How long a table lasts and how often you seat

A booking system asks you two numbers before it will do anything useful: how long a table lasts, and how many you are willing to seat at once. Almost every restaurant answers the first from memory and the second not at all — and those two guesses are why the same room can feel empty at seven, impossible at eight-thirty and empty again at ten, with the kitchen taking the blame for all of it.

⏱️ Measure your own turn times

Time twenty tables on a normal busy night: seated at, gone at, party size. It is the same exercise as the twenty tickets at the pass in the kitchen course, done from the door instead. The example restaurant came out like this.

The right-hand column is the measured time plus about fifteen minutes, and that buffer is the whole difference between a book that works and one that collapses at nine. It absorbs the table that lingers over coffee, the one that arrives ten minutes late, and the two minutes it takes to clear and reset — none of which are exceptions, all of which happen every night.

Three things to know about these numbers. They are yours: a bistro and a tasting menu do not share them, and neither do your own Tuesday and your own Saturday. They move when the room changes, which is why the previous course made a point of chairs being a decision — comfortable seating adds fifteen minutes to a two-top and that is either what you wanted or a surprise. And they are seasonal at the edges: a terrace in summer holds people longer than a dining room in February.

🧮 How many seatings a night actually is

Now turn the room into arithmetic. A hundred and twenty covers on a Friday, with an average party of just under three, is about forty-one seatings across a four-hour service. That is roughly ten tables an hour, or two and a half every fifteen minutes.

Write that number down, because it is the one that decides whether the night works. It is not a target — it is a ceiling that your kitchen, your floor and your dish pit have to be able to absorb, and it is remarkably easy to exceed for a single quarter of an hour while sitting comfortably under it for the night as a whole.

Which is exactly what a booking platform will do to you if you let it. Guests choose round times: eight o'clock, then eight-thirty, and nothing in between. Left alone, an online book will happily seat nine tables at eight and one at eight-forty-five, and the resulting service is the one where everybody says the kitchen was slow. The kitchen was not slow. It was handed four hours of work in ninety minutes.

🌊 Pacing: the kitchen sets the limit

The rule is to cap how many tables can be seated in each fifteen-minute slot, and the cap comes from the kitchen's own measurement rather than from a feeling. The kitchen course found the constraint by timing the pass and the stations; a restaurant that has done that knows what it can absorb, and a restaurant that has not should start with a conservative three tables per fifteen minutes and adjust.

• Set the cap in the system, not in somebody's head. Every serious booking tool has this setting and almost nobody turns it on.
• Cap covers as well as tables where you can: three twos and three sixes are very different fifteen minutes for the same "three tables".
• Protect the first thirty minutes of service. A kitchen that gets slammed at opening never recovers; a slightly softer start pays for itself all night.
• Offer the times you want filled. Seven and nine-thirty are far more attractive when they are the ones the site shows first, and a small incentive on the early slot moves more people than most promotions do.
• Watch the shoulders. The bookings that make a night work are the ones at the edges, not the ones at eight-thirty, which you were going to fill anyway.
• Re-check after any menu change. A new dish that takes twelve minutes changes what the kitchen can absorb, and nobody ever goes back to the booking settings when the menu changes.

🗺️ The book and the floor plan are the same document

The last piece is joining this to the room from the previous course. A booking is not just a time, it is a table — and the value of having squared off your tables so they combine is entirely lost if the book hands out slots without knowing which table each one lands on.

Two practices make it work. Assign bookings to tables, not just to times, at least for the busy nights, so that a party of five is not accepted when the only combination that seats five is already promised to a four at the same hour. And keep the flexible tables free as long as you can: the two-tops that push together are the room's ability to say yes, and giving them away first is how a restaurant ends up unable to seat a six with eight seats empty.

The other half of the same idea is deciding, in advance, which tables are bookable at all. Reserving the worst table in the room — the one topic 4 of the previous course told you to fix, repurpose or remove — is a way of guaranteeing that somebody's planned evening starts badly. If it is still in service, it is a walk-in table, not a booked one.