What are the key sales KPIs for the Online Travel Agency (OTA) industry in 2027?

What it is and why it matters

An Online Travel Agency does not own hotel rooms, rental cars, or flights — it owns a marketplace connecting travel demand to supplier inventory, and it gets paid a commission (the take rate) for closing the transaction. That structural fact is why the nine-KPI stack looks nothing like a SaaS or retail scorecard. Every OTA metric ultimately answers one of three board-level questions: is gross demand growing, is the platform keeping enough of that demand after paying for distribution, and is the underlying unit economics improving as the discovery layer shifts from search engines to AI agents.

The two-sided marketplace dynamic is the single most important thing to understand about this industry. Booking Holdings spent roughly $7.4B on marketing in 2025 — about 27% of revenue — with the overwhelming majority going to Google Ads and Meta. Expedia's marketing-to-revenue ratio runs closer to 50%. This is often called the "Google tax": because OTAs largely do not own organic discovery, they must buy it back at auction, and every basis point that AI Overviews, ChatGPT, or Perplexity redirect away from the traditional search results page has to be replaced with paid spend or the funnel shrinks. That single dynamic is why Take Rate % and Customer Acquisition Cost have to be read together, never in isolation — a rising take rate funded by a faster-rising CAC is not real margin expansion, it is a treadmill.

The second structural wrinkle is cancellations. Reported Gross Bookings Value is gross of cancellations, and free-cancellation policies mean 20-40% of "booked" room nights never convert to a stay. Booking Holdings' free-cancellation share sits near 50% of room nights booked, so the net-of-cancellation GBV that actually generates revenue is materially smaller than the headline figure investors see in a press release. Any operator or analyst who reports gross bookings without a cancellation-adjusted figure alongside it is presenting an incomplete picture of the business.

The third wrinkle, and the one reshaping 2027 planning cycles across the industry, is loyalty and direct traffic. The economics of a booking flip dramatically the moment a returning customer books through the app instead of clicking a paid search ad. Booking's Genius loyalty tier, Expedia's One Key program (which crossed 130 million members by mid-2026), and Airbnb's "Guest Favorites" curation all exist for the same reason: to move repeat demand off the paid-acquisition treadmill and onto owned channels. Airbnb's direct/unpaid traffic share, near 90% of total, is the highest in the category and the primary reason its EBITDA margin runs above 35% while competitors with heavier paid-search dependence sit meaningfully lower.

Finally, agentic AI search is now a fourth structural force, not a future one. Booking Holdings signed distribution and integration deals with OpenAI and Anthropic during 2025; Expedia launched its own agentic travel assistant, Romie; Kayak built a ChatGPT plugin for flight and hotel search. By 2027, the operating question for every OTA finance team is what share of bookings originate from an AI agent conversation versus a traditional search-engine results page, because the take rate, the cancellation profile, and the CAC all behave differently across those two channels — and getting the attribution wrong misallocates marketing budget at scale.

The step-by-step process

Tracking OTA sales performance is a pipeline, not a single dashboard number. Demand enters through a discovery channel, survives (or doesn't) a cancellation filter, and only then becomes revenue subject to a take rate. Mapping that flow end-to-end is the first thing any RevOps or finance team building an OTA metric stack should do, because it makes clear where each of the nine KPIs sits in the funnel and which ones are leading versus lagging indicators.

Reading the diagram left to right: demand always starts as a discovery-channel decision, and that decision is the single biggest lever an OTA has over its own cost structure. A customer who lands via a Google Ads click carries a fully loaded CAC that can run 30-40% of the net revenue that booking generates. A customer who opens the app directly because of a loyalty push notification carries a CAC in the low single digits. A customer arriving via an AI agent referral is, as of 2027, still being actively measured — early data suggests these bookings convert at a lower CAC than paid search because intent has already been pre-filtered by the AI's own research process, but they also cancel more often because the traveler did less independent verification before committing.

Once a booking is created, it enters the cancellation filter. This step is where a large share of "phantom" revenue exits the pipeline — free-cancellation policies mean anywhere from a fifth to nearly half of gross bookings never convert into a completed stay. Only bookings that survive this filter become Net Bookings Value, and it is Net Bookings Value, not Gross Bookings Value, that the take rate should be applied against when a team wants a true picture of sustainable revenue. The final step, loyalty re-engagement, is the pipeline's feedback loop: a well-run loyalty program pulls a meaningful share of the next booking cycle back into the low-CAC direct channel, while a weak one sends the customer back to a generic search query where a competitor can win the click.

Costs, timelines, and typical ranges

Every one of the nine KPIs has a real-world range that practitioners use to judge whether a platform is healthy, and those ranges differ meaningfully by business model. On Gross Bookings Value, Booking Holdings posted $186.1B in FY2025 (up roughly 12.3% year over year), Expedia ran close to $115B, Airbnb was near $89-90B, and Trip.com Group was around $62B with international growth above 30%. On Take Rate %, Booking sits near 14.5%, Airbnb near 13%, Expedia near 11%, and Trip.com in the 7-8% range — take rate has been climbing industry-wide through ancillary attach (flights, attractions, insurance, payments) and expanded merchant-of-record fee structures, and a 100-basis-point move at Booking's scale is worth roughly $1.9B of incremental revenue.

Room Nights Booked is the pure-volume metric underneath GBV: Booking reported approximately 1.16B room nights in 2025, Airbnb reported roughly 547 million Nights and Experiences, and Expedia reported about 409 million. Growth is uneven by geography — North American hotel nights are growing low single digits while APAC alternative-accommodation nights have been growing above 25% annually, which is why blended global growth figures can mask very different regional stories.

On Customer Acquisition Cost, the spread between operators is the widest of any metric in the stack. Booking spends roughly $6.40 in marketing per room night booked; Expedia is closer to $11; Airbnb spends under $2 because of its direct-traffic advantage. A blended CAC that stays under roughly 35% of net revenue is generally considered sustainable industry-wide; once paid CAC growth outpaces take-rate growth for more than two consecutive quarters, margin compression typically follows within two more quarters.

Cancellation Rate % typically runs in a 25-40% band for platforms with heavy free-cancellation mix, like Booking.com, while Airbnb's stricter host-cancellation policies keep it closer to 12%. Average Daily Rate has been rising roughly 3% year over year globally, with US ADR near $159 and EMEA near €145 as of the most recent STR/CoStar benchmarking cycle — occupancy has largely flattened across major markets, meaning ADR growth is now doing most of the work driving GBV growth rather than volume. Alternative-Accommodation Mix has reached roughly 36% at Booking Holdings, with Phocuswright projecting the category-wide mix could reach 40% of total OTA room nights by 2027, making it one of the fastest-moving KPIs in the entire stack.

Reporting cadence matters as much as the ranges themselves. Room nights, GBV, and paid-search CPC should be reviewed daily. GBV by region and accommodation type, trailing-7-day cancellation rate, paid CAC by channel, and loyalty enrollments belong in a weekly review. Take rate by product line, gross margin per booking, ADR by region, and AI-channel mix of bookings are monthly checkpoints. A full P&L, alternative-accommodation mix trend, repeat-booking cohort decay, loyalty-program lifetime value, and a CFO-level reforecast of the marketing-to-revenue ratio belong in the quarterly board packet.

Where teams get it wrong

Four failure modes recur across the Online Travel Agency industry, and each one is visible in the KPI data well before it shows up in a quarterly earnings miss. The first is paid-search dependency without a loyalty offset: when Google CPC rises 15% in a quarter and the repeat-booking rate stays flat, that gap compresses operating margin by roughly 50 basis points every quarter it persists, because there is no owned channel absorbing the increased acquisition cost. The second is take-rate optimism without a corresponding rise in merchant-of-record share — pushing take rate higher through ancillary attach only works sustainably if the platform is also capturing a larger share of the transaction as merchant of record; otherwise the higher take rate functions as a hidden price increase that hotel and airline suppliers eventually route around by pushing customers toward direct booking.

The third failure mode is reporting headline Gross Bookings Value without a net-of-cancellation figure alongside it. A 35% cancellation rate makes raw GBV growth close to meaningless as a health signal if the cancellation rate itself is rising in parallel — a board evaluating a business on gross figures alone is being told an incomplete, and sometimes misleading, story about the true growth of the platform. The fourth, and most 2027-specific, failure mode is missing the agentic-AI pivot entirely. OTAs that had not signed distribution or plugin integration deals with major AI platforms by 2026 are now watching a measurable share of organic search traffic erode into AI Overviews and conversational agents with no equivalent channel backfilling the lost volume — and unlike a paid-search CPC spike, this kind of structural traffic loss does not reverse on its own.

Decision framework: when to choose what

Not every OTA should react to a KPI signal the same way, because the right lever depends on which metric is moving and in what direction. A useful decision framework separates "channel-mix problems," which are solved with loyalty and product investment, from "unit-economics problems," which require pricing or cost-structure changes, and from "discovery problems," which require a new distribution deal or channel build-out entirely.

When CAC is outrunning take-rate growth, the fix is rarely more marketing spend — it is renegotiating merchant-of-record terms with suppliers or increasing ancillary attach (insurance, car rentals, experiences) so more margin is captured per booking without raising the headline commission. When the repeat-booking rate is flat, the fix is a loyalty and app-engagement investment, not a pricing change — Expedia's One Key expansion and Booking's Genius tier restructuring are both examples of platforms treating a channel-mix problem as a product problem rather than a marketing problem. When cancellations are climbing, the right response is tightening cancellation policy terms or adding light pre-commitment friction at checkout, rather than discounting harder to compensate for lost volume. And when AI-channel share is stagnant while competitors are gaining ground, that is a distribution problem that marketing spend cannot solve — it requires a direct integration or plugin deal with the relevant AI platform, the same category of deal Booking, Expedia, and Kayak all pursued during 2025 and 2026.

Related questions

What is a healthy take rate for an OTA in 2027?

Mature platforms typically run 11-15% on accommodations; Booking sits near 14.5%, Expedia near 11%. Rates below 9% usually signal a niche or low-margin business model rather than underperformance.

How much should an OTA spend on customer acquisition?

A blended CAC under roughly 35% of net revenue is considered sustainable. Platforms with strong loyalty and direct-app traffic, like Airbnb, can run well below that threshold.

Why do OTA cancellation rates vary so much between platforms?

Cancellation rate tracks policy flexibility, not just demand quality. Platforms with generous free-cancellation terms, like many Booking.com listings, run structurally higher rates than platforms with stricter host policies.

How is AI search changing OTA customer acquisition?

Agentic AI channels are shifting first-touch bookings away from traditional paid search, often at lower CAC but with higher cancellation rates, requiring OTAs to track AI-sourced bookings as a distinct KPI segment.

What does alternative-accommodation mix tell you about an OTA's growth?

A rising mix of apartments, villas, and short-term rentals versus hotels signals demand diversification and typically correlates with higher average daily rates and stronger differentiation against hotel-only competitors.

FAQ

What is the typical take rate for OTAs in 2027? Take rates vary by business model: Booking Holdings runs near 14.5%, Airbnb near 13%, and Expedia closer to 11%. Niche or emerging platforms can fall outside this range depending on cost structure and merchant-of-record share.

How do OTAs measure customer acquisition cost in 2027? Most platforms track paid-search cost per acquisition as the primary metric, targeting a blended CAC below roughly 35% of net revenue, with a growing share of spend and attribution work now allocated to AI-driven discovery channels.

What counts as a healthy repeat-booking rate for an OTA? Mature platforms typically see repeat or loyalty-driven bookings in the 20-40% range, with top performers like Airbnb exceeding that through high direct/unpaid traffic share. Rates below 20% usually signal excess reliance on paid acquisition.

Why does cancellation rate matter so much for OTA financials? Cancellations directly reduce net revenue and raise servicing costs even when gross bookings look strong. Keeping cancellations under roughly 25% is a common industry benchmark; sustained rates above that erode gross margin per booking.

How does alternative-accommodation mix affect an OTA's competitive position? Alternative accommodations have been growing roughly twice as fast as traditional hotel bookings. A mix above 30% tends to correlate with higher average daily rates and stronger demand diversification versus hotel-only competitors.

What role does AI/agentic-search conversion play in 2027 KPI tracking? It measures how effectively AI-driven booking tools — chat agents, voice assistants, embedded LLM widgets — convert searches into confirmed bookings. As this channel grows, it materially lowers blended CAC when tracked and optimized as its own funnel segment.

Sources

• https://ir.bookingholdings.com/
• https://ir.expediagroup.com/
• https://news.airbnb.com/
• https://www.sec.gov/
• https://www.str.com/
• https://www.phocuswright.com/
• https://skift.com/
• https://www.euromonitor.com/
• https://investors.trip.com/

Related on PULSE

• [What are the key sales KPIs for the Online Grocery and Q-Commerce Delivery industry in 2027?](/knowledge/ik0319)
• [Top 10 Real Estate Agency Revenue KPIs](/knowledge/ik0715)
• [Top 10 Home Health Care Agency Revenue KPIs](/knowledge/ik0662)
• [Top 10 Advertising Agency Revenue KPIs](/knowledge/ik0610)
• [What are the key sales KPIs for the Commercial Marketing Agency industry in 2027?](/knowledge/ik0014)