Overbooking
In hotel revenue management, overbooking is the deliberate practice of accepting more reservations than you have rooms, on the expectation that cancellations and no-shows will bring arrivals back within capacity.
It is standard practice at scale and a much harder trade at 46 rooms, for reasons that have nothing to do with courage and everything to do with arithmetic.
See how ampliphi prices for independent hotels
Get a free revenue audit · Book a demo
What is overbooking?
Selling rooms you do not have, on purpose, because experience says some of them will come back.
The logic starts from a fact about on the books: it overstates what will actually arrive. Reservations cancel, guests fail to show, and a property that sells exactly to capacity will finish below it almost every time. Overbooking is the attempt to reclaim those rooms.
There are two versions and confusing them is common. Deliberate overbooking is a decision, taken with a number attached, based on your own attrition history. Accidental overbooking is a systems failure, usually a sync gap between channels selling the same last room, and it is not a strategy at all.
The mechanics are simple enough. You forecast arrivals as on the books plus expected pickup minus expected attrition, compare that to capacity, and sell the difference. The difficulty is that attrition is a probability and capacity is a hard wall. Being right on average does not protect you on the night you are wrong.
Related to it are the terms that describe the failure. A walk is relocating a guest to another property at your cost. Denied service is the general category, and it is the outcome every overbooking policy exists to avoid.
Resources: Cancellation rate · Demand forecasting · Occupancy rate
How overbooking works in practice
The calculation is easy. The judgment about whether to act on it is not.
Worked example. A 46-room inn has all 46 rooms on the books for a Saturday, seven days out. Its history says roughly 8 percent of reservations at that lead time cancel or fail to arrive, so it expects about 42 guests and four empty rooms worth around $205 each. Selling three extra reservations would recover about $615. The risk sits on the other side: if attrition lands at 2 percent rather than 8 percent, the property is two rooms short and walking two guests, at a real cost commonly two to three times the room rate once alternative accommodation, transport and a refund are counted.
So the trade is roughly $615 of expected upside against a low-probability loss of $900 or more, plus a public review written by somebody who was turned away at 10pm.
That review is the part the arithmetic misses. At a 46-room independent, your rating is your distribution. A chain property absorbs one bad review into thousands. A small property carries it on the first page of its OTA listings for a year, where it is read by every guest deciding between you and the hotel next door.
The everyday version is more useful than the dramatic one. Most of the value at a small property comes not from overselling but from managing the attrition itself: confirming arrivals on tight dates, holding a card guarantee, and knowing which reservations are least likely to materialise.
Resources: On the books · Rate calendar
Why overbooking matters for independent hotels
Because the maths that makes it obviously correct at 300 rooms makes it questionable at 46.
At a large property, attrition is a stable percentage across a big population, the variance around the mean is small relative to capacity, and a walk can often be handled by a sister property down the road. Overselling by five rooms out of 300 is a rounding decision.
At 46 rooms the same percentage produces a handful of rooms, and a handful has real variance. The difference between 8 percent attrition and 2 percent on a single night is not a statistical footnote, it is two families in your lobby with nowhere to go. You also have no sister property, so a walk means calling competitors during their own busy evening.
None of which means the concept is useless to you. It means the profitable version at your size is usually the conservative one: understand your cancellation rate, protect against accidental oversells, and take one or two rooms of deliberate risk only on dates where your history is strong and consistent.
Resources: Revenue management for independent hotels
How to handle overbooking at your property
- Fix accidental oversells first. Know your channel sync speed and hold a buffer on dates that regularly sell out.
- Calculate attrition from your own data, split by lead time and channel. An industry average will not describe your property.
- Start at one room, not five. If your history supports more, you will see it within a season.
- Only oversell where your history is consistent. A date type with wide variance is exactly the wrong place to take the risk.
- Write a walk policy before you need one. Which property you call, what you pay, what you refund, who decides.
- Choose who gets walked on principle, not in the moment. Shortest stay, lowest rate, no loyalty history, arriving latest. Never a returning guest.
What overbooking will not tell you
The expected value calculation says nothing about the shape of the risk. A positive expected value made of small frequent gains and rare large losses is a different proposition from one made of steady gains, and the arithmetic treats them identically.
It cannot price the review. The direct cost of a walk is knowable and the indirect cost is not, which means every published overbooking model understates the downside for a property whose bookings depend on a rating.
And it assumes your attrition history predicts tonight. A date with an unusual mix, a first-time event nearby, or weather that keeps people home is a date where your normal percentage does not apply, and those are disproportionately the dates you are tempted to oversell.
How ampliphi approaches overbooking
Ampliphi prices rooms. It does not manage inventory, it does not set overbooking levels, and it does not decide how many reservations you accept above capacity. That decision stays with you and with your PMS.
Where the two touch is upstream. The everyday rate suggestion is demand-based, built on booking pace and occupancy, and its purpose is to sell the rooms you have at the right price rather than to sell rooms you do not have. A date priced correctly through its booking curve is less likely to arrive at the final week with the property sold out and the owner wondering whether to risk two more.
The suggestion covers your base rate and the differential between room types. You approve every rate before it publishes, and ampliphi runs on top of the PMS you already use.
Key takeaways: overbooking
- Deliberately accepting more reservations than you have rooms, expecting cancellations and no-shows to close the gap.
- Deliberate overbooking is a decision with a number behind it. Accidental overbooking is a systems failure.
- The arithmetic is expected attrition against capacity. The risk is that attrition is a probability and capacity is a wall.
- At 46 rooms the variance is large relative to capacity, which makes the trade much less attractive than at 300.
- A walk costs two to three times the room rate in direct cost, plus a review that sits on your listing for a year.
- The reliable gain at small properties is managing attrition, not overselling.
Frequently asked questions about overbooking
Should a small independent hotel overbook at all?
Cautiously, on a small number of dates, and only once you have measured your own attrition.
The case against is real at your size. Variance is large relative to capacity, you have no sister property to walk guests to, and a single bad review carries far more weight on a small property's listing than on a chain's. Many well-run independents choose not to oversell and lose very little by it.
The case for is that selling exactly to capacity means finishing below it almost every night. If your history on a specific date type is consistent across two years, one room of deliberate risk is a defensible decision. Five is not, and neither is any number chosen without the data behind it.
How do I calculate how many rooms to overbook?
Forecast arrivals, compare to capacity, and sell the gap, then cut the answer down for variance.
Arrivals are on the books, plus expected pickup between now and arrival, minus expected cancellations and no-shows at that lead time. If your capacity is 46 and your forecast arrivals are 42, the arithmetic says four rooms.
Then apply judgment the arithmetic does not contain. Four is the average outcome, not the floor, and at a small property the spread around that average is wide enough to produce a night with zero cancellations. Most independents should take a fraction of what the calculation offers.
What does it actually cost to walk a guest?
Commonly two to three times the room rate in direct cost, plus something you cannot put a number on.
The direct side is the room you buy at another property, often at short notice and at a rate you do not control, transport to get the guest there, a refund or waiver on your own charge, and frequently a goodwill gesture for a future stay.
The indirect side is the review, the guest who does not return, and whatever they tell the people they were travelling to meet. On a 46-room property with a few hundred reviews, one detailed account of being turned away late at night does measurable damage to conversion for a long time.
What is the difference between overbooking and an oversell?
Overbooking is the decision. An oversell is the situation, whether or not anyone chose it.
Deliberate overbooking means you sold above capacity with a calculation behind it and accepted a known risk. An accidental oversell means two channels sold the same last room inside a sync gap, or a mapping error left availability wrong, and nobody decided anything.
The distinction matters because the fixes are completely different. Deliberate overbooking is refined by better attrition data. Accidental oversells are fixed in your channel manager and by holding a buffer on tight dates.
Who should I walk if I have to?
Decide the rule in advance, because the moment itself is the worst time to be making policy.
The conventional order is the shortest stay, the lowest-rated booking, a guest with no history at the property, and the latest arrival, because walking a one-night guest disrupts less than walking someone three nights into a week. Never walk a returning guest and never walk a guest you have already confirmed personally.
Whoever it is, handle it well. Pay for the room, pay for the transport, refund the night, call ahead so they are expected, and follow up afterwards. A walk handled generously sometimes survives without a review. A walk handled defensively never does.
Related terms
Cancellation rate
The share of reservations cancelled before arrival. It is the input the whole overbooking calculation rests on, and it needs splitting by lead time and channel to be usable.
On the books
The count of rooms currently reserved for a future date. It reliably overstates arrivals, which is the observation overbooking is built on.
Demand forecasting
The prediction of how a future date will finish. Overbooking is one decision that falls out of a forecast once expected attrition is part of the arithmetic.
Occupancy rate
The percentage of available rooms sold. Overbooking exists to close the gap between the occupancy you sold and the occupancy that actually arrived.
Channel manager
The software distributing availability across channels. Sync gaps in it are the usual cause of an oversell nobody chose.