NO-SHOW RECOVERY
Two ways to lose a table, and only one of them is the no-show.
Restaurant no-show losses are reduced by two separate things: deterring the bookings that carry real risk, and refilling the table quickly when a booking is lost anyway. Fork Time scores each reservation as it is made and requests a card before confirming only when demand is high and the booking is higher risk, so regulars book with nothing asked and the friction lands where the risk is. When a cancellation or no-show happens, it checks the waitlist and recent enquiries, ranks the guests most likely to want that table at that time, and prepares a text offer your team sends. The first guest to accept is booked automatically. Recovery is the half most restaurants have no mechanism for, and the number that decides the loss is time-to-refill rather than the no-show rate itself.
THE RATE IS NOT THE PROBLEM
Published no-show figures vary enormously. Platform data from one large US point of sale reports around 2% of booked reservations, while reservation platforms report figures as high as one in five in peak-demand and fine-dining settings. The honest answer is that your rate depends on your room, your market and your night.
Which is why the rate is the wrong thing to manage. A venue with a 20% no-show rate and a ten-minute time-to-refill loses less money than a venue at 5% with no way to refill anything. Prevention has a ceiling. Recovery does not, and almost nobody measures it.
Time-to-refill is the median number of minutes between a table becoming available and that table being rebooked. It is the number that determines what a lost booking actually costs, and it is the number to start tracking this week whether or not you buy anything.
WHY BLANKET DEPOSITS BACKFIRE
The standard advice is to take deposits. The standard result is fewer bookings.
A blanket policy has inverted incidence. Call it the trust tax: the cumulative friction a single policy imposes on a restaurant's most loyal guests, who meet it most often precisely because they book most often. A regular booking twenty times a year is asked for a card twenty times. A first-timer is asked once. The friction lands hardest on the guests least likely to no-show.
The cost is measurable on the other side too. 44% of diners find a restaurant less appealing and stop trying when booking becomes difficult (Toast diner survey, 850 US adults). A policy written to deter no-shows also deters bookings you wanted.
There is a second-order effect worth knowing. Pricing an obligation can weaken it. In a well-known field study, ten daycare centres introduced a fine for late pickup and lateness increased rather than fell, and stayed high after the fine was removed; the researchers concluded the fine had converted a social obligation into a purchased option (Gneezy and Rustichini, 2000). A later survey-based study did not reproduce the effect and the mechanism is still debated. A reservation is an unpriced promise, and it is worth being careful about what attaching a price to it does.
WHAT THE EVIDENCE SAYS
Three interventions have real evidence behind them, in order of how well documented each one is.
Ask for an active commitment. The best-known case is Gordon's restaurant in Chicago, documented by psychologist Robert Cialdini in Influence. The receptionist changed "please call if you have to change or cancel your reservation" to "will you please call if you have to change or cancel your reservation?" and paused for the yes. No-shows fell from about 30% to 10%. One restaurant rather than a controlled trial, but the mechanism, that a spoken yes creates a commitment people act to keep, is among the most replicated findings in behavioural science. Fork Time's confirmations are worded to ask for a reply rather than announce a booking, and bookings that never confirm surface for follow-up before service.
Hold a card where the risk is. A card on file changes what a no-show costs the guest, and that is the deterrent. The blanket version of this policy carries the costs described above, which is why the card is requested per booking, on the reservations that carry the risk, and charged only when your team marks the no-show.
Remind with a question, not an announcement. The strongest reminder evidence comes from healthcare, where SMS reminders reduce missed appointments across dozens of published trials, and reminders that request a confirming action outperform reminders that only inform. The same shape applies to a table. A reminder that asks the guest to confirm or release does two jobs at once: it renews the commitment, and a released table comes back while there is still time to resell it, which is exactly what feeds recovery.
WHAT WORKS INSTEAD
Score the booking, not the policy. Fork Time reviews each reservation as it arrives, using guest history, demand on that service, party size and lead time. Low risk confirms with nothing asked. Higher risk is asked for a card before confirming, with the policy stated plainly. The card is charged only if your team marks the booking as a no-show. You set the thresholds and any team member with permission can override a recommendation.
Make cancelling easy. A guest who can cancel in one tap hands the table back while there is still time to sell it. A cancelled table you refill is worth more than a fee you collect.
Have a recovery mechanism at all. This is the gap. When a table opens, Fork Time checks the waitlist and recent enquiries, including guests who asked for that exact time and could not be accommodated, ranks them by fit and recency, and prepares the offer. Your team sends it. The first acceptance takes the table and the offer closes. Every recovered booking is recorded, so time-to-refill becomes visible from the first week.
WHAT THIS DOES NOT DO
It does not eliminate no-shows. A guest determined not to arrive still will not arrive, and a card hold is a deterrent rather than a guarantee.
It also cannot refill a table nobody wants. Recovery works by matching an open table against guests who already asked for that time, so a cancellation at eight on a quiet Tuesday with no waitlist is usually lost. And the offer needs a person to send it, which is deliberate: a system that texts your guest list unsupervised is a worse problem than an empty table.
BEST FOR
Restaurants where no-shows concentrate in peak services, new guests and large parties. Venues that dropped deposits because of the booking friction. Rooms that carry a waitlist and currently have no way to use it when a table opens.
NOT DESIGNED FOR
Ticketed events and full-prepayment set menus, where every seat is paid for by design and there is nothing to score. Venues with demand well below capacity, where the cost of asking anyone for a card is higher than the cost of a no-show.
EVERY CALL ANSWERED
See it working on your own service.
The demonstration uses examples based on your restaurant and service style.