How Many Rooms Can You Safely Oversell?

A note for revenue managers, GMs, and owners — for the season when holiday and New Year bookings start building, and cancellations build with them.

Two hotels on the same street, both full on paper for New Year’s Eve.

The first never oversells. On the night, four rooms sit empty: two cancellations that came in after the deadline and two guests who simply didn’t arrive. Nobody notices, because empty rooms don’t complain.

The second oversold by six, based on a feeling that “there are always some cancellations.” Five didn’t show. One did, at 11pm, and got sent to a hotel across town with an apology, a taxi, and a review that will still be visible in three years.

Both hotels guessed. One got lucky. The number is measurable, and this article is about measuring it — including the perfectly respectable case where the answer is zero.

Both mistakes cost money, but not the same way

The awkward thing about overbooking is that the safe-looking option isn’t free.

A hotel room is perishable inventory: the night expires. A room empty on New Year’s Eve because of a no-show is revenue you will never recover — and if you turned away a booking earlier that week for that date, you also created a denial you didn’t need to.

The other way costs more, but less often. Turning a guest away — “walking” them — costs you the alternative hotel, the transfer, whatever compensation you offer, the rest of that guest’s stay, and a public review written by someone who was told at 11pm that their reservation didn’t exist.

So this is a decision under uncertainty with asymmetric costs, which is exactly the shape of problem where instinct performs worst — and where a little measurement goes a long way.

Step one: measure what actually happens

You need one number per situation, and your PMS already contains it. For at least the last twelve months, for each arrival date, compare rooms on the books at the start of the arrival day against rooms actually occupied. The gap is your late cancellations plus no-shows.

Then break it down, because a single hotel-wide average is nearly useless:

• By segment. Corporate travelers with flexible rates cancel and no-show at very different rates from leisure guests on a non-refundable rate, and both differ from an allotment that releases automatically. This is usually the biggest driver, and it’s the one people skip.
• By day of week and season. A Tuesday in November behaves nothing like December 31.
• By lead time. Bookings made nine months out behave differently from ones made yesterday.
• By channel. Some channels carry systematically higher cancellation behavior — measure yours rather than trusting the folklore.

What you’re building is a small table: for this kind of date, with this mix, roughly this share of the booked rooms historically didn’t materialize. Twelve months is a minimum; two or three years is better, and if your mix changed substantially in that period, weight the recent data more.

A caution that matters more than any formula: this only works where the numbers are big enough to be stable. A 30-room hotel with four no-shows across a year has anecdotes, not a distribution. Which brings us to the most important section of this article.

Step two: price the cost of turning someone away

Do this before you set any limit, because it converts a scary abstraction into a number you can compare against.

Add up, for your hotel:

• The cost of the alternative room — often above your own rate, because you’re buying it at the last minute.
• Transport, and whatever compensation you offer (a voucher, a future free night, an upgrade on return).
• The remaining nights of that reservation, if the guest doesn’t come back.
• The lifetime value of a guest who won’t return, and — hardest to quantify, most persistent — the review.

Whatever your property’s total comes to, it will be a multiple of a room night rather than a fraction of one — and the multiple depends heavily on how you value the review and the lost relationship, which is why it has to be your number rather than a benchmark. Run it once and write it on the wall.

Now the trade-off is concrete. Say your arithmetic lands on six room nights per walk, against one room night for an empty room. Then overselling only makes sense where you’re confident of avoiding at least six empty rooms for every walk you cause. That’s a much higher bar than “there are usually some cancellations,” and it’s why a sober overbooking limit tends to be smaller than instinct suggests.

Step three: set the limit deliberately — or set it to zero

The mechanical version: for a given date, take your expected late cancellations and no-shows for that date’s mix, then hold back a safety margin. What remains is your overbooking limit.

Sensible practice, stated plainly:

• Be conservative at the top. The dates where overselling is most tempting (New Year, a trade fair, a festival) are exactly the dates where the whole city is full and there is nowhere to walk anyone to. The cost of being wrong peaks on precisely the nights the model says to be aggressive.
• Keep room types in mind. Your standard doubles are a pool; your two suites are not. Overselling a unique room type is a different and much worse bet.
• Watch the day itself. An overbooking limit set three weeks out should be revisited on the arrival day — pickup, early check-ins, and same-day cancellations all move the picture.
• Write down the decision and the outcome. How many you oversold, how many didn’t show, whether anyone was walked. Three months of that log teaches you more than any published benchmark, because it’s your hotel.

And the case for zero, which is legitimate and common: if you’re small, if your no-show data is thin, if your rooms are heterogeneous, if there’s no comparable hotel nearby to walk to, or if your brand promise makes a walked guest unusually damaging — don’t oversell. Use the tools in the next section instead.

The alternatives, which are often the better answer

Overbooking manages the symptom. These reduce the underlying uncertainty, and most independent hotels get more out of them:

• Deposit and guarantee policy on high-risk dates. A prepayment requirement for New Year’s Eve removes most of the no-show risk in a way overbooking never can.
• Non-refundable rate share. A modest discount for commitment converts uncertain inventory into certain inventory, and lets you sell the risky remainder with confidence.
• A cancellation deadline that leaves you time to resell. A 6pm same-day deadline on a peak date is a policy that hands you the problem too late to solve it.
• Release dates on group blocks and allotments. Contracted rooms nobody claims are the most predictable “cancellation” you have — the wash factor is the formal name, and unlike no-shows it’s negotiable in advance.
• A waitlist. Cheap, ancient, and effective: when a cancellation lands, you have someone to call.

None of these are exciting. All of them reduce the number of nights where the overbooking question even comes up.

(A note on our own product, since it usually appears here: Peaqplus doesn’t set overbooking limits and doesn’t have an overbooking module — this is a decision we think belongs to the hotel. What it does provide is the raw material: every imported PMS state is kept as a dated snapshot, so the difference between what was on the books on the morning of arrival and what was actually occupied is visible historically, by segment and by date, rather than reconstructed from memory. The table in step one is the goal; a PMS export and a spreadsheet will also get you there.)

Frequently asked questions

Is overbooking legal, and is it fair to guests?

Overselling is used in hospitality, but the remedy when you cannot honor a confirmed reservation depends on the booking and the jurisdiction. In the EU, the European Commission notes that a stand-alone accommodation booking depends on national contract law and the agreed terms; the Package Travel Directive applies only when the accommodation forms part of a qualifying package. Your contract, platform rules, or local law may require a refund, replacement accommodation, or related costs, but that is not one universal hotel rule. Check the rules for your country and each distribution contract before setting policy. Fairness starts with treating a walk as a last resort and having a clear guest-care plan — who finds a comparable room, pays for transport, and communicates with the guest — before it happens.

We’re a 25-room hotel. Should we oversell at all?

Usually not, and that’s a defensible strategy rather than a missed opportunity. Small properties have too few arrivals for cancellation rates to be statistically stable, fewer comparable rooms to shuffle, and a single walked guest can wipe out a month of overbooking gains. Put the same effort into deposit policy, cancellation deadlines, and a waitlist — you’ll capture most of the benefit with none of the tail risk.

What’s a normal no-show rate?

There isn’t a usable industry number, and any figure quoted without a segment, a market, and a rate mix should be treated as decoration. No-show behavior depends heavily on your channel mix, your guarantee policy, and your guest profile — a corporate-heavy city hotel with flexible rates and a resort selling mostly prepaid stays live in different worlds. Measure your own from your own history; that’s the only number you can act on.

Where to go from here

The glossary covers the vocabulary: overbooking, cancellation rate, no-show, denial, and wash factor. For the forecasting side of the same uncertainty, read your forecast is always wrong; for the segment differences that drive cancellation behavior, types of hotel guests. The free Peaqplus Academy covers on-the-books analysis from the ground up.

Overbooking isn’t courage and it isn’t recklessness. It’s a number, and the hotels that do it well are the ones that know what it costs them to be wrong.