Do no-show fees actually work? What the reservation data says

When we map a restaurant's reservation week, we expect the no-show problem to be a reminder problem. It rarely is. The booking that vanishes on a River North Friday at 7 PM usually got two texts and an email. The guest read them, meant to come, and then a better plan or a closer table won. A reminder informs. It does not commit. The confirmation text is one channel and the phone is another, and the line that rings out during the rush loses covers the same way a no-show does. The gap between informing and committing is where the money leaks, and it is why the fix is a policy question before it is a software question.

Why an empty table costs more than the cover

No-shows are not a rounding error. OpenTable reports that 28 percent of Americans say they skipped a reservation in the past year, and full-service margins leave nothing to absorb it. The National Restaurant Association's 2024 data puts income before taxes at a median of 1.1 percent of sales for full-service restaurants under 2 million dollars, and 4.3 percent above it. A four-top that books prime time and never arrives is not one lost check. It is the walk-in you turned away, the prep you bought, and the labor you scheduled, against a margin that thin. On a busy Saturday the seat cannot be resold at 8:15.

Test what a deposit policy would do to your book before you turn it on

Do fees and holds reduce restaurant no-shows?

Yes, and the size of the effect tracks how much the guest has committed. According to OpenTable, a credit card hold makes guests up to 16 percent less likely to no-show and 15 percent less likely to cancel late. A deposit does far more: the same data shows deposits cut no-show rates by 57 percent on average and make guests 72 percent less likely to cancel. Prepaid experiences, where the guest pays for the meal up front, cut no-shows 44 percent. Plain reminders and confirmations still help, mostly by catching the honest guest who forgot, but they move the number least because they ask for nothing. The pattern is consistent: friction at booking converts intent into commitment. Holds and deposits clearly work. The harder call is which reservations deserve that friction, because every card field you add also costs you some bookings.

The order we would install it

Booking friction is a scalpel, not a blanket. Applied to every reservation, it costs more covers than it saves. The order we would run it, the way we map an operating week, puts the cheapest tool first and reserves the blunt ones for the seats that actually bleed.

1. Fix the confirmation sequence and the data first. One clean guest record, a text the guest can reply to, a same-day reminder timed to service rather than to midnight.
2. Put a card hold on the slots that bleed. Prime-time Fridays and Saturdays, parties of six or more, holidays. Leave the Tuesday two-top alone.
3. Reserve deposits and prepayment for the highest-risk seats. Tasting menus, private events, New Year's Eve, the tables you cannot refill on the night.
4. Wire the waitlist to refill before service. This is the only tactic that recovers the seat instead of only deterring the loss.

Then measure by slot, not in aggregate. A single blended no-show rate hides the truth that a Saturday 8 PM four-top and a Wednesday walk-in are different risks. This is also where the reservation book sits among the workflows a restaurant group fixes first: it feeds the same guest record, the same floor plan, the same labor plan.

The fee that looks like a win

The no-show fee is the tactic operators reach for first, and it is the one that bites back. It works where it is enforced, but enforcement means charging cards on guests who feel punished, fielding chargebacks and disputes, and wearing the review that follows. There is a new cost too. In the second half of 2025 OpenTable began adding a 2 percent service charge on transactions that include no-show penalties and deposits, rolling out to most US restaurants by early 2026. A restaurant can absorb it or pass it to the guest, and either way the fee you set is not the fee that lands. None of this makes fees wrong. It makes them the last tool to reach for, and the one that most needs a human deciding when to waive it. A guest whose flight got cancelled is not the guest the policy was built for. Charge the seat you cannot refill. Forgive the one you can.