Walk-ins vs reservations: finding the right mix for your restaurant

Two guests reach your door at the same moment on a Friday night. One booked three days ago and expects her table to be ready. The other just walked past, smelled dinner, and wants in. Which one does your restaurant exist for?

Most owners never decide. They take reservations because guests ask for them, keep seating walk-ins because turning money away hurts, and end up with the worst of both: a floor that shows "full" on paper while real chairs sit empty, spontaneous guests bouncing off a "fully booked" that was not quite true, and a first seating that arrives all at once because nobody paced it.

Walk-ins versus reservations is not a war with a winner. It is a dial, and the right setting depends on your concept, your street, and even the night of the week. This guide walks through the three models, the honest math on no-shows and empty tables, how many tables to hold back if you go hybrid, why bar seating is the natural flex zone, and why overbooking, which works fine for airlines, is a trap for restaurants.

The three models, and what each one really optimizes

Full reservations: predictability, at the price of no-show risk

When every table is bookable, you know your night in advance: how many covers, at what times, in what party sizes. The kitchen preps to real numbers, staffing matches demand, and an online booking system with a free plan collects bookings around the clock, including from the guest deciding at 11 pm for next Saturday.

The price of that predictability is exposure to the guest who never turns up. A fully reserved floor has zero slack: every no-show is a table you refused someone else for, and it stays empty. The scale of the problem is documented on both sides of the Atlantic: in the US, 28% of adults admitted skipping a reservation without canceling in the previous year, according to a YouGov survey commissioned by OpenTable, and in the UK, research by Zonal and CGA put the cost of no-shows to hospitality at £17.6 billion a year in lost sales. A full-reservations model only works when you actively manage that risk: confirmations, reminders, and card holds where the stakes justify them. Our guide on reducing restaurant no-shows covers the whole toolkit.

Walk-in only: energy and simplicity, at the price of visibility

No phone to answer mid-rush, no no-shows by definition, a line out the door that markets the room better than any ad, and every seat sellable to whoever is physically there. Walk-in only is not a lesser model; for some concepts it is the stronger one. Fast-casual spots, counter-service concepts, tapas bars and high-traffic city locations often run better without a book.

The costs are just as real. You have no idea whether tonight brings 40 covers or 90, so prep and staffing are guesswork. Peak hours produce a queue you must manage well or lose (an honestly quoted wait beats a vague "shouldn't be long", every time). Quiet nights have no safety net: nobody committed in advance, so nobody comes. And guests planning an occasion, a birthday, a client dinner, a party of eight, simply cannot choose you.

Hybrid: most restaurants' real answer

The hybrid model takes reservations but deliberately holds a share of the floor back for walk-ins. You get advance visibility on most of the night, plus slack that absorbs both spontaneous demand and the damage from no-shows: a reserved table that stays empty can be given to the walk-in queue, and a walk-in surge can overflow onto tables freed by early finishers.

Hybrid is also the model that demands actual management. Somebody has to decide how many tables stay off the books, on which nights, and until what time, and revisit those numbers with data instead of gut feeling.

Walk-ins vs reservations at a glance

How many tables should you hold back for walk-ins?

There is no universal percentage, and any article that gives you one is guessing. What works is a starting point plus a feedback loop:

1. Start from your street, not from a benchmark. Heavy foot traffic (shopping street, tourist center, nightlife district) justifies holding back a meaningful block on peak nights, because spontaneous demand will reliably fill it. A destination restaurant on a quiet road should reserve nearly everything, because almost every cover is won online or on the phone, not off the sidewalk.
2. Vary it by night. Saturday 8 pm and Tuesday 7 pm are different restaurants. A common-sense pattern: hold tables back on the nights you regularly turn walk-ins away, and open the whole floor to reservations on the nights you need commitments in the book to guarantee any revenue at all.
3. Release the held tables at a set time. Tables kept for walk-ins until close are a bet that can lose. Pick a release point ("if the walk-in tables are still empty at 8:30 pm, they go to the waitlist and back on sale online") so slack never hardens into waste.
4. Count two numbers every week: refused walk-ins and empty reserved tables. Lots of refusals and few empty tables: hold more back. The reverse: reserve more of the floor. Two or three weeks of honest tallies beat a year of impressions, and a reservation history gives you half of this data automatically.

Bar seating: the flex zone that makes hybrid work

If you have a bar or counter, you already own the easiest version of the hybrid model: keep the dining room bookable and the bar first come, first served.

The bar absorbs walk-ins without costing you a single reservable table, gives waiting guests somewhere to spend money instead of standing by the door, and keeps the room looking alive from the street. Full-menu bar seating in particular turns "we're fully booked" into "the dining room is booked, but you can eat at the bar right now", which is a completely different sentence to hear.

Pacing: the part both models get wrong

Whether guests book or walk in, the enemy is the same: everyone arriving at once. A kitchen hit by thirty covers in fifteen minutes serves all of them worse than sixty covers spread across ninety minutes.

Reservations give you the tool to fix this, if you use it: instead of accepting any time to the minute, offer arrival slots (7:00, 7:30, 9:00, 9:30 pm) and cap how many parties each slot takes. You are no longer just recording demand, you are shaping it around what your kitchen can actually deliver. As a reference for sizing slots, TheFork Manager observes that dinner guests typically stay two to two and a half hours, so a 7 pm slot naturally feeds a second seating around 9:15 pm. Measure your own durations before copying anyone's numbers.

On the walk-in side, pacing means running the door honestly: quote waits on the high side, take names and phone numbers, and let people wait somewhere pleasant rather than in a scrum by the host stand. A well-run restaurant waitlist is the walk-in equivalent of arrival slots: it converts a chaotic queue into an ordered flow, and it doubles as the refill mechanism when a reserved table opens up.

Overbooking: why airlines can and restaurants mostly should not

The comparison comes up constantly: airlines sell more seats than the plane holds, count on no-shows, and it works. Why not restaurants?

Because the airline version rests on machinery restaurants do not have. Airlines predict no-show rates from millions of past flights on the same routes, and when the bet fails, a whole US federal regulation, 14 CFR Part 250 "Oversales", organizes the fallout: carriers ask for volunteers, apply published boarding-priority rules, and hand involuntarily bumped passengers a written explanation of the denied boarding compensation they are owed. The overbooked passenger is inconvenienced, compensated, and put on another flight.

Now run the restaurant version of a failed bet. You took 110 covers for 100 seats on the theory that 10% would not show. Tonight everyone shows. There is no later flight to your dining room, no compensation schedule, no volunteer system: just parties standing at your door holding a confirmation you cannot honor, on what was probably a birthday or an anniversary, ready to describe the evening in a one-star review. You do not lose a booking, you lose the guest, publicly.

The honest exception: a very high-volume operation with years of its own data might shade capacity slightly on patterns it truly knows. For everyone else, the same fill rate is available without the downside, and it is simply the hybrid toolkit: reminders that turn silent no-shows into early cancellations, card holds on high-stakes nights, a waitlist that refills freed tables in minutes, and software that puts a canceled slot back on sale online the moment it opens. Overbooking transfers your no-show risk onto your best-behaved guests; the toolkit removes the risk instead.

Match the model to the concept, then to the night

Pulling it together, the dial position mostly follows from what you are:

• Fine dining, tasting menus, destination restaurants: full reservations, protected by reminders and card holds. Your guests plan; spontaneity is marginal; a no-show costs an entire seating at your highest check (or bill, in the UK).
• Neighborhood full-service restaurants: hybrid. Reserve most of the floor, hold a block or the bar for walk-ins on proven busy nights, open everything to booking on slow ones.
• High-footfall casual spots, counters, fast turnover: walk-in first, with a managed queue. Consider taking reservations only where they add real value: large groups, off-peak seatings, quiet weeknights you want to guarantee.
• Everyone, on slow nights: bookings are your friend. A Tuesday with twelve covers in the book at noon is a Tuesday you can staff, prep and market; a walk-in-only Tuesday is a coin flip.

And revisit the setting seasonally. A terrace (or patio) in summer, a tourist wave, a new office building across the street: each one moves the right answer.

Where ViteUneTable fits, and where it does not

ViteUneTable is reservation software for restaurants, so it powers the reservation side of whichever mix you choose. The free plan takes unlimited online bookings through your own link, sends automatic confirmation emails and keeps your reservation book, with 0% commission, no per-cover fees and no commitment. The Standard pack, at €29 excl. VAT per month, adds automatic email reminders before service and the Reserve with Google button on your listing. The Standard + Anti No-Show pack, at €49 excl. VAT per month, adds credit card holds for the nights where a no-show would really hurt. Hybrid setups are just how you use the book: whatever you do not open to online booking stays yours for the door.

Let's be honest: if your restaurant is genuinely walk-in only, a counter spot with a line and no book, you may not need reservation software at all, ours included. A queue at the door is managed at the door. The moment you start taking bookings for groups, quiet nights or a reservable dining room alongside a walk-in bar, that is when a reservation system starts paying for itself, and starting on a free plan means finding out costs you nothing.

Frequently asked questions

Should a small restaurant take reservations or stay walk-in only?

Decide from your street and your slow nights. If foot traffic reliably fills you, walk-in only keeps operations simple. If some nights are unpredictable, or guests ask to book for groups and occasions, take reservations at least for those cases: advance bookings are the only tool that guarantees covers on a quiet Tuesday.

What percentage of tables should be held for walk-ins?

There is no universal figure. Start from your location (heavy foot traffic justifies a bigger held-back block), vary it by night of the week, and adjust with two weekly numbers: walk-ins you refused and reserved tables that sat empty. Release held tables at a set time so unused slack goes back on sale.

Do walk-in-only restaurants avoid no-shows completely?

By definition yes: nobody can fail to honor a booking that does not exist. The trade-off is that walk-in-only restaurants absorb all demand uncertainty instead, with unpredictable covers, guesswork staffing and no committed revenue on slow nights. Reservations move the risk; reminders and card holds then shrink it.

Why is overbooking worse for restaurants than for airlines?

Airline overbooking runs on huge historical data and a regulated safety valve: US rules on oversales organize volunteers, boarding priorities and denied boarding compensation, and a bumped passenger gets another flight. A restaurant has no equivalent: when everyone shows up, guests with confirmations are turned away with nothing, and the damage lands in public reviews. A waitlist plus fast online re-release of canceled slots achieves the fill rate without the betrayal.

How do arrival slots help with walk-ins and reservations?

Slots pace the reserved side of the night so the kitchen is not hit by every party at once, and a predictable reserved flow is exactly what lets the door say yes to walk-ins: the host can see when tables free up instead of guessing. Cap how many parties each slot takes and base slot spacing on your measured meal duration.

Can I run different models on different nights?

Yes, and most hybrid restaurants should. Open the whole floor to reservations on quiet nights, when the book is your safety net, and hold tables or the bar back for walk-ins on peak nights, when spontaneous demand is guaranteed. Your reservation history tells you which night is which.