Is your hotel overselling, and is it using the best overbooking strategy it could? Two of the possible answers cost money. A hotel that sells exactly its capacity and then takes the no-shows and the last day cancellations has chosen an overbooking level of zero, and pays for it in empty rooms on the nights it could have sold twice. A hotel that oversells by feel pays for it the other way: a guest with a confirmed reservation standing at the desk with no room, a taxi to a competitor, and a review that outlives the night.
Overbooking is not a bet. It is a decision with numbers, and the numbers are yours: your no-shows, your rate, your costs, your arrivals for that night. This process gives you the level, the timing, and the protocol for the day it goes wrong. It is written for an independent hotel working from a PMS and a spreadsheet.
Who this is for
Revenue managers, front office managers and hotel directors at independent hotels that today oversell by intuition, or that do not oversell at all because the last walk was painful. It applies to any property with high occupancy nights: long weekends, city events, high season.
What you need before you start
- Your no-show history: for each arrival date, how many guaranteed reservations did not arrive. Thirty comparable dates are enough to start; separate weekdays from weekends and event nights from normal nights.
- Your late cancellations: the ones that arrive inside the last 48 to 72 hours, by channel. A cancellation at 6 pm frees a room you can no longer sell; it behaves like a no-show.
- ADR (average daily rate) and the variable cost of an occupied room: cleaning, amenities, breakfast if included, commission if the booking came through an agency.
- The real cost of walking a guest: the room at the partner hotel, transport, a meal, and an honest figure for the goodwill you lose. Hotels that have measured it put the total well above the room rate.
- The share of guaranteed versus non-guaranteed reservations in your arrivals, and your room types with their upgrade path.
The process, step by step
Step 1. Measure your no-shows honestly
Count, for each of the last thirty comparable arrival dates, how many reservations did not show. Do not use a single average: the useful figure is the distribution. How many nights had zero no-shows, how many had one, two, three, four or more. Write it as a table.
Keep two things in mind while you count. No-shows depend on arrivals: a night with ten arrivals cannot produce eight no-shows, so a busy night and a quiet night are different samples. And a no-show on a three night stay frees three nights, not one; count the room nights it released, not only the arrival.
If your history is thin, start with the guarantee policy. Public research on US hotels found that requiring a credit card guarantee cut no-show rates from a range of 10% to 15% down to around 4% (Toh and DeKay, Cornell Hotel and Restaurant Administration Quarterly, 2002). If a large share of your reservations are non-guaranteed, your first overbooking decision is to change that.
Step 2. Put a cost on an empty room and on a walk
Two costs drive the whole calculation.
The cost of an empty room is the revenue you lose when a room stays unsold because of a no-show: ADR minus the variable cost of occupying it. With an ADR of 150 and a variable cost of 30, an empty room costs you 120.
The cost of a walk is what you pay when a guest with a reservation cannot be accommodated: the room at the alternative hotel, transport, a meal, the staff time, and the goodwill. A worked example: room and tax at the partner hotel 140, taxi 20, breakfast 20, and 150 for the goodwill you lose, which is a conservative figure for a guest who tells the story. Total 330. Many hotels use a rule of thumb of two to three times the cost of an empty room. Use your own numbers, and revisit them after every walk.
Step 3. Calculate the overbooking ratio
The overbooking ratio tells you how sure you need to be that a no-show will happen before you sell one more room than you have. It is the cost of a walk divided by the sum of the cost of a walk and the cost of an empty room.
With the numbers above: 330 divided by (330 plus 120) equals 0.73. The more a walk costs you relative to an empty room, the higher the ratio and the more cautious your overbooking. If your walk cost rises to 500 because your partner hotel is expensive, the ratio becomes 500 divided by 620, or 0.81, and you oversell less.
Step 4. Read the ratio against your no-show distribution
Turn your table from step 1 into cumulative probabilities: the share of nights with one or more no-shows, two or more, three or more, and so on. With thirty nights where six had zero no-shows, nine had one, eight had two, five had three and two had four or more, the cumulative table is: one or more 80%, two or more 50%, three or more 23%, four or more 7%.
Now read it from the bottom up and stop at the first line that equals or exceeds your ratio of 0.73. Four or more, 7%, no. Three or more, 23%, no. Two or more, 50%, no. One or more, 80%, yes. Your overbooking level for that type of night is one room. That is the number of rooms you can sell beyond capacity while the expected cost of walks stays below the expected cost of empty rooms.
If your walk cost were lower, say 150, the ratio would drop to 0.56, the first line to clear it would still be one or more, and the level stays at one. If it fell further, two rooms. The method is not perfect on any single night; it gives you the best result over the year.
Step 5. Adjust the level night by night
The ratio gives you a level per type of night, not a fixed number for the hotel. Adjust it with what you know about the specific date.
- Arrivals: the level is a share of arrivals, not of capacity. A night with few arrivals and many stayovers has few no-shows to absorb.
- Guarantee: reservations guaranteed by card or prepaid rarely fail; non-guaranteed ones are where the no-shows live.
- Lead time and channel: long lead time bookings cancel more; late direct bookings show up. Look at where your no-shows actually come from.
- Event nights: guests who paid a non-refundable rate for a concert or a fair arrive. Lower your level on those nights, and never oversell them by feel.
- Room type: oversell only the categories with an upgrade path above them. Overselling the top category leaves you with a walk, not an upgrade.
Step 6. Decide when, not only how much
Overbooking is set by night and revised every day from seven days out, with the cancellations and the pickup of the last 24 hours. Two rules protect you.
Do not open the overbooking for the first time on the day before or on the day of arrival. Last minute overselling has no time to be absorbed by cancellations and produces walks.
On the arrival day, run the count in the afternoon. Confirm the arrival time of every guaranteed reservation that has not checked in by 4 pm, release the non-guaranteed ones after their hold time, and close sales the moment confirmed arrivals reach capacity. From that point, the desk manages the night, not the channel manager.
Step 7. Run the walk protocol when the night goes over
Some nights you will still be over. The difference between a walk that costs 330 and one that costs a regular guest is preparation.
- Decide before the guest arrives who is walked. Priority goes to one night stays booked through third parties with late arrival times, never to a regular, a group leader, a family with children or a guest with a special need.
- Have the partner hotel agreed in advance, same category or better, with the room already reserved and paid by you, and the transport arranged.
- Call the guest before arrival when you know in time; nobody should learn at the desk.
- A manager delivers the apology, in person if possible, and offers a reason to return: the first night on the next stay or an upgrade, not a discount.
- Log every walk with its full cost and the reservation that caused it, and feed it back into step 2. The ratio only stays honest if the walk cost is real.
Situations you will meet, and what to do
- A long weekend at 98% on the books on Thursday afternoon. Count the arrivals for Friday, apply the level for weekend nights, confirm every non-guaranteed reservation, and keep the top category closed to oversell.
- A group with a history of wash. A group that blocks 40 rooms and consumes 32 is a no-show factor of its own. Cut the block at the cut-off date, and do not oversell on top of an unreleased block.
- A last minute cancellation at 7 pm on a full night. That room will not sell tonight. Your overbooking level exists to cover exactly this; if you are at zero, the room stays empty.
- Two walks in one week. Stop and check the walk cost you are using. If the real cost is higher, the ratio rises and the level drops by itself.
The three mistakes that ruin overbooking
- Overselling on a fixed number, the same for a Tuesday in February and a concert Saturday, without looking at the arrivals or the guarantee of that night.
- Overselling the top category, or overselling by room count without checking that the upgrade path exists.
- Never logging the walks, so the walk cost in the calculation stays a guess and the level never learns.
How many rooms should a hotel overbook?
As many as your own no-shows justify, and no more. Count the no-shows of your last thirty comparable arrival dates and build the distribution: how many nights had one or more, two or more, three or more. Calculate the cost of an empty room (ADR minus variable cost) and the cost of a walk (the alternative hotel, transport, a meal and the goodwill lost). The overbooking ratio is the walk cost divided by the walk cost plus the empty room cost; with a 330 walk and a 120 empty room it is 0.73. Read your cumulative no-show table from the bottom and stop at the first line that reaches the ratio: if 80% of nights had one or more no-shows and 50% had two or more, you oversell by one room. Adjust that level by night for arrivals, guarantee and events, oversell only categories with an upgrade above them, never open it on the arrival day, and log every walk at its real cost.
Check your work
You have done this well when the no-show table is refreshed every month, when the overbooking level is written per night in the pickup sheet and revised daily from seven days out, when the walk cost in the calculation matches the last real walk in the log, when the top category is never oversold, and when a full night with a 7 pm cancellation still closes at 100%.
This content was created with the help of artificial intelligence and reviewed by Rafael Osborne (Profit Guest Services) before publishing.
