Table turnover restaurant

Table turnover in a restaurant — how to measure and speed rotation without spoiling the guest

Saturday night, the floor is full, a queue at the door, and the day's sales look like an average Thursday. Or the opposite: fewer guests than usual, yet the kitchen and waiters can barely keep up. In both situations, "whether it is full" alone says little. You need a simple number: how long a table is occupied and where the visit lengthens for no good reason.

Table turnover is how many times in a given time window (lunch, dinner, the whole day) a table — or a seat at the table — is served from seating to release. Do not confuse it with staff turnover or inventory turnover in the storeroom. It is about how long a guest occupies a table and how many times that same table takes the next party before the venue closes.

Below: how to calculate it from sales data, how to find the bottleneck (waiter, kitchen, reset, check), what really shortens a visit without ushering the guest out — and when to deliberately not speed up. No promises of magical revenue growth percentages.

A full floor does not guarantee results — what table turnover is about

A full floor means occupied seats. It does not tell you whether a guest spends 70 or 140 minutes, whether an empty table sits between reservations for half an hour, or whether the check closes quickly or the waiter is hunting for a terminal.

A manager who only watches occupancy easily confuses three different problems:

• low turnover with a full floor — visits are long or table release drags on; the queue grows despite occupancy;
• high turnover, weak check — tables free up quickly, but the guest never gets to add dessert or wine, and the kitchen runs on the edge of errors;
• empty seats despite reservations — the calendar looks full, yet gaps sit between slots and nobody closes them.

Turnover helps you see where the floor stalls. The goal is not "as many turns as possible at any cost," but as much rotation as the venue can carry without spoiling the guest and without cutting margin on dishes that need time (alcohol, dessert, celebration).

How to calculate turnover

You do not need a separate "turnover" report to start. What you already have in the sales system is enough: the moment a table check opens and the moment it closes.

Open-to-close check time

The simplest proxy for visit length:

average check time = sum (close time − open time) / number of closed table checks

in a chosen window (e.g. lunch 12:00–15:00 on weekdays).

This is not a perfect "time at the table from entry to exit." The system usually does not see the queue before seating or the reset after the check. It does show how long a table is occupied in sales — and that is exactly what blocks the next guest.

Example (illustration, not a market benchmark):
you have 20 lunch checks; the sum of open-to-close times is 1,900 minutes.
Average check time = 1,900 / 20 = 95 minutes.

If a week later, with the same number of tables and a similar menu, the average drops to 75 minutes, something in the flow changed — for the better, or by shortening the experience (e.g. dessert fell out of the rhythm). The number alone does not judge quality; for that you need floor context.

How many times a table takes guests at lunch / dinner

Second metric:

table turns in the period = number of closed checks assigned to a table (or zone) / number of tables in that zone

or, if you prefer seats:

turns per seat = number of guests (or checks × average party size) / (number of seats × number of hours in the window) — that is G / (seats × hours). You divide by the product of seats and hours, not by seats alone. Be careful with assumptions about party size.

Example (illustration):
lunch 3 hours, 15 tables, 36 closed checks in the zone.
Turns = 36 / 15 = 2.4 (on average a table takes guests about 2.4 times at lunch).

How to read the result:

• office lunch / casual — shorter visits, higher turnover can be desirable if the kitchen keeps up and the guest does not feel rushed;
• casual dinner — a longer check is normal; compare yourself to yourself (same day of week), not to lunch;
• celebration dinner / "occasion" venue — low turnover is often a feature of the model, not a failure; what matters is revenue per visit and guest return, not how many times the table freed up.

What the sales system does not see

Leave two measurement gaps in the manager's notebook:

1. Time from guest departure to table ready — clearing dishes, wiping, resetting covers, sometimes waiting for the person who resets. In the system the table is already free or still occupied — depending on when the waiter taps release. A delayed tap inflates "check time"; tapping before reset gives a false picture of a free table.
2. Queue and wait before seating — the guest stands at the bar for 15 minutes, and the check starts only at the table. System turnover looks better than what the guest feels.

That is why a time export alone is not enough. It also helps to run 1–2 shifts with a note: how many minutes a table really sits dirty before it is ready for the next guests.

Revenue per seat per hour — an extra metric (RevPASH)

To start, average check time and table turns are enough. When you want to set time against money, you can add a second metric: how much revenue one available seat brings in during one hour the venue is open.

In hospitality literature this metric is called RevPASH (Revenue Per Available Seat Hour) — revenue per available seat-hour. The definition comes from restaurant revenue management (including HSMAI / Cornell). You do not need it in the first week of measurement; it is a complement, not a replacement for turnover.

Formula (per HSMAI):

RevPASH = venue revenue (in the period) / (number of available seats × open hours in the period)

Example (illustration):
day: PLN 12,000 revenue, 40 seats, 8 open hours.
RevPASH = 12,000 / (40 × 8) = 12,000 / 320 = PLN 37.50 per seat-hour.

What that means in practice:

• you can have high turnover and a low result per seat-hour — tables free up, but checks are thin;
• you can have lower turnover and a higher result — a longer visit with wine and dessert "earns" the seat better than a quick lunch with no add-ons;
• a result calculated for the whole day blurs lunch and dinner — count windows separately if you run both modes.

Turnover shows how fast a table returns to availability for guests. This metric says whether that rhythm actually shows up in the till. If you prefer one number to start, stay with average check time.

Four places where a visit lengthens: waiter, kitchen, reset, check

When average check time is too long or the queue grows with a full floor, do not guess. Check four zones.

For 1–2 shifts a simple list is enough:

1. Log 10 checks: open, first dish on the table, check request, close, table ready for the next guest.
2. Count which segment eats the most minutes.
3. Pick one zone to improve next week — not all at once.

What really shortens a visit without ushering the guest out

It is not about slipping the check between courses. It is about removing empty waiting: for the menu, for the waiter, for the kitchen, for the terminal.

Ordering and add-ons

The sooner the guest has a menu (or knows the lunch set) and the smoother the waiter takes the order with variants, the less dead time at the start. Add-ons (another round of drinks, a side) should not require hunting for staff for a quarter of an hour — that lengthens the visit with no value for the guest and no control for the floor.

On the floor a shared table view helps: which is free, occupied, who is asking for the check, how long since the order was placed. In OrderNow that view comes from the waiter panel: table grid, statuses, add-ons on an open check.

Orders in the kitchen

The floor can run smoothly and the table still sits for an hour if the kitchen loses priority or a comment. An ordered queue and prep status do not shorten conversation at the table — they shorten the wait for the plate. A fuller picture of the process from table to settlement is in the piece on a restaurant system with waiter service.

Fast settlement and table release

The last 10–15 minutes of a visit are often not "experience," only logistics: check request, payment, receipt, goodbye, reset. Here turnover gains the most without pressure on a guest who is still eating dessert.

In practice:

• the waiter must know right away that a table is asking for the check;
• settling cash or card and marking the check paid should close the table in one move;
• releasing the table in the system must go together with a real surface reset — otherwise the next guest sits down to a mess, or the floor map lies.

In the OrderNow waiter panel, settlement and table release sit in one flow (cash or card payment, check close, table returns as free). The guest can also send a check request from their phone via QR menu — the waiter gets a notification instead of scanning the room by eye. That does not replace payment at the table; it speeds the moment the team knows it is time to close the check.

Reservations: time windows and gaps between visits (a separate matter)

This is different from floor service. Reservations decide when a guest is entitled to take a table and how large the holes between visits are. Service decides how long the table is occupied once the guest is seated.

Industry writing often mentions time windows (e.g. about 2 hours per slot) and fitting the next reservations so an empty table does not sit between departure and the next arrival. That is sound occupancy planning — provided the slot matches your real visit time and you are not copying someone else's scheme.

What not to do based on reservation marketing: do not treat promised revenue lifts of tens of percent from blogs and system comparisons as fact. Those numbers circulate in reservation content, but they rarely have an independent, primary source for your venue. Treat them as a sales line, not as your forecast.

In OrderNow table reservations mean an online form, party and hour limits, auto-confirm, a calendar, and a view of upcoming reservations on the waiter panel. The system does not take deposits for a reservation and does not auto-seat guests on the floor plan — seating and floor organization stay with the team. Reservations organize the calendar; they do not replace the host or the waiter.

Two separate matters side by side:

• Reservations — fewer gaps between slots, better fit of the next visits, fewer surprise walk-ins at peak (or deliberate windows for walk-in guests).
• Floor service — faster ordering, a readable kitchen, a smooth check and table release.

Fixing only the calendar with a slow kitchen, or only the kitchen with holey reservations, leaves half the problem.

When not to speed up turnover

Faster rotation hurts when:

• the model rests on a long visit — tasting menu, an "occasion" venue, a celebration dinner; the guest pays for time and rhythm, not for throughput;
• high margin sits in the second half of the visit — wine, a drink, dessert, coffee; forcing a shorter stay cuts exactly what builds the check and revenue per seat;
• the guest is celebrating — birthday, business dinner, date; a sense of rush comes back as a weak review and no return visit;
• the kitchen is already at the ceiling — tightening turnover without pass capacity ends in errors, dishes going cold on the pass, and shift stress;
• you already cut empty waiting — further speeding up means pressure on the guest, not on the process.

First cut empty waiting (menu, waiter, kitchen, check). Touch the time when the guest sits, eats, and finishes drinks (wine, dessert, conversation) only when you have a queue, the kitchen keeps up, and shortening does not kill margin from add-ons.

An office lunch with a queue at the door is a different world from a Saturday dinner with full glasses. The same "95-minute average check" can be a problem in the first scenario and a target in the second.

Manager checklist for 7 days

Day 1–2 — measure
Pull open-to-close check times from the sales system separately for lunch and dinner (same day-of-week type). Calculate average check time and turns per table in the zone. Optionally note revenue per seat and hour (revenue / seats / hours) — as context, not as the only goal.

Day 3 — floor check
Walk the four-bottleneck table. For 8–10 tables log segments: seating–order, order–dishes, last dish–check request, check–table ready.

Day 4 — one decision
Pick one zone and one shift (e.g. lunch only Tuesday to Thursday). Examples: earlier menu or lunch set delivery; a clear "check request" signal; a dedicated person for table prep at peak; cleaning up the kitchen queue.

Day 5–6 — introduce the change
No full floor rebuild. A short brief for waiters and kitchen: what we change and what we deliberately leave alone (e.g. we do not shorten dessert).

Day 7 — review
Compare average check time and turns with day 1–2. Also check the queue, reviews, and whether average check fell because add-ons were cut. If the number improved and the floor atmosphere got worse — ease the pressure and keep looking in logistics, not at the guest.

If you want to set turnover against what the guest orders from the menu (not how long they sit), a separate topic is menu engineering on sales data — that is a different problem from table time.

Table statuses, check, and reservations in one workflow

If you want table statuses, time since the order was placed, settlement with table release, a view of reservations on the floor, and optionally a check request from QR menu in one waiter workflow — see the waiter panel and table reservations. OrderNow does not promise "+X% turnover"; it organizes the floor work you still have to measure in your own venue.

Start with average check time from last week. One number and one zone where the visit lengthens usually tell you what to fix first.

Sources

• HSMAI Academy: RevPASH — definition of revenue per available seat-hour
• HorecaBC: restaurant operations KPIs — measurement in a venue
• OrderNow: Waiter panel — table statuses and settlement on the floor