What 3.8M Reservations Say About No-Shows

The number you have heard is not the number anyone measured

If you run a Las Vegas restaurant, someone has told you that one diner in five books a table and never shows up. That figure comes from surveys where people admit to past behavior. The Resos No-Show Index 2026 measured something else entirely: 3,768,761 real reservations at 2,417 restaurants, from 1 August 2025 to 31 July 2026. The recorded no-show rate was 2.33 percent. That is 87,953 no-shows and 283,728 booked covers that never walked through the door.

Both numbers can be true at once. A diner who ghosted one table three years ago is a yes on a survey and a rounding error in a booking system. The operational question is not how many people have ever done it. It is how many tables you lose this month, and the index is the better answer to that question.

Read the method before you use the rate

Recorded means a staff member marked the booking as a no-show in the system. Nobody marks every one, especially on a full Saturday. Resos says so plainly, which means 2.33 percent is a floor rather than a true rate. Among the 1,761 locations that actively record no-shows, the rate rises to 2.49 percent.

Two more adjustments before a Nevada operator borrows this number. The UK makes up 54 percent of the dataset, so the global average is mostly a UK average. The United States slice ran 3.17 percent across 309,604 reservations at 316 restaurants, meaningfully worse. Walk-ins are excluded throughout, so all of this describes your booked business only. If you want a single anchor, use 3.17 percent and treat it as conservative.

The average restaurant in the index recorded about 36 no-shows a year, roughly 117 booked covers.

The digest

Notice the party size column runs the opposite direction from the panic. Small tables ghost more often, big tables ghost more expensively. A party of ten no-shows at 1.15 percent but costs 13.44 covers when it does, which is why the expected loss per booking is three times higher than a deuce.

The bigger hole is the cancellation that arrives too late to matter

The index also counted 566,049 cancellations. Of those, 56 percent, or 317,541, landed within 24 hours of the booking or after it had already started. 40,360 were cancelled after the start time. The full timing distribution runs 7.1 percent after start, 17.4 percent zero to two hours before, 15.0 percent two to six hours, 16.6 percent six to 24 hours, 11.6 percent 24 to 48 hours, 17.4 percent two to six days, and 14.6 percent seven days or more.

A cancellation seven days out is inventory handed back to you. A cancellation 90 minutes before a Friday seating is a no-show wearing a polite hat. Counting both, the restaurants in the index lost about 405,000 tables in a year, roughly 168 per restaurant against 36 recorded no-shows. The late cancellation is four to five times the size of the problem you are already tracking, and almost nobody logs it separately.

A Las Vegas dinner service, worked out

These assumptions are ours, not the index's. Assume a room taking 120 dinner reservations a week, an average party of three, an average check of 55 dollars a cover, and no deposit policy. Dinner runs 2.63 percent in the index.

120 times 0.0263 gives 3.2 no-shows a week. At three covers each that is 9.5 covers, about 521 dollars of seated revenue, or roughly 27,100 dollars a year if those seats never resell. Now apply the index ratio of total lost tables to recorded no-shows, 168 to 36, which is 4.7 to 1. Weekly exposure lands closer to 15 tables, not 3.

Two honest caveats. Some of those seats do resell, especially Strip adjacent on a Friday with a live waitlist, so 27,100 dollars is an upper bound on the no-show piece alone. And 2.63 percent is a recorded rate at restaurants that are mostly not in Nevada. Run this arithmetic on your own cover count and your own average check and you will have a number you can defend in a staff meeting. Run it on the 20 percent survey figure and you will not.

What actually moved the number

Three interventions show up in the data, and all three are observed associations rather than proven causes.

Delivered SMS reminders came out best. Within the same restaurants, reservations that received one had a 16 percent lower recorded no-show rate, a risk ratio of 0.836, across 46,884 exposed reservations at 257 locations. Comparing inside a restaurant rather than across restaurants is the right way to run that test, but the restaurants still chose which bookings got a reminder.

Deposits and no-show protection show the widest gap and the weakest evidence. Reservations that required them sat at 0.80 percent, 1,907 out of 238,539 at 327 restaurants, against 2.47 percent on the 3,461,164 that did not. Resos ran a matched study on that comparison and it came back inconclusive. Restaurants that ask for a card also have a different guest, a different price point, and a different booking culture.

Confirmation prompts came in at a risk ratio of 0.983, near enough to nothing.

The reminder result matches what we found looking at reminder timing and no-shows: delivery and timing carry the effect, not the wording. Pair that with the lead time rows above. Bookings made two to six days out are the worst at 2.67 percent, worse than bookings made 30 or more days ahead at 1.96 percent, which means the riskiest reservation is the one made just far enough in advance to be forgotten and not far enough to feel like a plan.

What to do this week

1. Split your reporting so late cancellations are counted separately from early ones. Use a 24 hour cutoff, because that is where the index draws the line and where 56 percent of cancellations landed.
2. Confirm that your reminder texts are actually being delivered, not merely queued, and send them to dinner bookings made two to six days ahead first. The same mechanics apply to any appointment scheduling system, restaurant or otherwise.
3. Price your own exposure with the arithmetic above, using your covers and your check average, before you consider a deposit policy. The deposit evidence is descriptive, so treat it as a hypothesis to test on one service, not a rollout.

If you want a second set of eyes on those numbers, our free 15-minute audit walks your booking flow and shows what an AI system can and cannot do for a Las Vegas restaurant.