct channel at scale. Two thirds of room nights at the majors come from programme members. Anything that increases enrolment and repeat rate compounds; anything that only wins a single booking does not.

Agentic booking is the risk the filings actually name. Both large OTAs have written AI disintermediation into their risk factors for 2025. Structured, machine-readable property and rate information is becoming a distribution requirement rather than an SEO nicety.

B2B distribution is where the growth is. Expedia's B2B gross bookings grew 20% against 3% for B2C. The traveller-facing brand is no longer the only route to volume, and white-label distribution deserves a line in more travel marketing plans than it currently gets.

Measurement will get worse before it gets better. Longer booking windows, social video research and AI-mediated answers all sit outside conventional attribution. Plan on incrementality testing and geo experiments rather than platform-reported conversions. If you want help working out how findable your travel or hotel brand really is in search and AI answers, you can book a call with me.