Your OTA Cancellation Rate Is Nearly Double Your Direct Rate. Here's the Real Cost.

I was going through Cloudbeds’ 2026 State of Independent Hotels Report, built from roughly 90 million real bookings, and one line stopped me: OTA bookings cancel at nearly double the rate of direct bookings. It’s the kind of number I think is genuinely useful for clients and readers alike to see laid out clearly, so it seemed worth pulling out and explaining on its own.

Picture a 30-room independent hotel that gets most of its bookings through OTAs. Its OTA channel shows $180,000 in confirmed reservations for the next quarter. That number looks solid, until the cancellations start coming in, and a meaningful share of it never actually arrives as revenue.

Cancellation rate is the OTA cost most hotel owners never calculate. Commission gets attention because it shows up on every statement. Cancellation risk is quieter. It shows up later, as a forecast that turns out to be wrong.

A few terms come up throughout this piece, so in plain terms:

What cancellation rate actually measures

Cancellation rate is the share of confirmed reservations that get cancelled before the guest arrives. A hotel with 100 room-nights booked in a month and 20 cancellations has a 20% cancellation rate.

This is a different kind of risk than commission. Commission is a known cost, paid on every booking that actually happens. Cancellation is uncertainty. It affects how much you can trust a number on your books before the guest actually shows up.

The real numbers

Cloudbeds’ 2026 State of Independent Hotels Report, built from roughly 90 million real bookings across tens of thousands of properties in 180 countries, found OTA cancellation rates at 21.8% in 2025. Direct bookings cancelled at 10.6% over the same period.

That’s close to double. A booking made through an OTA is roughly twice as likely to fall through as a booking made directly on a hotel’s own website.

A booking through an OTA is nearly twice as likely to disappear before the guest arrives as a booking made directly.

Why OTA bookings get cancelled more

Nobody has published a definitive answer, but a few explanations show up consistently.

Free cancellation makes speculative booking cheap. A guest can hold a room on an OTA with no commitment, keep comparing prices, and cancel later if something better turns up. Most OTA rate types make that free.

Cancelling on a large platform also feels less personal than cancelling with a hotel directly. There’s no relationship to consider, just a button.

Free cancellation doesn’t cost the guest anything. It costs the hotel a forecast it can’t fully trust.

Mirai, a hotel booking-engine company, studied this back in 2016. It’s a small sample (40 hotels, four months) and a decade old, so treat it as historical texture, not a current benchmark. It found Booking.com cancellations ran 104% higher than a hotel’s own direct channel, and Expedia ran 31% higher. Mirai’s own explanation pointed to the same speculative-booking behavior, plus a share of fraudulent bookings concentrated on the largest platform.

The pattern isn’t new. It’s been showing up in real data for at least a decade.

What the gap actually costs

Picture a hotel with $500,000 in annual OTA revenue and $300,000 in direct revenue. Apply the Cloudbeds rates to each: roughly $109,000 of that OTA figure represents bookings that get cancelled before arrival, against about $32,000 on the direct side.

That’s not lost revenue outright. A cancelled room can often be resold, especially if the cancellation comes early. It’s forecasting error: the gap between what your booking calendar says and what actually walks through the door. For this property, that gap is more than three times larger on the OTA side than the direct side.

The number on your booking calendar and the number that actually walks through the door are not the same number, and the gap is bigger on your OTA channel than you probably assume.

The bigger a hotel’s OTA share, the more of its total booked revenue sits inside this less-reliable channel.

Guests are giving more notice

There’s a real mitigating factor here. Cloudbeds’ data also shows the average cancellation lead time (how far before arrival a guest cancels) rose from 34.6 days in 2023 to 38.7 days in 2025.

Guests are cancelling further in advance than they used to. That gives a hotel more time to resell a cancelled room before it goes empty.

This softens the cost of the cancellation gap. It doesn’t close it. A room cancelled 39 days out is easier to resell than one cancelled 3 days out, but a 21.8% cancellation rate is still a lot more forecasting noise than a 10.6% rate, regardless of how much notice you get.

What to actually do about it

A few things follow from this, and none of them are exotic.

Don’t treat OTA-channel bookings as reliably booked revenue until much closer to arrival. If you’re forecasting cash flow or staffing off total booked revenue, weight the OTA portion down.

Size your overbooking buffer around the real gap, not a guess. A property that leans heavily on OTAs can reasonably carry a larger buffer than one that leans direct, because more of its booked revenue sits in the less-reliable channel.

Give guests a reason to book direct with a non-refundable or more-restrictive rate option. A guest who picks a direct rate specifically because it’s cheaper in exchange for less flexibility is a guest whose booking is far more likely to actually happen.

Not every dollar of booked revenue is equally reliable, and pretending otherwise is where the forecasting error comes from.

None of this replaces OTAs. See reduce OTA dependence for the fuller case on that. This is about not treating every dollar of booked revenue as equally reliable, because it isn’t.

This is one number out of Cloudbeds’ report. There’s more in it worth digging into, and I’ll be pulling out what stands out to me in future posts.

And if you want a second opinion on what your own cancellation numbers say about your booking mix, let’s talk - fifteen minutes, no obligation.