What it is
A cancellation removes a booking that was counted in pace and forecasts. The rate can be measured on bookings created in a period (cohort basis) or on stays scheduled in a period (stay basis).
Revenue-based rates matter most for forecasts; count-based rates are easier to explain.
Why operators track it
- Forecast accuracy. Pace and OTB overstate demand if cancellations are high.
- Policy choices. Strict vs flexible policies trade conversion against risk.
- Channel comparison. Channels can differ in cancellation behavior.
- Owner cash. Cancellations after payout create clawbacks.
How it is calculated
Cancellation rate = cancelled bookings ÷ bookings created (or by nights or revenue)
- Cohort basis: bookings created in the period, cancelled at any time so far.
- Stay basis: bookings for stays in the period.
- Cohort rates rise over time as more bookings cancel. Report with an as-of date.
Worked example
Example data The same made-up portfolio is used on every hospitality page: a manager with 120 homes, four of them shown in detail.
The four example homes received 30 bookings created in September 2026. As of October 20, 4 had cancelled: 13.3% by count. They covered 14 of 110 nights (12.7%) and $5,200 of $31,000 revenue (16.8%).
How to read it
- State the basis and as-of date.
- The revenue rate can be much higher when long, high-rate stays cancel.
- Separate guest and host cancellations.
- We do not publish a benchmark. Levels vary by market, season, and property type.
Definitional choices and common pitfalls
- Cohort or stay basis. Pick one and label it.
- Count, nights, or revenue. Show at least two.
- Who cancelled. Guest, host, or channel. Host cancellations are an operations issue.
- Modifications. A date change is not a cancellation.
How different roles use it
| Role | How they use it |
|---|---|
| Revenue manager | Adjusts forecasts for expected cancellations. |
| Finance / trust accounting | Handles refunds and clawbacks on statements. |
| Operations | Tracks host-side cancellations caused by maintenance. |
| Product / engineering lead | Keeps cancelled reservations with dates instead of deleting them. |
Questions it answers
- What share of bookings cancel, by channel and policy?
- How much revenue did we lose?
- Are host cancellations rising?
How to present it in a report
- Rate by count and revenue, by booking month.
- By channel and policy.
- Host vs guest cancellations.
CancellationsBookings created in September 2026, as of Oct 20
By bookings13.3%4 of 30
By revenue16.8%$5,200
Host cancellations1maintenance
| Channel | Created | Cancelled | Rate | Revenue lost |
|---|---|---|---|---|
| OTA A | 14 | 2 | 14.3% | $2,650 |
| OTA B | 7 | 1 | 14.3% | $2,250 |
| OTA C | 4 | 1 | 25.0% | $300 |
| Direct | 5 | 0 | 0.0% | $0 |
Data you need
Fields
| Field | What it means | Example value |
|---|---|---|
| Created date | When booked | 2026-09-04 |
| Cancelled date | When cancelled | 2026-10-02 |
| Cancelled by | Guest, host, channel | guest |
| Policy | Cancellation policy | moderate |
| Refund | Amount refunded | $2,250 |
What the raw data looks like
Example data A few rows as they might come out of the source system.
| Reservation | Created | Cancelled | By | Nights | Rent | Refund |
|---|---|---|---|---|---|---|
| R-88620 | 2026-09-04 | 2026-10-02 | guest | 5 | $2,250 | $2,250 |
| R-88633 | 2026-09-10 | 3 | $600 | |||
| R-88641 | 2026-09-12 | 2026-09-30 | host | 2 | $400 | $400 |
Where this data usually lives
- Property management system (PMS) holds units, the calendar with block types, reservations, rates, fees, and often owner statements and trust accounting. For example: Guesty, Hostaway, OwnerRez, Track, Streamline, and Escapia. Some larger managers run their own in-house PMS.
- OTAs and booking channels (Airbnb, Vrbo, Booking.com, and partner distribution sites) hold payout reports, host fees, reviews, and cancellations. Payouts are usually net of host fees.
The same field can live in more than one system. Agree with the customer which one is the source of truth.
Common data gaps and traps
- Cancelled bookings deleted.
- Cancellation reason missing.
- Rebooked stays created as new bookings.
Related metrics
FAQ
- Cohort or stay basis?
- Cohort for channel and policy comparison; stay basis for monthly reporting.
- Do date changes count?
- No, unless the stay is cancelled and rebooked.
- What is a normal cancellation rate?
- We do not publish a benchmark. Levels vary by market, season, and property type.
For your data teamFormula variants, tables, SQL sketch, and dashboard spec
Formula details
cancel_rate_count = COUNT(cancelled) / COUNT(created)
cancel_rate_rev = SUM(rent of cancelled) / SUM(rent of created)
Worked example detail
| Basis | Cancelled | Created | Rate |
|---|---|---|---|
| Bookings | 4 | 30 | 13.3% |
| Nights | 14 | 110 | 12.7% |
| Revenue | $5,200 | $31,000 | 16.8% |
Tables and fields
| Table | Fields | Notes |
|---|---|---|
reservations |
reservation_id, created_at, cancelled_at, cancelled_by, nights, rent_total, policy, channel |
SQL sketch
SELECT COUNT(*) FILTER (WHERE cancelled_at IS NOT NULL)::numeric / COUNT(*) AS rate_count,
SUM(rent_total) FILTER (WHERE cancelled_at IS NOT NULL) / SUM(rent_total) AS rate_revenue
FROM reservations
WHERE created_at >= '2026-09-01' AND created_at < '2026-10-01';
Build spec
Copy this into a ticket. It lists the grain, inputs, logic, and edge cases.
metric: hospitality.cancellation_rate
grain: one row per reservation
inputs:
- created_at
- cancelled_at
- cancelled_by
- nights
- rent
logic:
- rate by count, nights, revenue
- cohort by created month, with as-of date
edge_cases:
- rebooked: link and exclude
- host cancellation: separate
- partial cancellation: nights basis
As a dashboard card
- Rates by booking month with count and revenue lines, channel filter.
Common mistakes
- Deleting cancelled bookings.
- Comparing young and old cohorts.
- Counting modifications as cancellations.