One picture repeats in the hotel forecasts I open during audits. A Saturday six weeks out shows 91 percent on the books, the team has already pushed the rate up twice, and everyone in the revenue meeting is relaxed. When I split that Saturday by channel and rate plan, close to half of the rooms usually sit on fully flexible OTA rates that guests can cancel free until the day before arrival. That night is not 91 percent sold. It is 91 percent promised.
That gap between promised and real is the single most common forecasting error I find in hotel audits in 2026, and it is getting more expensive. Cloudbeds' 2026 State of Independent Hotels report, built on 90 million bookings, puts the OTA cancellation rate at 21.8 percent against 10.6 percent for direct bookings. In this article I explain what a hotel cancellation rate really measures, what a normal rate looks like by channel and lead time, the audit mistakes that turn cancellations into lost revenue, a worked example with the arithmetic, and the fixes I put in place first.
What Is a Hotel Cancellation Rate?
A hotel cancellation rate is the share of confirmed reservations for a set of stay dates that guests cancel before arrival. Hotels usually measure the hotel cancellation rate by reservations, room nights or room revenue, and each version tells a different story about how much of the business on the books will actually arrive and pay.
The basic formula is simple. Cancelled reservations divided by total reservations created, multiplied by 100. If a hotel created 400 reservations for November 2026 stay dates and 88 of them were later cancelled, the hotel cancellation rate by count is 22 percent.
The simple version is also where most hotels stop, and that is the first problem. I always ask which of three versions the report is showing, because they rarely match.
Three ways to measure cancellations
| Measure | Formula | What it tells you | Where it misleads |
|---|---|---|---|
| By reservation count | Cancelled reservations divided by reservations created | How often guests change their minds | Treats a one-night booking and a seven-night booking the same |
| By room nights | Cancelled room nights divided by room nights booked | How much inventory comes back to you | Ignores the rate the cancelled nights were sold at |
| By room revenue | Cancelled revenue divided by revenue booked | How much of the forecast revenue is at risk | Can look worse in peak periods when cancelled rooms carried high rates |
For forecasting, the room-night version matters most, because it tells you how many rooms will come back into inventory. For owner reporting, the revenue version matters most, because it shows how much of the revenue on the books in 2026 is soft. I track both, and I track them by stay date, not by the date the cancellation arrived. A cancellation report sorted by cancellation date tells you how busy the reservations team was. A report sorted by stay date tells you what will happen to a given night.
Gross pickup versus net pickup
Gross pickup is every new booking for a stay date. Net pickup is new bookings minus cancellations for the same stay date over the same period. If you have read my guide on reading a hotel pickup report, you know I only make pricing decisions on net pickup. A Saturday that picked up 14 rooms and lost 9 to cancellations did not have a strong week. It had a net week of 5.
Bottom line: Measure the hotel cancellation rate by room nights and revenue for each stay date, and price only on net pickup.
What Is a Normal Hotel Cancellation Rate?
A normal hotel cancellation rate in 2026 sits at roughly 18 to 22 percent of reservations across all channels, based on the largest booking datasets published this year and last. Normal varies a lot by channel, lead time and market, so a hotel should compare its own rate with its own history first and with industry averages second.
Here is what the large datasets say. SiteMinder's Hotel Booking Trends research, based on more than 130 million hotel bookings, put the global cancellation rate for 2025 at about 19.15 percent, and the United States at 18.78 percent, down from 19.22 percent the year before. SiteMinder's mid-year 2026 report, covering June to September 2026 stays across 22 markets, found cancellation rates rising in most regions: up 0.81 percentage points year on year in Asia Pacific, up 0.5 points in Europe and up just 0.1 points in the Americas.
Older European data shows how high the number can climb when free cancellation becomes the default. D-EDGE's study of its European client base from 2014 to 2018 found the average cancellation rate rose from 32.5 percent of on-the-books revenue in 2014 to 39.6 percent in 2018. That is a revenue-based measure, so it is not directly comparable to the 2025 reservation counts, but the direction was clear and it was driven by flexible OTA policies.
How I read a hotel's own number
When I audit a hotel, I do not judge its blended cancellation rate against a national average. I judge it against three things:
- The same months last year, by stay date, so seasonality is removed.
- The channel mix, because a hotel with 70 percent OTA share will run a higher rate than one with 40 percent.
- The lead time mix, because a resort booked 90 days out will always cancel more than a city hotel booked 9 days out.
A city hotel at 25 percent can be in worse shape than a resort at 32 percent if the city hotel's cancellations arrive inside 72 hours, where there is no time to resell. Timing matters as much as the rate.
Bottom line: A rate near 20 percent is typical in 2026, but the real test is whether your cancellations are predictable and early enough to resell.
Hotel Cancellation Rate by Booking Channel
The hotel cancellation rate by booking channel varies more than any other cut of the data. OTA reservations cancel far more often than direct reservations, because OTAs promote free cancellation heavily and guests use it to hold rooms while they keep shopping. Every channel needs its own cancellation assumption in the forecast.
The Cloudbeds 2026 report is the clearest current source on the gap. OTA bookings at independent hotels cancelled at 21.8 percent in 2025, against 10.6 percent for direct bookings. The same report found OTAs produced 63.4 percent of independent hotel bookings, with some markets approaching 80 percent. Put those two numbers together and the conclusion is simple: for most independent hotels in 2026, the majority of the rooms on the books come from the channel most likely to cancel.
Within OTAs, the spread is wide too. In D-EDGE's 2018 European data, reservations from Booking Group cancelled at 49.8 percent, and roughly two-thirds of reservations from HRS Group were abandoned. Those figures are old, and policies have changed since, but in the hotel audits I run today, the OTA that defaults guests to free cancellation is still the one with the highest cancellation rate in almost every portfolio.
Why OTA guests cancel more
- Free cancellation is the default display. Guests learn they can book without risk and treat the reservation as an option, not a commitment.
- Multiple holds. A traveller planning a trip often books two or three hotels on flexible rates, then cancels the losers once plans firm up.
- Price tracking. Price-drop alert services and rebooking apps watch refundable reservations and prompt guests to cancel and rebook when the rate falls, which means your own discounts can trigger your own cancellations.
- Loose loyalty. An OTA guest has no relationship with the hotel, so cancelling costs them nothing emotionally.
Direct guests cancel less because they have usually chosen the hotel already. Many come back as repeat guests or book after a search that started with the hotel's name. That is another reason a strong direct channel lowers your cost of distribution twice: once in commission and again in cancellation risk. I covered the commission side in detail in what OTAs really charge hotels.
Bottom line: Never apply one cancellation rate to all channels, because in 2026 OTA rooms cancel about twice as often as direct rooms.
Cancellation Risk Rises With Lead Time
Cancellation risk in hotels rises with lead time, because the further away a stay is, the more time a guest has to change plans, find a cheaper option or drop the trip. A reservation made four months ahead on a flexible rate is far less certain than one made four days ahead, even at the same price.
The trend in 2026 is toward longer windows on both sides. Cloudbeds found travellers booked independent hotels an average of 40 days ahead in 2025, up from 38 days in 2023, with North America at 48 days and EMEA at 47 days. Over the same period, the average cancellation window expanded to 39 days, up from 35 days in 2023. Guests are booking earlier and cancelling earlier.
At the far end of the booking window, the risk is much higher. Profitroom's 2026 analysis of European hotels, as reported in the trade press, found that OTA bookings made 180 days or more before arrival cancelled at rates between 28 percent and 47.6 percent depending on the market, with direct bookings lower in every region. Profitroom calls this phantom demand: reservations that look like demand on the forecast but were never firm.
What longer windows mean for pricing
Longer cancellation windows are not all bad news. If a hotel knows that a share of its far-out business will cancel around 39 days before arrival, it also knows it will get those rooms back with five weeks left to sell them. The danger is only for hotels that do not expect the cancellations and price as if the rooms were gone. I wrote about the booking curve itself in whether the hotel booking window is shrinking, and the same logic applies here: you cannot read pace without reading cancellations on the same curve.
In practice, I build a simple matrix for every hotel: lead time bands down the side (0 to 7 days, 8 to 30, 31 to 90, 91 to 180, over 180) and channels across the top, with the historical room-night cancellation rate in each cell. That matrix becomes the cancellation assumption for the forecast. It takes an afternoon to build from twelve months of reservation data, and it changes how the team reads every future date.
Bottom line: Treat far-out flexible OTA reservations as soft demand, and expect a large share of them to return to inventory about five weeks before arrival.
Cancellation Mistakes I Find in Audits
Cancellation mistakes in hotel revenue management are rarely about the cancellation policy itself. The mistakes I find in audits are about how the hotel reads, forecasts and prices around cancellations: gross occupancy treated as firm, one blended rate for every channel, no overbooking plan and discounts that trigger rebooking.
These are patterns I see across the hotels our team reviews, not one property. Most hotels have at least two of them.
Mistake 1: Pricing the last rooms off gross occupancy
This is the pattern from the opening of this article. The hotel sees 90 percent on the books, raises the rate for the remaining rooms and closes the discounted plans. Then 12 percent of the rooms cancel inside the last ten days, and the hotel is suddenly selling those rooms at last-minute prices on the busiest discount channels. The rate that looked strong six weeks out turns into a weak ADR on the night.
Mistake 2: One cancellation assumption for every channel
When a forecast uses a single cancellation factor, it overstates the risk on direct and corporate business and understates it on flexible OTA business. In audits I often find the forecast is close on average and wrong on every individual night, because the channel mix changes night by night.
Mistake 3: No overbooking plan at all
Many independent hotels refuse to overbook because one walked guest once caused a bad review. The result is that they sell 100 percent and arrive at 93 percent after cancellations and no-shows, every high-demand night of the year. A small, controlled overbooking level based on real cancellation data protects far more revenue than it risks. The decision belongs in the same meeting as pricing.
Mistake 4: Discounts that cause cancel and rebook
When a hotel drops its public rate close to arrival, guests holding refundable reservations, or the rebooking services watching for them, cancel and rebook at the lower price. The hotel records a cancellation and a new booking, gross pickup looks healthy, and the actual result is the same guest paying less. If cancellations spike in the days after a rate drop, this is usually why.
Mistake 5: Too many flexible rate plans
Some hotels load four or five flexible plans and only one non-refundable plan, often with a discount so small that no guest sees the point. If your rate structure is crowded, my guide on how many rate plans a hotel needs explains how I simplify it.
My cancellation audit checklist
- Pull 12 months of reservations with created date, cancelled date, stay date, channel, rate plan and room nights.
- Calculate the room-night cancellation rate by channel and rate plan.
- Calculate the cancellation rate by lead time band.
- Measure how many days before arrival each cancellation arrived.
- Check whether cancellation spikes follow public rate drops.
- Compare forecast occupancy 30 days out with actual arrivals for the last 20 high-demand nights.
- Review the non-refundable discount and how many guests choose it on each channel.
- Confirm the overbooking level, if any, is based on data rather than habit.
Bottom line: Most hotel cancellation losses come from reading the forecast wrongly, not from the cancellation policy itself.
What Do Cancellations Really Cost a Hotel?
Cancellations cost a hotel the revenue it fails to recover when a cancelled room is resold late at a lower rate, or not resold at all. The cost is rarely the cancelled booking itself. The real cost is the pricing and selling decisions the hotel made while it believed that room was sold.
The clearest way to show this is a worked example. The numbers below are an example, not a client, and are rounded for readability.
Worked example: one Saturday, two ways of reading it
Take an 80-room independent hotel in 2026. Sixty days before a Saturday, it has 72 rooms on the books at an ADR of $180. Of those, 45 rooms are on flexible OTA rates and 27 are direct or non-refundable. From its own history, the hotel knows flexible OTA rooms on that lead time cancel at about 22 percent and the rest at about 10 percent.
Expected cancellations are 45 times 0.22, which is 9.9 rooms, plus 27 times 0.10, which is 2.7 rooms. That is 12.6 rooms, call it 13. The gross forecast says 72 rooms, or 90 percent. The net forecast says about 59 rooms, or 74 percent.
Version A, priced on gross. The team treats 72 rooms as firm. It raises the rate on the last 8 rooms to $240 and closes discounts. It sells 4 of them. Then 13 of the original rooms cancel, most inside the last two weeks. With little time left, the hotel resells 6 rooms at $150 through last-minute promotions.
- Retained original rooms: 59 times $180 equals $10,620.
- High-rate rooms: 4 times $240 equals $960.
- Last-minute rooms: 6 times $150 equals $900.
- Total: 69 rooms and $12,480. Occupancy 86.3 percent, ADR $180.87, RevPAR $156.00.
Version B, priced on net. The team reads 59 net rooms and 21 rooms still to sell. It keeps selling at $195, a modest increase over the existing ADR, for the remaining weeks. It adds 15 more reservations, and about 3 of those also cancel, so it nets 12 rooms at $195. Close to arrival, it sells 6 more at $150.
- Retained original rooms: 59 times $180 equals $10,620.
- Net new rooms: 12 times $195 equals $2,340.
- Last-minute rooms: 6 times $150 equals $900.
- Total: 77 rooms and $13,860. Occupancy 96.3 percent, ADR $180.00, RevPAR $173.25.
The difference is $1,380 of room revenue on one night, a RevPAR gap of $17.25. Version A had the higher-looking price and the lower result. If a hotel has 40 nights a year where this pattern repeats, the example gap is $55,200 a year, before the extra food, beverage and parking revenue from 8 more occupied rooms each night.
The commission angle
There is a quieter cost too. OTA commission is usually charged on stayed revenue, so a cancelled OTA booking costs no commission. But when the replacement room is sold at the last minute through an OTA promotion, the hotel pays full commission on a discounted rate. That is why I always look at net revenue after distribution costs, not just room revenue. My piece on hotel net RevPAR walks through that calculation.
Bottom line: In this example, reading cancellations correctly was worth $17.25 of RevPAR on one Saturday, with no change to the cancellation policy at all.
How to Reduce Hotel Cancellations in 2026
Hotels reduce cancellations in 2026 by making the committed booking more attractive than the flexible one, not by punishing guests. The levers that work are a meaningful non-refundable discount, deposits on far-out flexible bookings, a stronger direct channel and stable public pricing that removes the reason to cancel and rebook.
Here are the changes I make first, in order.
1. Make the non-refundable rate worth choosing
A non-refundable rate at 3 percent off does nothing. Guests compare the saving with the risk and choose flexible. In most hotels I set the non-refundable discount between 8 and 15 percent, then test it. Booking.com's partner guidance says adding a non-refundable rate plan alongside flexible rates can reduce cancellations by at least 9 percent and increase bookings by at least 5 percent. That is Booking.com's own figure, and it matches the direction I see in practice.
2. Tighten flexibility only where cancellations cluster
The cancellation matrix from the lead time section tells you where the risk sits. If the problem is flexible bookings made more than 90 days out, add a deposit or a longer free-cancellation cut-off for that window only. Do not tighten the policy for bookings made 5 days out, because those guests rarely cancel and a strict policy just costs conversion. Revenuenaire has a detailed breakdown of the break-even math behind cancellation policies if you want to test policy changes properly.
3. Keep public prices stable close to arrival
Every sharp public rate drop close to arrival invites cancel and rebook. Where you need to stimulate demand late, use closed or fenced offers (member rates, packages, length-of-stay offers) rather than cutting the public rate that existing guests can see.
4. Price competitively from the start
Booking.com's partner guidance also warns that unattractive prices can increase cancellations by up to 30 percent, because guests keep shopping after they book. A rate that is clearly out of line with the competitive set is not just a conversion problem. It is a cancellation problem, because the guest who booked it is still looking.
5. Grow the direct channel
Every reservation moved from a flexible OTA booking to a direct booking cuts the expected cancellation rate roughly in half, if the Cloudbeds 2026 figures of 21.8 percent and 10.6 percent hold for your hotel. Direct guests also give you their contact details, which lets you confirm plans, sell add-ons and spot likely cancellations early.
6. Contact far-out flexible guests before the cut-off
A short pre-arrival message two weeks before the free-cancellation deadline, confirming the stay and offering an upgrade or package, does two things. It firms up guests who were going to come anyway, and it brings forward cancellations from guests who were not, which gives you more time to resell.
Bottom line: Reward commitment with a real non-refundable discount and target stricter terms only at the lead times where cancellations cluster.
Building Cancellations Into the Forecast
Building cancellations into the hotel forecast means converting every stay date from gross rooms on the books to expected arrivals, using historical cancellation rates by channel, rate plan and lead time. The expected-arrivals figure, not the gross figure, should drive pricing, overbooking levels and staffing decisions for 2026 and 2027.
The method I use is practical and works from a plain reservation export. For each future stay date:
- Group the reservations on the books by channel, rate plan type (flexible or non-refundable) and lead time band.
- Multiply each group by its historical survival rate, which is one minus its cancellation rate for the remaining time to arrival.
- Add expected no-shows, which I keep separate from cancellations because they behave differently.
- Add expected remaining pickup, net of the cancellations that new pickup will also produce.
- Compare the result with capacity to set the price and the overbooking level.
Setting a safe overbooking level
Overbooking is the direct counterpart of cancellation forecasting. If a sold-out Saturday reliably loses 4 to 6 rooms to cancellations and no-shows inside the last 48 hours, a hotel can safely accept 3 extra reservations and still have a buffer. Going beyond what the data supports creates walked guests and bad reviews. Staying at zero means accepting the 4 to 6 empty rooms every time. I set the level per day of week and season, and I review it after every walk, never after every quiet night.
Review the assumptions monthly
Cancellation behaviour changes. SiteMinder's mid-year 2026 data showed cancellations rising in Asia Pacific and Europe while staying almost flat in the Americas, which tells me no hotel can set its cancellation assumptions once and leave them. I refresh the matrix every month, and every time a channel changes its default policy display.
Bottom line: Forecast expected arrivals, not reservations, and set overbooking from the same cancellation data that drives your pricing.
Frequently Asked Questions
What is a good cancellation rate for a hotel?
A good cancellation rate for a hotel in 2026 is one that is stable and predictable, usually around 15 to 22 percent across all channels. SiteMinder put the 2025 global average at about 19.15 percent. A rate below that is good only if it does not come from policies so strict that they cost bookings.
How do you calculate hotel cancellation rate?
Divide the number of cancelled reservations for a set of stay dates by the total reservations created for those dates, then multiply by 100. For forecasting, use room nights instead of reservation count, and calculate the rate separately for each channel, rate plan and lead time band so the result reflects real risk.
Why are Booking.com cancellations so high for hotels?
Booking.com cancellations are high for many hotels because free cancellation is widely displayed and guests use it to hold several hotels while they keep shopping. D-EDGE's 2018 European data showed Booking Group reservations cancelling at 49.8 percent. A meaningful non-refundable discount and competitive pricing usually bring the rate down.
Do OTA bookings cancel more than direct bookings?
Yes. OTA bookings cancel far more often than direct bookings. Cloudbeds' 2026 State of Independent Hotels report found OTA reservations cancelled at 21.8 percent in 2025, compared with 10.6 percent for direct reservations. Direct guests have usually chosen the hotel already, while many OTA guests are still comparing options.
Should a hotel overbook to cover cancellations?
A hotel should overbook modestly on high-demand nights when its own data shows reliable late cancellations and no-shows. The overbooking level should be smaller than the typical late loss, set by day of week and season, and reviewed after any walked guest. Zero overbooking on sold-out nights usually leaves rooms empty.
When should a hotel hire a revenue management consultant for cancellations and forecasting?
A hotel should hire a revenue management consultant when sold-out forecasts regularly arrive several points lower, or when nobody can show cancellation rates by channel and lead time. Below about 25 rooms with steady demand, an owner can usually manage this alone. Above that, Alaa Elhadi and Revenuenaire build the forecast and pricing around real cancellation data.
Does lowering my rate cause more cancellations?
Lowering a public rate close to arrival can cause more cancellations, because guests on refundable reservations, and price-drop rebooking services acting for them, cancel and rebook at the cheaper price. The hotel records extra activity but earns less from the same guest. Fenced offers stimulate late demand without this effect.
My Verdict
A high hotel cancellation rate is not a disaster in 2026. An unexpected one is. Cancellations near 20 percent are normal, OTA business cancels about twice as often as direct, and far-out flexible bookings are the softest demand you hold. None of that hurts a hotel that forecasts expected arrivals, prices on net pickup and rewards guests who commit. It hurts the hotel that reads 90 percent on the books as 90 percent sold and prices the last rooms accordingly. Build the cancellation matrix, fix the non-refundable offer, and stop cutting public rates close to arrival.
If your sold-out nights keep arriving short, talk to Alaa's team about your forecast and we will show you where the rooms are going.
