Your no show rate: measure it before you fix it

A no show rate is one of those numbers everyone in hospitality talks about and few measure properly. Ask a manager how bad no shows are and you will usually get a story about last Saturday rather than a figure. Ask for the figure and you will often get one that mixes no shows with cancellations, counts tables instead of guests, and averages a busy Friday with a quiet Tuesday until both disappear.

That matters, because almost every decision about no shows depends on the number. Whether to send a second reminder, whether to hold back tables for walk-ins, whether to ask for deposits, whether to overbook: each of these is a trade-off, and you cannot weigh a trade-off without knowing what you are trading against. A badly measured rate leads to measures that are either too weak to help or too strong for the guests who were always going to come.

This guide explains how to calculate your no show rate so that it means something. It covers what counts as a no show, why cancellations must be kept apart, why covers are better than tables, how to split the rate by service, how to read it over time, and the point at which no shows stop being an ordinary cost and start being a problem worth fixing.

A no show is a confirmed booking where the guests neither arrived nor cancelled. That definition has three parts, and each one matters.

The booking was confirmed. A provisional enquiry, a booking still awaiting verification, or a waitlist request that was never converted is not a no show, because the guest was never promised a table.

The guests did not arrive. A party that arrives forty minutes late is a late arrival, not a no show, even if it causes problems. Decide in advance how long you hold a table, and record a booking as a no show only after that window has passed.

The guests did not cancel. This is the part most often blurred. A guest who cancels, even at the last minute, has told you the table is free. That is a different event with a different fix.

Cancellations and no shows look similar in the diary, because both leave a table empty. They are not the same problem.

A cancellation gives you information and time. Even a late cancellation, an hour before the booking, lets you offer the table to the waitlist or a walk-in. A no show gives you nothing until the hold window has passed, by which time the best moment to resell the table may have gone.

Their fixes are also opposite. Most measures that reduce no shows, such as easy online cancellation and reminders with a cancellation link, increase cancellations, because they turn guests who would have silently stayed away into guests who tell you. If you measure the two together, a successful change can look like a failure: no shows fall, cancellations rise, and the combined number barely moves.

So track them separately. Measure the no show rate on its own, and measure last minute cancellations alongside it, defined by a cut-off that suits your venue, such as cancellations made within a few hours of the booking.

You can calculate the rate by bookings or by covers. Both are useful, but covers usually tell you more.

A rate by bookings treats a no show of two guests the same as a no show of twelve. A rate by covers weighs each no show by the number of seats it left empty, which is closer to what it actually cost you. If large parties no show more often than small ones, or the reverse, a booking-based rate will hide it.

Calculate both if you can, and look at the cover-based rate first.

The calculation itself is simple:

Notice the denominator. It excludes cancelled covers, because a cancelled booking was never expected to arrive. Including cancellations in the denominator makes the no show rate look smaller than it is, and the more successful you are at converting no shows into cancellations, the more distorted it becomes.

A matching formula for late cancellations:

And it is useful to track the seated rate too, which is the share of booked covers who actually sat down:

The three together tell the full story. A good change to your booking process should lower the no show rate, may raise the late cancellation rate, and should hold or raise the seated rate.

An overall no show rate hides most of what matters. The problem is almost always concentrated: on particular days, at particular times, for particular party sizes, or from particular booking channels.

Split the rate in at least these ways:

Each split points at a different fix. A high rate on bookings made far in advance suggests a reminder closer to the date. A high rate from one booking channel suggests that channel is producing low-commitment bookings. A high rate for large parties suggests a deposit or a confirmation call for those parties only.

Here is an illustrative month for a restaurant with roughly 3,000 booked covers. The figures show how the rates are built, not what is typical.

The no show rate is 138 divided by 2,760, which is 5.0 percent. The late cancellation rate is 90 divided by 3,000, or 3.0 percent. The seated rate is 2,622 divided by 3,000, or 87.4 percent.

Now split the no shows by service. Suppose 96 of the 138 came from Friday and Saturday dinners, which held 1,100 of the 2,760 expected covers. The weekend dinner no show rate is about 8.7 percent, while the rest of the week is about 2.5 percent. The overall 5 percent was hiding a weekend problem and a weekday non-problem. Only the weekend dinners need stronger measures.

Every venue that takes bookings has some no shows, and a small, steady rate is part of the cost of running a diary. The question is when the rate has become a problem worth acting on.

There is no universal threshold, but there are useful signals. A rate that is rising month after month is a problem even if it is still low. A rate on one service that is several times the rest of the week is a problem for that service. And a rate high enough that an empty table at peak is a regular sight, rather than an occasional surprise, is costing you money you could recover.

The surest way to decide is to put a value on it. Multiply the no show covers on a service by the average spend per head you would have taken, and you have an approximate cost. Our guide to what a no show costs works through the sum in detail, including the difference between a table you could have resold and one you could not.

A rate tells you how often no shows happen; a cost tells you how much they matter. To turn one into the other, multiply the number of no show covers on a service by the average spend per head on that service. Then ask a second question: could the table have been resold if you had known in time? On a quiet weekday, often not, so the real loss is the preparation and staffing for guests who did not come. On a full Saturday, usually yes, so the loss is the full spend of the guests you could have seated instead.

That second question is why the same rate can be harmless on one service and expensive on another. A 5 percent rate on a half-empty Tuesday costs little. A 5 percent rate on a fully booked Saturday, with a waitlist of guests you turned away, costs a great deal.

Lead time is the gap between when a booking was made and when it was for. It is one of the most useful splits, and one of the least used. Bookings made on the day tend to behave differently from bookings made weeks in advance, because the longer the gap, the more chance plans have of changing.

Group your bookings into a few bands, such as same day, within a week, one to four weeks, and more than four weeks, and calculate the no show rate for each. If the rate climbs steeply with lead time, the fix is usually a reminder closer to the date and an easy way to change or cancel, rather than any stricter policy.

The people taking bookings and running the door affect the no show rate more than anyone. Share the weekly rate with them, by service, and explain what moves it: verified contact details, clear confirmations read back on the phone, reminders that go out on time, and a waitlist kept alive through the evening. When the team can see the number fall after a change they made, they keep doing it.

A key performance indicator only changes behaviour when people know it, understand it and can influence it, which is the general lesson of any performance indicator a business tracks. The no show rate is no different.

A single month can be misleading. Holidays, weather, local events and school terms all move no shows around. Compare each service with the same service in previous months, and look at the trend over a quarter rather than reacting to one bad weekend.

When you make a change, such as adding reminders or introducing a cancellation link, note the date and watch the next two months. A successful change usually shows up as a falling no show rate and a rising late cancellation rate within a few weeks.

Calculating these rates by hand from a paper diary is possible but slow, and it rarely happens consistently. Reservation reports that record every outcome automatically, and split them by service, source and party size, make the numbers routine. Whatever system you use, check that it records cancellations and no shows as separate outcomes, lets you choose the time window, and shows the rates by meal period. A report that only counts arrivals cannot give you a no show rate at all.

With the rate measured properly, the response becomes proportionate. Services with a low rate need nothing more than good confirmations and a reminder. Services with a higher rate need a reminder closer to the booking, an easy cancellation link and a live waitlist. Only the services where the rate is high and a no show is expensive need deposits or a stricter cancellation policy. Our guide on how to reduce no shows without annoying guests sets out those levers in order.

Once a week, look at the previous week's rates by service. Note any service that is well above its normal level, and check whether something explains it: a local event, a holiday, a large party that did not arrive. Once a month, compare each service with the same service in previous months and decide whether any service needs a different set of measures. The routine takes minutes once the numbers are recorded automatically, and it keeps no shows a managed cost rather than a recurring surprise.

The booking source split deserves a little more attention, because it is where many venues find their biggest surprise. Bookings made directly on your own page, by phone, through a third-party platform and through a concierge or hotel desk can behave very differently. A source that produces a high rate is telling you something about how much the booking means to the guest at the moment it is made. Look at the confirmation and reminder each source sends, whether its contact details are verified, and whether guests booking through it can cancel easily. Sometimes the fix is not stronger policy but a better process for that one source.

In servearound, every booking records its outcome: seated, cancelled or a no show, with walk-ins counted in or left out as you choose. The reservations report in Analytics shows covers and bookings, the no show, cancellation and seated rates, and the median lead time, split by meal period, by where the booking came from and by party size, with every figure set against the period before. Small samples are flagged, so a rate built on a handful of bookings does not mislead you.

On the booking side, the reservations module verifies guests' email addresses before a booking is confirmed, sends a reminder as many hours ahead as you choose, and gives guests a link to change or cancel, which turns forgetful no shows into cancellations you can see.

No. Walk-ins were never booked, so they cannot no show. They do matter for the seated rate and for covers, which is why it helps to read covers with and without walk-ins.

Fifteen to twenty minutes is common, but choose what suits your service and stick to it. Changing the hold window changes the rate.

Not necessarily. If it rises while no shows fall, guests are telling you about changed plans instead of staying away, and those tables can often be resold.

Two or three months by service is enough to see a pattern. A single month can be distorted by holidays or events.

Sometimes, sometimes not. Measure it in your own venue by party size before deciding on a policy for large groups.

Use the same definitions everywhere: the same hold window, the same cancellation cut-off and covers rather than tables. Then compare like with like, weekend dinners with weekend dinners. A city-centre venue and a neighbourhood venue may have very different patterns for good reasons, so look at each venue's trend before comparing their levels.

Usually a reminder about a day before the booking with a one-click way to change or cancel. It costs almost nothing, it suits every service, and it tends to convert a share of silent no shows into cancellations you can resell.

Track them separately. A private event that does not happen is usually handled through its own contract and deposit, and it behaves nothing like an ordinary booking. Mixing it into the main rate distorts the figure for that month.

A rate near zero is good news, but check how it is being achieved. If it comes from strict deposits on every booking, you may be turning away guests who would have come. Watch total bookings and the seated rate alongside it.

Start recording them now, even by hand. Without separate outcomes, the no show rate cannot be measured accurately.

Your no show rate is the share of booked, uncancelled covers that did not arrive. Count covers rather than tables, keep cancellations and last minute cancellations separate, mark a no show only after the hold window, and split the rate by service, party size, source and lead time.

Read it alongside the late cancellation and seated rates, watch the trend rather than single weekends, and put a value on it with your own average spend. Measured this way, the number tells you exactly which services need stronger measures, and which need nothing more than a good reminder.