The No-Show You Could Have Predicted

A table booked for eight on a Saturday that does not arrive is one of the most expensive events of the night. The section was held, the staffing was set for it, and both are gone.

The usual response is a deposit policy and a shrug.

But no-shows are not evenly distributed. They cluster, and the clusters are visible in the booking record before the night happens.

What predicts a no-show

Four things, and most venues have all four already recorded without ever looking at them together.

Lead time. A table booked three weeks out no-shows at a far higher rate than one booked the same day. Plans change; the further out, the more change.

Channel. A booking made by phone, by a form, by a message, and through a third party do not behave the same. The order varies by venue, but the spread is usually large — and consistent enough to act on.

Whether it is a first booking. A guest with history is a much better bet than a name with none. This is the strongest single signal in most rooms.

Whether anyone confirmed. A booking that was confirmed by a reply from the guest — not just a message sent to them — behaves differently from one that was not.

None of these require judgment. They are fields you already have.

The measurement

For the last three months, split your bookings into no-show and arrived, then compare the two groups on each of the four.

You do not need statistics. If the no-show group averages eleven days of lead time and the arrived group averages four, you have found something.

Do it once. It takes an afternoon and it usually changes what you do about the confirmation call more than any policy would.

The confirmation that works

Most venues send a confirmation. Fewer require a response.

A message that asks nothing is a notification. A message that requires a reply is a filter. The second one produces a signal: the tables that do not reply are your at-risk list, and they are usually a small enough number to call.

The timing that works in most rooms: far enough ahead that a cancelled table can be resold, close enough that the guest's plans are settled. For a weekend booking that is usually the same week, not the same day.

Same-day confirmation is nearly useless for resale. By then you cannot fill an eight-top.

What a deposit actually does

Two things, and only one is the one people expect.

It does not primarily recover money. The amount is usually small relative to what the table would have spent, and collecting on it creates friction with the guest who does show up next time.

It filters at the point of booking. Someone who will not put a card down was less likely to arrive. The deposit removes the booking before it occupies your section.

That is the real value: fewer held tables, not more recovered revenue.

⚠️ What you may charge, when, and what you must disclose is governed by consumer rules that differ by market and by card scheme. Get the policy wording checked before you apply it — that is a question for your counsel, not for an operations decision made after a bad Saturday.

The overbooking question

Once no-show rates are measured, someone will suggest overbooking.

Be careful. In a restaurant a no-show frees a table that turns; in a club a held section that arrives late still occupies the space. Overbooking a room with fixed sections and minimums produces a worse failure than a no-show — two groups, one section, at 23:00 on a Saturday.

If you overbook at all, overbook the smallest tables, and only where you have somewhere to put the overflow.

The part that is not the guest's fault

Worth naming before blaming the booking.

Some no-shows are yours. A booking taken on a busy night by someone who did not write down the details, a time that was misheard, a table that was given away and the guest arrived to find nothing.

The signal for this is a no-show that turns out to be a guest who did arrive and left. Those are recoverable and they are frequently miscounted as no-shows, which makes your rate look worse and points you at the wrong fix.

Separate "did not arrive" from "arrived and was not seated." They are opposite problems.

Three numbers to hold

No-show rate by lead time band. Same day, this week, further out. Tells you where the risk actually is.

No-show rate by channel. Frequently the largest single difference, and the easiest to act on.

Confirmation reply rate. Your at-risk list, available before the night rather than after it.

Where the record has to sit

The booking is in one place, the arrival in another, the guest's history in a third. That is why the four fields never get compared, and why no-shows stay filed under bad luck.

tasteck keeps bookings, arrivals, and guest history on the same record, so which bookings look like the ones that did not arrive is a list you can work through on Friday afternoon rather than a post-mortem on Sunday.

The booking already told you. It just was not being read.