In this article
Every owner checks tonight's reservation list every morning. Almost nobody checks when those bookings were actually made — same day, last week, or three months ago. That one number, the booking lead time, is why some reservations almost never fall through and others almost always do.
A booking system shows you two things very well: who's coming, and how many. It rarely shows you a third thing that matters just as much: when that booking was placed relative to the day itself. A table for two booked yesterday behaves statistically completely differently from a table for two booked three months ago — and yet most calendars treat them identically.
That is not an academic distinction. It is a foundational revenue-management concept every hotel has used for decades to decide when a room gets sold, how strict a cancellation policy needs to be, and how much capacity to hold back for latecomers: the booking curve, or pickup analysis. A restaurant has the same curve — almost nobody in hospitality has ever measured their own.
This guide walks through the seven numbers that together make up that curve: how many of your bookings come in same-day, how many plan months ahead, and — the number with the most money riding on it — how the odds a guest simply doesn't show up rise the further ahead the booking was made. Further down, plug in your own figures: your own split across five windows, your covers per night and your average spend. You'll get the expected number of no-shows, what that costs, when a deposit starts making sense, and how many seats to hold back for walk-ins.
Everything runs in your own browser: nothing is sent and nothing is stored. The figures in this guide are reference values based on the booking-curve pattern that revenue-management research in hospitality and hotels has described for years — your own booking system holds the real answer for your venue, and that answer is probably already sitting in data nobody has ever pulled.
Why nobody has ever measured their own booking lead time
A booking system is built to answer one question: who's coming tonight. That's the question the floor manager asks at 5pm, so that's the question every dashboard puts front and centre. "When was this booking placed" rarely sits on that same screen, and even more rarely as a figure you can total across hundreds of bookings — it's in the database, but nobody has ever asked for a report that shows it.
The second reason is that the two signals that would reveal lead time — no-shows and booking date — are almost never looked at together. No-shows get counted per night ("we had three tonight"), not broken down by when those three booked their table. Without that link, every no-show looks the same, even though the pattern behind it is entirely predictable.
And then there's the third reason: most hospitality businesses have no separate policy per window. There's one cancellation policy, one deposit rule — or none — applied to every booking regardless of when it was made, when it's precisely that number, the lead time, that best predicts which booking is actually carrying risk.
The ultimate guide Managing Reservations: The Complete Guide From no-shows to wait times to walk-ins: everything your booking system can do for you, in one guide. Open the guideThe 7 numbers, and what each one tells you
They run in the order that builds your booking curve: first how much books close and how much books far out, then the risk behind it, and finally what you do with it — a deposit cutoff, a walk-in cushion, and a calendar that's neither too short nor too long.
1. More than half your bookings come in same-day or the week before
Line up every booking from the past few months against the date it was placed, and at an average restaurant one thing jumps out immediately: how little advance planning most of it actually involves. Roughly 55% of all reservations get made the same day or within the seven days before — that's not the exception, it's the majority.
That figure surprises most owners, because it runs against how a calendar feels: the bookings that come in two months ahead stand out precisely because they're so far in advance, while the thirty bookings that came in yesterday and today just feel like "business as usual." That's exactly why this number never gets counted separately — it doesn't feel remarkable, even though it's the majority of your revenue.
The chart below sets the five windows side by side so you can see at a glance where your own booking curve's centre of gravity sits — and, just as important, how that share per window relates to the risk that the table stays empty.
Five windows, from same-day to three months and beyond — the share of your bookings alongside the illustrative no-show risk for each window.
The pattern repeats in almost every dataset: the share of bookings falls as the lead time grows, while the no-show risk climbs. That crossover — lots of bookings with low risk at the near end, few bookings with high risk at the far end — is the whole reason one deposit policy for every booking never works well.
2. Barely 5% books three months or more ahead — and they're often your biggest tables
At the other end of the curve sits a small group planning weeks to months out: often groups, birthdays, anniversaries and business dinners where several people's calendars need to line up. On illustrative figures that's around 5% of all bookings — small in count, but rarely small in size: these reservations tend to be bigger than a booking made yesterday for two.
That's exactly why this segment can't just be ignored, even though its share is low. A party of ten that books three months ahead and then doesn't show costs you not one table but a whole night's worth of capacity you could have sold to someone else — and that risk is real, even though the odds of it happening are small.
The question this figure raises comes back at point seven below: how far do you actually open your calendar, and what do you lose if you close that window too early.
3. The further ahead a booking is made, the higher the no-show risk
This is the number the other six revolve around. Revenue-management research in hotels and hospitality — the work of Sheryl Kimes among others, published extensively in Cornell Hospitality Quarterly, and the "pickup" analysis every hotel PMS has run for decades — describes a consistent pattern: the further ahead a booking is made, the more time there is for the guest's circumstances to change, and the higher the share that ultimately doesn't show up.
Translated into illustrative figures for an average independent restaurant: a same-day booking usually carries the lowest no-show risk, somewhere around 3%, because the guest quite literally decided to come today. That risk climbs as the window grows — a booking made a week ahead already sits higher, a booking made a month ahead higher still, and a booking made three months ahead carries the highest risk of the five, at over 20%. These are reference values, not iron law — your own curve lives in your own booking data, and you'll find it in the calculator further down.
Why does the risk climb like that? The longer the gap between booking and arriving, the more likely something else comes up — another invitation, illness, a change of plans — and the weaker the sense of commitment at the moment of booking itself. It's exactly the same logic behind why hotels tighten their cancellation terms the further ahead a room is booked: the risk isn't a guess, it's a measurable pattern.
4. Bigger groups book further ahead than a table for two
A table for two can be decided on the spot — call, book, show up. A table for eight needs eight calendars lined up, and that takes time. The result is a second, smaller curve layered over the first: the bigger the party, the more likely the booking was made weeks to months ahead rather than same-day.
That relationship is intuitive — anyone who has ever tried to plan a ten-person birthday dinner knows the hassle of getting everyone free on the same night — but it has a direct practical consequence: the finding in point 3 above (higher risk at longer lead times) and the finding in this point (bigger groups book further out) reinforce each other. Your biggest tables are exactly the tables carrying the longest risk window.
That's not a reason to turn groups away — it's exactly why a separate policy for group bookings (a deposit, a confirmation call the day before) earns more than applying the same policy to every booking, small or large, near or far.
5. The deposit cutoff sits where the risk doubles
If risk climbs with lead time, when does a deposit actually earn its keep? Not on every booking — that scares off spontaneous, last-minute guests, exactly the segment carrying the lowest risk. It starts making sense from the point where the no-show risk clearly jumps above the short-term baseline.
A usable, inspectable rule: take the no-show percentage of your shortest window (same day) as the baseline, and find the first window whose percentage climbs above double that baseline. On the illustrative figures in this guide, that baseline sits at 3%, so the threshold sits at 6% — and the first window that clears it is bookings made from eight days ahead. In practice: bookings up to a week out stay deposit-free, bookings from just over a week out get one.
That's not a fixed cutoff for every restaurant — it's a method. Plug in your own no-show figures per window (pulled from your own booking system, not from this guide) and the same rule gives you your own cutoff, instead of a rule of thumb that never fits anyone exactly.
A timeline from today to three months and beyond, with the point where the no-show risk doubles relative to same-day bookings.
On the illustrative figures in this guide, that cutoff sits at 8 days: bookings up to a week out stay deposit-free, bookings from just over a week out get one. Plug your own no-show figures into the calculator below for your own cutoff.
6. How many seats you hold back for walk-ins depends on your own short-term share
Walk-ins — guests with no reservation — aren't a separate phenomenon sitting outside the booking curve; they're its far end: people deciding at the exact moment they step through the door, without even a call or an app. A venue where a large share of bookings already arrives same-day typically sees more walk-in traffic too — it's the same spontaneous demand, just without a reservation.
That gives a concrete, inspectable rule of thumb for how much capacity to hold back: roughly the same share of your seats as your same-day booking share. On the illustrative split in this guide (22% of bookings arrive same-day), that means a venue with 45 seats holds back a cushion of roughly 10 seats it doesn't fully commit through the calendar.
That's not seats sitting empty — it's capacity you're deliberately not pre-assigning, so it's there for the demand that shows up regardless, booked or not. If you'd rather predict tonight's demand itself instead of only holding a cushion for it, that's exactly what the covers forecast is for: it predicts how many covers to expect tonight from your own history, the weather and recurring events.
7. A calendar that opens too short loses groups; one that opens too long is betting on demand that doesn't exist yet
This last number isn't a cost but a trade-off, and it decides how far you actually open your booking calendar. Close it too early — say, only four weeks out — and you lose exactly the segment from point 2: the groups, birthdays and anniversaries wanting to lock in months ahead, who simply book elsewhere while your calendar still shows nothing open.
Open it too long — say, a full year ahead — and you're making staffing, menu and pricing decisions for a peak night based on bookings that can still change for months, while the real demand only becomes visible in the final weeks (exactly the biggest chunk of your curve from point 1). A calendar that stays open too long is betting on demand that doesn't exist yet.
There's no universally correct horizon — it depends on how much of your revenue comes from groups and how predictable your peak nights already are without far-out bookings. What works for every restaurant is asking the question again every quarter, instead of setting the horizon once and never revisiting it.
Put your own booking curve into the calculator
Fill in how your own bookings split across the five windows (in %, pre-filled with the illustrative figures from this guide), plus your covers per night, your seats and your average spend.
You'll get the expected number of no-shows per night, what that costs you per month and per year, from which lead time a deposit starts making sense, and how many seats to hold back for walk-ins.
Booking Lead-Time Scan
Your split across five windows, your covers and your spend — the expected loss to no-shows in euro per night, month and year.
Split of your expected no-shows
—
The ratios in this scan are reference values based on the booking-curve pattern from revenue-management research, not a measurement of your specific venue. Everything runs in your browser; nothing is sent or stored.
Two things worth taking away. The figure at the top — your expected no-shows per night — usually runs higher than most owners would guess off the top of their head, precisely because nobody has ever lined up the five windows side by side. And the deposit cutoff isn't a penalty on the guest: it's a threshold that only sits where the risk actually justifies it, not on every booking.
What you CAN act on is where you set that cutoff and how much cushion you hold back — both follow directly from your own curve, not from a rule of thumb that could just as easily apply to the venue next door. Plug in your own figures once you've pulled them from your system, and the calculator above gives you your own answer.
What to do this week, this month and this quarter
Tackling seven numbers at once works for nobody. This order does, because each step makes the next one measurable.
This week — pull your own curve from your booking system
- Export the last three months of bookings with two dates: when the booking was made and which date it was for. Most systems have that report, even though almost nobody ever asks for it.
- Split the bookings across the five windows in this guide and plug your own percentages into the calculator above.
- Mark against each booking whether the guest showed up or not — that gives you your own no-show percentage per window, instead of the illustrative figures.
- Compare your own curve with the illustrative split above: where's the biggest difference, and why?
This month — put the deposit cutoff and walk-in cushion on paper
- Work out your own deposit cutoff with the rule from point 5: double your same-day no-show percentage, applied to your own windows.
- Agree with your team from which lead time a deposit or a confirmation call becomes standard — and state that clearly at the point of booking.
- Lock the walk-in cushion from point 6 into your booking system, so the calendar doesn't automatically fill those seats.
- Review your current cancellation policy: does it apply equally to every booking, or does it already account for lead time?
This quarter — reset your calendar's horizon
- Decide how many weeks or months ahead your calendar stays open, and check that against the share of group bookings you already get further out.
- Missing groups or anniversaries: consider extending the horizon for that category specifically, without opening your whole calendar.
- Repeat the scan every quarter — a terrace restaurant's curve looks different in summer than in winter.
- Put your booking lead time next to your deposit and cancellation policy: one number decides when the other one makes sense.
Lead time isn't a footnote — it's the number that already knew your no-shows before they happened
Most owners who measure their own curve find the same pattern: no surprise, no randomness, just a risk that was sitting in the data the whole time, waiting for someone to pull the report.
That's good news, because the biggest lever costs nothing: a deposit that only applies where the risk actually justifies it, a walk-in cushion that matches your own short-term share, and a calendar horizon you revisit every quarter instead of setting once and forgetting.
Put that next to your no-show policy and your approach to waiting — the three belong together, because all three are about the same moment: between booking and showing up.