In this article
A kitchen brigade is written down: chef, sous-chef, chef de partie, commis, and everyone knows whose name belongs on which station. The floor is written down nowhere. "We'll manage with whoever's in" is the most common staffing rule in the trade, and it's exactly why Friday night always feels one server short.
There's a rule of thumb everyone knows and nobody can defend: four tables per server, or twenty covers per server, or whatever number has been circulating in your kitchen since you started. The problem isn't that the number is wrong — it's that it can't be right for every restaurant, because it knows nothing about your concept, your course count, your floor layout, or whether you run with a food runner.
This article works it out differently. The capacity of one server — how many covers that person can actually handle per hour — isn't a fixed number. It's a base rate for your service style, multiplied by four factors you already know: how many courses per cover, whether you run a dedicated runner, how your room is laid out, and whether tableside service is part of the job. From that follows how many servers your peak hour needs — not your whole shift.
The calculator below takes your own service: your concept, your busiest hour, your room, your labor cost and your average spend. It gives you the number of servers your peak hour needs, what that shift costs in wages, and — the part that's usually missing — what it costs you per year if you're structurally one server over or under. The answer isn't symmetric, and that's exactly why the question matters.
Everything runs in your own browser: nothing is sent or stored. Once you have a server count and a peak hour, turn it straight into a floor section plan for tonight — which server takes which tables, who covers the bar, who covers the door.
Why "four tables per server" is the wrong question
A flat ratio treats every table as the same amount of work. It isn't. A table that orders one main course and the bill costs a server two visits. A table on a seven-course tasting menu with wine pairing costs twenty — taking the order, explaining each course, pouring each glass, clearing each plate. The same four tables are comfortable in one restaurant and impossible in another, and the number "four" says nothing about which.
The second mistake isn't in the number, it's in the timing: almost every operator staffs to the shift's average, when the room is only actually full at 8:30. A three-hour dinner service with ninety covers total sounds like thirty an hour — but if sixty of those land between 8 and 9pm, your peak hour is the only hour that decides your staffing. Staff to the average and your floor collapses exactly when it matters.
And then there's the part almost nobody prices: what getting it wrong actually costs, in both directions. One server too many is a visible labor cost you'll see on next week's payroll. One server too few is an invisible one — tables that turn slower, shorter conversations about the wine list because nobody has time to make them, and in the worst case a guest who leaves after a ten-minute wait. Both deserve a number, and further down they each get one.
The ultimate guide Finding and Keeping Hospitality Staff: The Complete Guide From staffing levels to scheduling to retention: everything about your team in one guide. Open the guideThe 5 numbers, and what they tell you about staffing
They build on each other. The first two set what your floor can handle in theory, the next two adjust that to your own situation, and the last one puts a price on it.
1. Your base capacity depends on what kind of restaurant you run
Every service style has its own ceiling on how many covers one server can comfortably handle per hour, and the spread is wide: from around twenty covers an hour in a fast-casual counter-service concept down to barely six in fine dining with full tableside service. That isn't a judgment on skill — a fine-dining server is doing fundamentally different work, with more visits per table and more explanation per course.
Five starting points, in covers per server per hour: fast-casual and counter service around 20, café and bakery around 16, bar and pub with a kitchen around 14, bistro with full table service around 11, and fine dining around 6. These are starting points, not laws — a well-laid-out bistro sits higher, a fine-dining room pouring wine at every course sits lower.
Find your own concept on the ladder below and remember the number: everything that follows is a multiplication on it, never a replacement for it.
Covers per server per hour, before any of the four adjustments above. Find your own concept — everything else in this article is a multiplication on this number.
These are starting points from how most independent restaurants actually run, not regulatory standards. A particularly efficient bistro layout sits higher on the ladder; a fine-dining room pouring wine at every course sits lower.
2. You staff to the peak hour, never the shift total
Never work backwards from a shift's total covers divided by its hours — that number doesn't exist anywhere in the real room. A three-hour dinner service that's quiet at 7pm and slammed by 8:30 has one peak hour that decides your staffing, not three average ones.
Ask "how many covers in my busiest hour", never "how many covers per shift". At thirty covers in that hour and a bistro capacity of eleven per server, you need three servers to handle that hour — even though the rest of the shift never comes close to that pace. If you don't know your own peak hour off the top of your head, a covers forecast built on your own reservation data gives you a sharper number than a guess.
This is the single most common mistake in restaurant staffing: a rota that adds up on paper for the whole shift, and still falls apart every Friday at 8:45.
3. A dedicated runner adds about a third to your capacity
A server who carries every plate to and from the kitchen themselves loses a large share of their time to walking rather than serving. Add a dedicated food runner or busser whose only job is bringing food and drink to the table and clearing it, and the same server can handle roughly 35% more tables — not because they work faster, but because they stop walking and start serving: taking orders, making recommendations, bringing the bill.
Run that on the bistro capacity of eleven: with a runner it becomes roughly fifteen covers per server per hour. At the same thirty covers in your peak hour, you now need only two servers instead of three — one fewer labor cost, every busy shift.
That makes a runner one of the cheapest staffing decisions there is: it costs less than a full extra server, and it raises the capacity of everyone already on the floor.
4. A sprawling room and tableside service both cost you capacity
Two things pull the other way. A compact room with every table a few steps from the bar and the kitchen keeps your base capacity intact. A spread-out room — a terrace on the other side of the building, multiple floors, a garden out back — costs a server walking they wouldn't lose in a compact room: budget roughly 20% less capacity per floor or separate space.
Tableside service — gueridon service, carving meat at the table, finishing a sauce in front of the guest — costs another roughly 25%, because every step that would normally happen in the kitchen now happens in front of the guest and takes longer as a result.
Neither is a reason to cut them — a spread-out terrace and tableside service are often exactly what sets a restaurant apart — but both are a reason to count them into your staffing, rather than discovering afterwards that the terrace needed its own server.
5. The cost of getting it wrong runs in two directions, and they aren't equal
One server too many is the mistake operators avoid hardest, because the cost is visible — it shows up on next week's payroll. One server too few feels free, because no invoice arrives for it — the cost hides in tables that turn slower, guests who order less because nobody had time to pitch the wine list, and in the worst case a guest who leaves after a ten-minute wait.
Run both for the same bistro from the steps above — thirty covers in the peak hour, eleven per server, three servers needed — and the pattern is strikingly consistent: at a fully-loaded labor cost of €19 an hour, a three-hour peak service, and three services a week hitting that same peak, one server too many costs roughly €8,900 a year in wages that never needed spending.
One server too few — two instead of three, capacity dropping to twenty-two against thirty covers arriving — leaves a shortfall of eight covers in that hour. Assume conservatively that forty percent of that is actually lost to waiting or walking out, at an average spend of €36, and that shortfall runs to roughly €18,000 a year — more than double the cost of one server too many. The calculator below runs both numbers side by side for your own service.
Thirty covers in the peak hour, eleven per server. Two servers is too few, four is too many — and the two mistakes don't cost the same amount.
Amounts per year, at a labor cost of €19/hour, a 3-hour peak service, 3 services a week hitting that peak, and an average spend of €36. The shortfall conservatively counts 40% of missed covers as actually lost — the rest wait longer or order less, which isn't counted here.
Run your own service through it
Enter your concept, your peak hour and your room and you get the number of servers that peak needs, straight away. The fields start filled in with the bistro example above so you can see how it reads — overwrite them with your own numbers.
Below that sit three amounts: what that shift costs in wages at the right headcount, and what it costs you per year if you're structurally one server over or under. Compare those two before you draw a conclusion — the answer is almost never symmetric.
Staffing calculator
Your concept, your peak hour, your room — and the difference in money per year.
—
The base rates per concept and the four adjustments are working figures from how independent European restaurants actually run, not regulatory standards — your own room, your own team and your own guests are the final judge. Everything runs in your browser; nothing is sent or stored.
Got a server count for tonight? Turn it straight into a floor section plan: which server takes which tables, who covers the bar, who covers the door — that's a different question from this one, and the section plan answers it in a few clicks once you have your number here.
And don't forget the kitchen: the same peak that decides how many servers you need also decides how many cooks. The kitchen brigade guide is this article's mirror image, for the other side of the pass.
What to do with this this week, this month and this quarter
Overhauling your whole staffing model in one go works for nobody. This order does, because each step makes the next one measurable.
This week — know your own peak hour
- Count three Fridays in a row how many covers arrive in your busiest hour, not your whole shift divided by its hours.
- Find your concept on the capacity ladder and work out your base capacity with your own course count, runner, room and tableside service.
- Compare that number with who's actually on the floor on Friday — most restaurants are closer to the right number than they think, or further off than they'd like to admit.
This month — price the mistake in both directions
- Fill in the calculator above with your own labor cost, average spend and number of peak services a week.
- Compare the cost of 1 server too many against the risk of 1 too few — that decides which way you round when unsure.
- Turn your server count into a floor section plan so the staffing stops being just a number and gets a name per section.
This quarter — build it into your rota
- Use a covers forecast built on your own reservations instead of a fixed peak assumption per weekday.
- Consider a dedicated runner if you're structurally on the edge of being 1 server short — it's usually cheaper than a full extra server.
- Repeat the exercise every season: a summer terrace or a winter menu with more courses shifts your base capacity, even if your guest count stays flat.
Staffing isn't a feeling, it's a multiplication
Almost every restaurant that runs this finds the same thing: the problem was never "not enough staff" in general — it was a peak hour that was never looked at on its own, or a runner that was never hired, or tableside service that was never counted into the staffing.
Four tables per server isn't a bad rule of thumb because the number is wrong — it's a bad rule of thumb because it isn't a number that belongs to your restaurant at all. Your own base capacity, multiplied by your own courses, your own runner, your own room and your own tableside service, is.
Do the same thing on the other side of the pass with the kitchen brigade, and turn both into an actual rota with staff scheduling — that's the whole team, not just the floor.