How many tables can one server handle? Work out your own section size

There is no verified standard for how many tables one server can handle. Staffing blogs quote neat numbers, but we found no study behind them. A section size comes from your own minutes: how long a table stays, how many minutes of a server’s time it needs, and how much slack you want for the moments that go wrong. This post gives the formula, a worked example with every assumption stated, and the factors that move the answer up or down.

Is there a standard number of tables per server?

Short answer: no. The numbers you’ll find online are rules of thumb without a study behind them.

That doesn’t make them useless, but a rule of thumb from someone else’s restaurant ignores everything that makes yours different: how far the kitchen is, whether anyone runs food, how big your parties are and how many courses you serve. A server with four 2-tops near the pass and a server with four 8-tops up a flight of stairs are not doing the same job.

There is also a guest-side reason not to stretch sections. In the National Restaurant Association’s 2024 survey, 64% of full-service diners said they would choose “traditional service from employees” over a restaurant that “incorporates technology” (NRA, 2024). Guests notice when their server has too many tables.

What decides how many tables a server can handle?

Short answer: anything that changes the minutes each table needs or how long tables stay.

Notice that faster turns make a section harder, not easier: every new party needs a greeting, a first order and a check. A busy bistro turning 2-tops every 45 minutes can be more work than a slow dining room full of 6-tops.

How do you work out section size?

Short answer: with one line of arithmetic, once you have measured your own minutes.

Tables per server = share of time busy × minutes a table stays ÷ server minutes per table

• Minutes a table stays: from seating to leaving, on your busy shift.
• Server minutes per table: every minute the server spends on that table or walking for it, across the whole visit.
• Share of time busy: nobody can be busy 100% of the time. Two tables will want something at the same moment, and side work doesn’t stop. Using 0.8 keeps a fifth of the time spare. That is an assumption; choose your own.

Round down. A section of 3.6 tables is three tables, with help on the fourth.

A worked example at Casa Olivo

Short answer: four tables in the base case, ranging from three to five as the factors change.

Casa Olivo is a fictional restaurant (an example, not a customer). The minutes below are assumptions to show the method, not measurements or standards.

In the base case, a table stays 75 minutes and needs 15 server minutes: 3 to greet and take drinks, 3 to take the food order, 3 to carry food, 1 to check back, 2 to clear and 3 for the check and payment.

The same server, on the same night, can sensibly take anywhere from three to five tables depending on how the room is set up. That’s why copying a number rarely works.

How do you measure server minutes per table?

Short answer: with a stopwatch and a tally sheet, on two busy shifts, for a handful of tables.

1. Pick two busy shifts and one experienced server who agrees to be followed.
2. Follow four or five tables from seating to leaving. Write down the time seated and the time they leave.
3. Tally every visit to the table and roughly how long it took: greeting, drinks, order, food, check-back, clearing, dessert, check, payment.
4. Add walking. Time one round trip from the section to the pass, the bar and the till, and add a trip for each visit that needed one.
5. Average it. Total server minutes ÷ number of tables, and total minutes stayed ÷ number of tables.
6. Run the formula, then sanity-check it against what the server says felt manageable.

Repeat it after any change: a new menu, a runner on Fridays, a new ordering tool. The number is only as good as your last measurement.

How many servers should a restaurant have?

Short answer: tables occupied at your busiest hour, divided by section size, rounded up, plus cover for breaks and sick calls.

At Casa Olivo, 20 tables are full at the Saturday peak. With a section size of 4, that is 20 ÷ 4 = 5 servers on the floor. Add one for breaks and a sick call and the rota says 6. At a Tuesday lunch with 8 tables occupied at peak, 8 ÷ 4 = 2, plus one, makes 3.

Do this by shift, not by week, and check what the rota does to your labor cost percentage. Your till’s sales by hour, or your ordering tool’s busy-hours report, tells you when the peak actually is, which is often half an hour later than the rota assumes.

How many servers do you need for 150 guests?

Short answer: work out the tables first, then use a section size for the style of meal.

Events work the same way, with bigger tables. These ratios are assumptions to show the method, not standards.

Add bartenders and a captain or manager separately. If the plated courses all go out at once, add runners for the main course; that is the moment sections overflow.

Who does what on the floor?

Short answer: the more of the carrying and clearing others do, the more tables a server can look after well.

Which tasks can come off a server’s section?

Short answer: the ones that don’t need a person at the table, such as taking a routine order, re-explaining the menu or fetching the check.

A Cornell report timed servers in one restaurant: time per table fell from 10 minutes 16 seconds to 3 minutes 30 seconds when guests ordered and paid on a tabletop device (Susskind and Curry, Cornell, 2016, p. 6). It was one restaurant and only 7 tables in that group, so read it as a hint about where the minutes go, not a promise.

With WaiterPilot, guests scan the QR code on their table, ask menu questions and order by chat in their own words, and the chat answers from your menu. They call a waiter or ask for the check from their phone, and the check request says cash or card, so the server brings the card machine with the check. What stays with your team:

• Confirming each table’s first order on the staff board; later orders from that table go straight to the kitchen until staff close the table.
• Carrying food, checking on guests and clearing.
• Allergies. WaiterPilot gives what the menu says and asks the guest to confirm with staff; your team always has the final word.
• Payment. WaiterPilot takes no payments, and it has no POS connection, so someone rings orders into your till if you use one.
• Watching the board. Waiting orders and calls change colour at 1, 5 and 10 minutes, so somebody needs to be looking.

We don’t promise bigger sections. If guests ordering from their phones frees minutes, you choose how to spend them: more time with each table, a calmer night when someone calls in sick, or a section change you have measured first. Our list of jobs to take off servers’ plates goes through nine of these tasks, and how it works shows the staff screens.

What to do next

1. Time four or five tables on two busy shifts, using the six steps above.
2. Put your numbers into the formula and round down.
3. Check the result against the busiest hour, not the average.
4. Re-measure after any change to the menu, the team or your tools.

For the night when every section fills at once, use the rush-hour playbook. If service feels slow even with sensible sections, the 9-point diagnosis of slow service finds the gap, and waiter call and bill shows how guests reach staff from the table. To weigh what a tool costs against what your floor costs, see pricing.