10 Hotel Cancellation Statistics Every Hotel Owner Should Know in 2026

Hotel cancellation risk is not one percentage. A reservation can be canceled, modified, canceled after a deadline, marked as a no-show, resold, or left unrecovered. The correct rate changes with the channel, property type, booking window, rate plan, season, and whether the denominator is bookings or room nights.

We reviewed Cloudbeds' 2026 State of Independent Hotels findings and ZentrumHub's Hotel Distribution 2026 report. Cloudbeds says its analysis covers 90 million bookings across tens of thousands of independent properties in 180 countries, measuring 2025 performance. ZentrumHub reports more than 1.5 million aggregated and anonymised bookings across 90 or more OTAs and 15 or more countries, with network activity covering May 2025 to May 2026. The sources provide useful channel and timing observations, but they do not provide one universal hotel no-show rate.

The figures below are a source-led reference for hotel owners, operators, and hospitality writers. The Cloudbeds cancellation rates are platform-based independent-hotel observations, not a universal average for all hotels. The ZentrumHub comparison is a separate vendor network observation. The original sources, definitions, and limitations are listed at the end.

Back to Hotel statistics

The table keeps the required numerator and denominator visible. "Not reported" means the source does not publish a comparable rate. The booking counts below are not room-night denominators unless explicitly stated, so they should not be combined with room-night recovery metrics without a property-level calculation.

OTA bookings made up 63.4% of independent hotel bookings in 2025

Cloudbeds reports that OTAs accounted for 63.4% of independent hotel bookings in 2025, with some markets approaching 80%. This is the most important context for reading the cancellation data: the channel with the higher cancellation rate also represents a large share of the booking volume in the independent-hotel dataset.

The statistic is an OTA share of bookings, not a share of room nights, revenue, or arrivals. It applies to Cloudbeds' independent-hotel benchmark, not to branded hotels or every hotel in every market. An operator should pair channel share with cancellation, no-show, resale, and net-revenue fields before deciding whether a high OTA share is profitable.

OTA cancellations reached 21.8%, versus 10.6% for direct bookings

Cloudbeds reports that 21.8% of OTA bookings were canceled in 2025, compared with 10.6% of direct bookings. The comparison is based on 90 million bookings across 180 countries and tens of thousands of properties in the independent-hotel report. It is a strong channel-specific observation because both figures come from the same published benchmark.

The result does not mean that 21.8% of OTA room nights were lost, or that 10.6% of direct bookings became arrivals. The source release does not publish the associated no-show rate, resale rate, booking-window distribution for each channel, or room-night exposure. Those are the fields a property must add before turning cancellation rate into expected revenue loss.

The channel gap is 11.2 percentage points, or about 2.1 times the direct rate

Subtracting the Cloudbeds rates gives an 11.2 percentage point gap between OTA and direct cancellation rates. Dividing 21.8% by 10.6% gives approximately 2.1 times. These are calculations from the reported rates, not separate Cloudbeds findings.

The calculation is useful for prioritising measurement, but it is not a causal claim about OTAs. Channel mix can be associated with different booking windows, countries, rate plans, traveler types, cancellation policies, and room types. A hotel should therefore reproduce the comparison with its own data using the same stay period, channel definitions, and cancellation cutoff.

Hotel bookings were made 40 days in advance on average, up from 38 days in 2023

Cloudbeds reports that travelers booked an average of 40 days before arrival in 2025, up from 38 days in 2023. A longer booking window gives a hotel more time to price the room, monitor pickup, revise restrictions, and prepare for a cancellation without assuming the room will be resold.

The booking-window statistic should not be mistaken for the cancellation window. Some bookings are canceled soon after they are made, while others remain active until close to arrival. The property should report both the days between booking and arrival and the days between cancellation and arrival, ideally by channel and rate plan.

The average cancellation window grew to 39 days, up from 35 days in 2023

Cloudbeds reports that the average cancellation window expanded to 39 days in 2025, up from 35 days in 2023. This measures when a cancellation occurs relative to arrival, not how many bookings are canceled. It suggests that operators may receive more advance notice on average, even while the overall OTA cancellation rate remains high.

Advance notice creates an opportunity only if the hotel can use it. The useful operating measures are canceled room nights by days before arrival, the share returned to inventory, the time to resale, the final achieved rate, and the share that remains unrecovered. A 39-day average can hide a small group of last-minute cancellations that creates most of the operational pain.

North America led the booking window at 48 days, while EMEA averaged 47 days

Cloudbeds reports average booking windows of 48 days in North America and 47 days in EMEA in 2025. These regional figures are close, but they are longer than the report's global average of 40 days. Regional booking behavior therefore changes how early an operator can act on a cancellation.

The comparison should not become a universal rule for every property in those regions. City hotels, resorts, independent inns, event-driven properties, and different room types can have different booking windows. The right local cut is region, property type, channel, arrival date, length of stay, rate plan, and whether the booking is refundable.

More than two-thirds of bookings were one or two nights, while seven-night bookings rose 25%

Cloudbeds reports that more than two-thirds of bookings were for one or two nights in 2025. It also reports that bookings of seven nights increased 25% year over year. These two findings describe different cancellation exposures: a one-night cancellation has a smaller room-night value, while a longer booking can leave more inventory to replace.

The property should therefore count both canceled bookings and canceled room nights. A one-night reservation and a seven-night reservation are each one booking, but they do not create the same revenue exposure or the same resale problem. Length of stay should be a required field in any hotel cancellation table.

The B2B hotel-API channel was shown at 16.3% in ZentrumHub's 2025 network comparison

ZentrumHub's Hotel Distribution 2026 report presents a 16.3% cancellation figure for the B2B hotel-API channel in its Network 2025 comparison. The report places that figure between direct hotel bookings at 10.6% and OTA bookings at 21.8%, while labeling the Cloudbeds and Network figures separately.

This is useful as a distribution-channel hypothesis, not as a universal industry rate. ZentrumHub's own network is more than 1.5 million aggregated and anonymised bookings across 90 or more OTAs and 15 or more countries, while the Cloudbeds figures come from a 90 million-booking independent-hotel benchmark across 180 countries. The samples, channel definitions, property sets, and periods are not identical, so the figures should remain in separate rows.

Cancellation risk matters more when independent hotel performance is already under pressure

Cloudbeds reports that global occupancy at independent hotels slipped 0.6% year over year in 2025, while ADR declined 5.8% and RevPAR declined 5.4%. In North America, the report says RevPAR declined 4.4%. These are context figures, not cancellation effects.

The connection is operational: when demand or room revenue is under pressure, an unrecovered canceled room night has a larger opportunity cost. A hotel should calculate expected room revenue at risk from canceled and no-show room nights, then subtract revenue from resold rooms and the cost of acquiring or servicing the replacement booking. That is more useful than applying a generic cancellation rate to total annual room revenue.

No universal hotel no-show rate exists, so operators need a room-night recovery model

The public Cloudbeds and ZentrumHub sources reviewed publish cancellation rates, booking windows, and channel comparisons, but not a comparable universal hotel no-show rate. A cancellation is a reservation removed before arrival. A no-show is a reservation that remains active past the property's arrival or check-in cutoff and does not produce an arrival. A late arrival, modification, and canceled-and-resold room must be classified separately.

Use consistent formulas:

• Cancellation rate = canceled reservations divided by eligible reservations.
• No-show rate = no-show reservations divided by reservations still active at the arrival cutoff.
• Canceled room-night exposure = canceled room nights not excluded by the property's policy.
• Recovered-room rate = canceled or no-show room nights resold divided by room nights available for resale.
• Unrecovered room-night rate = canceled or no-show room nights not resold divided by room nights available for resale.
• Expected room revenue at risk = exposed room nights multiplied by expected ADR.
• Net unrecovered revenue = expected room revenue at risk minus resale revenue and directly attributable recovery costs.

The denominator must be fixed before a policy comparison. Report bookings and room nights separately, then cut the results by channel, booking window, rate plan, cancellation deadline, property type, season, and length of stay. This makes the article useful without pretending that a vendor benchmark can answer a property's no-show question.

Sources

• Cloudbeds 2026 State of Independent Hotels report findings, published March 25, 2026. Provides the 90 million booking, 180-country, independent-hotel scope and the 2025 OTA share, cancellation, booking-window, cancellation-window, length-of-stay, occupancy, ADR, and RevPAR figures.
• Cloudbeds 2026 State of Independent Hotels report hub. Provides the report landing page and access to the broader benchmark.
• ZentrumHub Hotel Distribution 2026 benchmark. Provides the report scope, methodology, 1.5 million-plus network booking context, 90-plus OTA coverage, 15-plus country coverage, and the separate network comparison.
• ZentrumHub Hotel Distribution 2026 report PDF. Provides the page-level comparison showing 16.3% for the B2B hotel-API channel, 21.8% for the Cloudbeds OTA figure, and 10.6% for the Cloudbeds direct figure.
• CoStar with STR data and insights. The public benchmark page reviewed describes hotel data coverage but does not provide a comparable universal hotel no-show rate, so no STR no-show percentage is presented here.