Restaurant No-Shows in 2026: What 641,000 Reservations Reveal About Who Books and Never Arrives

Ask any operator about no-shows and you will hear the same story: it is the big Saturday-night party that books an eight-top and vanishes, and the answer is a credit card hold. The reservation platforms have built entire deposit products around that story.

The reservation data tells a different one. We analyzed 641,000+ advance reservations across roughly 50 restaurant locations in the Bloom Intelligence network and found that the guests most likely to no-show are not big parties at all. They are two-tops, booking online, for a Sunday or Monday. By the end of this post you will have the real 2026 no-show benchmarks by channel, day, and party size, a dollar figure for what empty tables cost per location, and a prevention playbook that does not punish your best guests.

What Is the Average Restaurant No-Show Rate in 2026?

In Bloom Intelligence network data, 3.7% of advance restaurant reservations end in a no-show and 15% are cancelled, so 18.7% of booked tables, nearly 1 in 5, never seat a guest. That is measured from 641,000+ POS-linked reservation records since January 2025, not from surveys.

You have probably seen industry no-show figures as high as 15% or 20%. Those numbers usually come from operator surveys, which blend no-shows with cancellations and rely on memory of the worst nights. Reservation-system records tell a calmer but more expensive story: the pure no-show rate is lower than the folklore, but it is relentless, it hits every day of the week, and unlike a cancellation it gives you zero chance to rebook the table. For context on what your other guest numbers should look like, see our 2026 restaurant benchmarks report.

What Does a No-Show Actually Cost?

The average seated reservation in the Bloom Intelligence network carries a check of roughly $165, so the 23,792 no-shows in our data represent about $3.9 million in vanished bookings, approaching $49,000 per location per year. That figure counts only the empty table, not the guests turned away while it sat reserved.

The margin math makes it worse. As we covered in our restaurant profit margins benchmark, full-service restaurants keep 3 to 5 cents of every revenue dollar. A no-showed $165 table does not cost you $165 of profit, but the labor was scheduled, the prep was done, and the seat inventory expired the moment service ended. Revenue that never arrives against costs already spent is the most expensive kind of loss a restaurant has, and it compounds every week of the year.

It also lands hardest on the nights you can least absorb it. A no-show on a slammed Saturday might be backfilled from the walk-in line. The same no-show on a quiet Monday is simply gone, and as the next section shows, Monday is exactly when no-shows peak.

Web Reservations No-Show More Than Phone Reservations

Online reservations no-show at 3.9% versus 3.0% for phone reservations, and cancel at 16.3% versus 9.3%. The channel with the least booking friction is also the channel with the least commitment.

Nobody dials a restaurant, waits on hold, and gives their name to a host casually. But tapping a 7:30 slot in an app takes eleven seconds and costs nothing emotionally, so the booking is abandoned just as lightly as it was made. Two decades of the industry pushing every reservation online has been great for volume and quietly corrosive for commitment. The point is not to bring back the phone. It is that a web reservation arrives with less attachment built in, so the attachment has to be built after the booking, through confirmation and reminder touchpoints that give the guest a moment of human connection the app never did.

There is one more wrinkle: the phone reservations you do get are disproportionately valuable. Phone bookings in our data average 7.2 seats per party versus 3.3 for web. Large groups still pick up the phone, which means a single phone no-show walks out with roughly $375 of booked revenue, more than double the network-average table. Fewer phone reservations fail, but each failure hurts more.

The Party-Size Surprise: Two-Tops Are Your No-Show Problem

Parties of one or two no-show at 4.5%, while parties of seven or more no-show at just 1.8%, and over half of all no-shows in our data are two-tops or singles. The smallest parties are two and a half times more likely to vanish than the largest ones.

It makes intuitive sense once you see it. An eight-top is a birthday, an anniversary, a client dinner. It has a group chat, a guest of honor, and social pressure to show up. A two-top made on a Tuesday for the following Sunday is a loosely held plan between two people, and loosely held plans lose to a rainy evening and a couch. The day-of-week data confirms it: Sunday and Monday reservations no-show at 4.2%, the highest of the week, while Wednesday runs 3.1%. The reservations guests forget are the ones they made for the days that felt optional.

The volume math is what should change your prevention strategy. Big parties no-show rarely but expensively. Small parties no-show constantly, and because two-tops dominate reservation volume, they account for 52% of all the no-shows we measured. Any prevention system that only guards large parties is guarding the wrong door.

Why Deposits and Blacklists Backfire

Deposits, card holds, and no-show blacklists target large parties, but Bloom Intelligence data shows parties of 7+ are the least likely group to no-show, so the industry’s default defense adds friction for the guests most likely to arrive while ignoring the two-tops driving half the problem.

Deposits do work, in the narrow sense that a guest with $25 per head on the line shows up or cancels early. But look at what they cost. They add a payment wall to the booking flow, which suppresses reservations from exactly the high-intent large parties that were already 97% reliable. They create hostile moments at the host stand when plans legitimately change. And no restaurant is going to demand a card hold from every two-top on a Monday, which is where the actual no-shows live. Blacklists are worse: they punish a guest after the revenue is already lost, and in our first-time guest research, where 8 in 10 first-time guests never return anyway, burning a bridge over one missed booking is a strange trade.

Deposits guard the eight-top that was always coming. Nobody is guarding the two-top that was always a coin flip.

The Real Fix: Know Who Is Booking Before They Arrive

No-show prevention works when the reservation is connected to a guest profile, because a first-time booker, a slipping regular, and a loyal weekly guest need three different messages before the same 7:30 slot. A reservation system alone sees a name and a party size. A customer data platform sees the guest.

When reservations flow into unified guest profiles alongside WiFi visit history, POS spend, and email engagement, the no-show problem changes shape. That Sunday two-top stops being an anonymous coin flip. It is either a first-time guest whose visit deserves a warm confirmation and a reason to keep the plan, or a regular whose booking pattern just changed, which is worth knowing for entirely different reasons. Prevention becomes a personalized touchpoint instead of a blanket policy, and the same connected data then powers everything downstream, from the second-visit campaigns in our sales playbook to at-risk recovery, where Bloom recovers an average of $53,000+ per location per year.

Automating Prevention With Reservation Triggers

Bloom’s Summer 2026 release added workflow triggers for all five reservation actions, so restaurants can automatically fire the right message at booking, confirmation, seating, completion, and cancellation without anyone touching a send button.

In practice, a no-show prevention workflow looks like this. A guest books online and instantly gets a branded confirmation. The day before, a reminder goes out by SMS, the channel with 98% open rates, with one-tap options to confirm, modify, or cancel. Because the trigger knows the guest profile, a first-time booker gets a different message than a regular. If the guest cancels, a cancellation trigger immediately invites them to rebook, converting a lost table into a future visit. And when the reservation completes, the post-visit sequence starts working on the second visit, which our data shows is the single most decisive moment in guest lifetime value. Behavior-based triggers with n’th-occurrence thresholds even let you treat the third no-show differently from the first, quietly, without a blacklist.

How to Reduce Restaurant No-Shows: The Checklist

Step 1: Measure your real baseline. Pull your no-show and cancellation rates by channel, day, and party size for the last 90 days. You cannot manage a 3.7% problem you have been guessing is 15%.

Step 2: Connect reservations to guest profiles. Route reservation data into your CDP so every booking resolves to a known guest with visit history, spend, and contact permissions.

Step 3: Automate the confirmation and reminder sequence. Instant confirmation at booking, an SMS reminder the day before with one-tap modify and cancel, and messaging that differs for new versus returning guests.

Step 4: Make cancelling effortless. Every silent no-show you convert into an early cancellation is a table you can rebook. Friction in cancelling does not save the booking, it just hides the loss until service.

Step 5: Watch Sundays, Mondays, and two-tops. Aim your reminders where the data says no-shows live, not where the folklore says they do.

Step 6: Reserve deposits for true high-stakes bookings. Holiday buyouts and large-format events justify card holds. The Tuesday four-top does not.

Restaurants do not have a no-show problem. They have a stranger problem. Nobody stands up a place that knows their name.

Methodology

This analysis covers 641,036 advance reservations (web and phone origin) recorded across approximately 50 restaurant locations in the Bloom Intelligence network between January 1, 2025 and August 6, 2026. Walk-in records were excluded because a walk-in cannot no-show. No-show and cancellation figures use the explicit status reported by the integrated reservation systems. Average check figures come from the subset of completed reservations with linked POS transactions (587,047 covers across the same window) and are stated as blended network averages. Earlier reservation records exist in the network but predate no-show status reporting and were excluded. These are network-level figures: the Bloom Intelligence network skews toward full-service and multi-location operators, and individual restaurants, segments, and markets will see different numbers. Nothing here is a guarantee of results for any specific restaurant.