What OTA Commission Actually Costs You
The real cost of an OTA booking is commission plus payment fees plus the cancellation gap — often more than 20% of the rate. Here is the full arithmetic.
The commission percentage on your contract is not what an OTA booking costs you. Add payment processing, then add the fact that OTA bookings cancel roughly twice as often as direct ones, and a headline 18% commission lands closer to 20-25% of the money that actually reaches your bank. A direct booking at the same rate typically costs you a third of that. This post works the whole sum out with real numbers so you can put it in front of an owner.
The number on your contract is the smallest part of the bill
Ask an operator what OTA commission costs and you get one figure — 15%, 18%, sometimes 25% with a visibility booster switched on. That figure is real, but it is the first line of a longer invoice.
Three things sit underneath it. Payment processing, which someone pays regardless of who collects. The cancellation rate, which decides how many of those bookings turn into actual occupied rooms. And the marketing and technology cost of the direct channel you are comparing against — because direct is cheaper, not free.
Until you total all four, you are arguing about a percentage instead of about money.
What the acquisition-cost research found
The best public work on this is not from a vendor. The American Hotel & Lodging Association, STR and Kalibri Labs studied roughly 25,000 US hotels over 2014-2016 in Demystifying the Digital Marketplace, Part 2.
Their finding: customer acquisition cost — everything you spend to get one booking, including commission, transaction fees, loyalty costs and marketing — rose from 5-10% of guest-paid revenue in the 1990s to 15-25% by 2016. The average sat at 16-18%. Some hotels reached 35-40%.
Two other numbers from that work matter for a small independent. Acquiring a guest through an OTA cost roughly 2.5 times what acquiring one through the hotel's own brand.com channel cost. And Expedia plus Booking Holdings accounted for about 96% of US OTA business — so "shop around for a better commission" is not a strategy that exists.
Commission ranges reported by Cloudbeds and Preno put independents at 15-30% and large brands at 10-15%. You are on the wrong side of that gap by default.
The worked comparison: one $180 booking, two channels
Here is a single night at $180 booked two ways. The percentages below the rate are a labelled worked example, not published figures — swap in your own contract terms and your own merchant statement.
| Line item | OTA booking | Direct booking |
|---|---|---|
| Rate paid by guest | $180.00 | $180.00 |
| Commission @ 18% | −$32.40 | — |
| Booking engine fee @ 2% | — | −$3.60 |
| Card processing @ 2.5% | −$4.50 | −$4.50 |
| Digital marketing allocation @ 3% | — | −$5.40 |
| Net to hotel | $143.10 | $166.50 |
| Effective acquisition cost | 20.5% | 7.5% |
The gap is $23.40 per booking — 13% of the rate, gone. The ratio between the two channels is about 2.7 to 1, which sits close to the 2.5x Kalibri measured. That is a sanity check, not a coincidence.
Two honest caveats. If the OTA collects payment under a merchant model, the card fee is inside their margin rather than on your statement — the total still comes out of the same $180. And your digital marketing allocation is genuinely arguable; some of that spend would exist anyway.
The cancellation differential nobody prices in
A booking that cancels is not a booking. The Cloudbeds 2026 State of Independent Hotels panel — 90 million bookings across 180 countries, 2025 data — put the cancellation rate for independent hotels at 21.8% on OTA bookings against 10.6% direct.
That changes the arithmetic. Take 100 bookings at $180 through each channel:
| OTA | Direct | |
|---|---|---|
| Bookings taken | 100 | 100 |
| Cancellation rate (Cloudbeds panel) | 21.8% | 10.6% |
| Bookings that actually stay | 78.2 | 89.4 |
| Net per stayed booking (table above) | $143.10 | $166.50 |
| Net revenue per 100 bookings taken | $11,190 | $14,885 |
| Net per booking taken | $111.90 | $148.85 |
Per booking taken, direct is worth about a third more. That is before you count the forecasting damage — a channel that cancels one booking in five makes your pickup numbers less trustworthy, which pushes you toward pricing decisions you would not otherwise make.
The caveat is real and cuts the other way: on a high-demand date, a cancelled room usually resells, so the loss is smaller than the table suggests. On a soft midweek night it usually does not, and the table understates it.
NRevPAR: the metric that actually captures this
RevPAR — revenue per available room — is room revenue divided by rooms available (STR's definition). It is the industry's default because it is the only metric everyone benchmarks. The HSMAI APAC and Singapore Institute of Technology metrics study in 2018 found 77.4% of respondents tracked RevPAR, against 20.4% for GOPPAR and 13.7% for TRevPAR.
RevPAR has one blind spot that matters here: it counts the gross rate and ignores what you paid to get it. Two hotels with identical RevPAR can be $20 a room apart in what they keep.
NRevPAR — net revenue per available room — is room revenue minus distribution costs, divided by rooms available. Same denominator, honest numerator.
A 40-room hotel over a 30-night month, using the figures above and the Cloudbeds panel's 63.4% OTA share of independent bookings:
| Line | Value |
|---|---|
| Rooms available (40 x 30) | 1,200 |
| Occupancy 70% → rooms sold | 840 |
| ADR | $180.00 |
| Room revenue | $151,200 |
| RevPAR | $126.00 |
| OTA rooms (63%) x $36.90 cost | −$19,520 |
| Direct rooms (37%) x $13.50 cost | −$4,199 |
| Net room revenue | $127,481 |
| NRevPAR | $106.23 |
Nearly $20 of RevPAR per available room, every night, is distribution cost — about $23,700 that month. You cannot manage a number you do not calculate.
What a mix shift is actually worth
Now the useful question. Move ten points of that mix from OTA to direct — 84 room nights — and you save $23.40 on each. That is $1,966 a month, roughly $23,600 a year, on the same occupancy and the same ADR. NRevPAR moves from $106.23 to $107.87 without selling one extra room.
That is the real prize, and it is also why the honest answer to "should I quit the OTAs" is no. The OTAs are delivering volume you would not otherwise get, and there is credible research that listing on them also drives direct bookings — covered in the companion post on why your OTA commission is partly a marketing bill. The goal is a better mix at a defensible price, not a boycott.
The mistakes that cost the most
Quoting the commission percentage as the cost. It is roughly three-quarters of the cost. Add processing and cancellation before you make a decision on it.
Comparing OTA cost to zero. Direct bookings carry booking engine fees, card fees and marketing. Compare 20.5% to 7.5%, not to nothing.
Ignoring the cancellation gap in your forecast. If a fifth of your OTA bookings will cancel and you forecast as if they will not, you will discount dates that were never soft.
Chasing direct share at any price. A 25% discount to win a direct booking costs more than the 13% you saved. The mix is only worth shifting while the shift is cheaper than the gap.
Tracking RevPAR alone. It is the right benchmark against your comp set and the wrong number for a distribution decision.
The bottom line
Take your contracted commission, add your payment processing, then adjust for the fact that OTA bookings cancel at roughly twice the direct rate. On a $180 room that is about $36.90 of cost against $13.50 direct — the gap Kalibri's 2.5x ratio predicts. Multiply the gap by your monthly OTA room nights and you have the annual number to show an owner. Then track NRevPAR alongside RevPAR so the cost stays visible every month instead of once a year.
Working out the gross numbers is the easy half. If you want ADR, RevPAR and occupancy against your market without rebuilding the spreadsheet each month, our analytics and reporting module produces them alongside STAR benchmark indices, exportable to Excel so you can add your own cost lines.