The Busiest and Slowest Days for Restaurants, and How to Staff Each One

What are the busiest days for restaurants?

The busiest days for restaurants are the days at the top of each restaurant’s own sales report, and no government source ranks them nationally. The U.S. Census Bureau counts restaurant sales by month, not by day of the week. So any busiest day for restaurants you read about online, including the usual answer to what is the busiest day of the week for restaurants, was measured on other restaurants’ sales, not yours.

Your own ranking depends on what brings people through your door. A dinner restaurant fills up when people have evenings off. A lunch spot near offices lives on the working week. A brunch place lives on the weekend. Two restaurants on the same street can have opposite busy days, so the only ranking worth staffing to is the one in your own sales.

For example, let’s say a full-service restaurant sells $100,000 in a normal week. Averaged over eight normal weeks, its days look like this:

Schedule every weeknight the same, and you pay for Monday as if it sold like Thursday. The ranking is the first thing to write down, because every staffing decision after it runs off those seven numbers.

Delivery has its own busy days. Delivery orders do not always peak on the same days as the dining room. So pull them from your reports as their own line and rank them separately. A night that is quiet at the tables can still keep the kitchen busy.

Ranking your days is one part of building a week. The full restaurant employee scheduling guide covers the rest of the job.

What is the slowest day of the week for restaurants?

The slowest day of the week for a restaurant is the day at the bottom of its own daily sales, and that least busy day is the easiest one to overstaff.

Restaurant overstaffing on a slow day happens for a simple reason. A slow night is usually copied from a normal weeknight. Nobody notices an extra person with nothing to do the way they notice a slammed rush with nobody free. The quiet cost never draws a complaint, so it never gets fixed.

For example, let’s say the Monday in that week sells $8,000, the smallest day of the seven, and its schedule was copied from a busier night.

How to bring guests in while a restaurant is slow

Once a slow day is staffed to its sales, you can work on growing it. Common ideas are a set menu for the slow nights, a regulars’ night, and a push for the slow days on the email list and social pages you already have. Try one idea at a time. Compare that day’s sales for the next four weeks with the four weeks before. Then add staff only when the sales actually move, so restaurant slow days never get their extra people before they have the extra guests.

What overstaffing slow days and understaffing busy days costs a restaurant

Staffing every day the same costs a restaurant twice in the same week. It pays for hours nobody needed on the slow days. It runs the busy days short. And the weekly total barely moves, so neither problem shows up in the number you check.

The way to see it is to measure staff hours against sales. Take the busy day that ran best, with no long waits and nobody standing around. Divide its staff hours by its sales in thousands. That gives you staff hours per $1,000 of sales, your own measure of how many people your sales actually need.

For example, let’s say that restaurant’s best Saturday sold $23,000 on 230 staff hours. That is 10 staff hours for every $1,000. Now compare what each day needs with how the week is really built, where one normal weeknight is copied from Monday to Thursday.

Let’s say the average wage across the kitchen, the floor and the bar is $20 an hour. The week as built holds 1,020 hours, or $20,400 of wages. The sales need 1,000 hours, or $20,000. The week is only 2 percent over, so nothing looks wrong.

But Monday, Tuesday and Wednesday carry 80 hours nobody needed. At $20 an hour that is $1,600 a week, and $83,200 a year, at one restaurant. Across five locations it is $416,000 a year.

And Thursday to Saturday run 60 hours short. That is six 10-hour shifts missing on the three nights that pay for the week. The guests wait longer, and a manager ends up working the line instead of running the room.

Move those hours to the days that sell, and the week costs about the same while the rush gets its people.

If you also want a hard dollar ceiling on your slow days, the guide on how to reduce restaurant labor costs shows how to set one.

How to find your restaurant’s busiest and slowest days

To find your restaurant’s busiest and slowest days, average each weekday’s sales over your last eight normal weeks, rank the days, and multiply each day by the staff hours your best-run busy day needed per $1,000 of sales. It takes five steps and a spreadsheet, and it is restaurant sales forecasting in its simplest form, built from your own register instead of a guess.

1. 1

   Pull daily sales for eight normal weeks

   Export daily sales from your register reports for the last eight weeks. Leave out holidays, closures and big local events, because each of those gets its own plan. For example, let’s say that leaves you 56 ordinary days.

2. 2

   Average each weekday and rank the days

   Add up the eight Mondays and divide by eight, then do the same for every other day. Put the seven averages in order. For example, let’s say Monday comes out lowest at $8,000 and Saturday highest at $23,000.

3. 3

   A simple restaurant sales forecast from your own register

   Start each day of next week at its average. Move a day up or down for anything you already know about, such as a holiday, a home game or a patio closed for the weather. For example, let’s say there is a home game on Friday, so Friday’s forecast goes above its $20,000 average.

4. 4

   Find your staff hours per $1,000 of sales

   Pick the busy day that ran best, with no long waits and nobody standing around. Divide its staff hours by its sales in thousands. For example, let’s say that Saturday used 230 staff hours on $23,000, which is 10 hours per $1,000.

5. 5

   Multiply each day’s forecast by that number

   Each day’s forecast in thousands, times your hours per $1,000, is the staff hours that day needs. For example, let’s say Monday is forecast at $8,000, so it needs 80 hours, and Saturday needs 230.

A forecast that lives in the manager’s head changes with whoever built last week’s schedule. Seven numbers from your own sales, and every day gets the hours it actually needs.

Once a week is published, check what it cost. XShift AI’s Workforce Insights shows each week’s labor cost, scheduled hours, overtime hours and overtime premium, person by person, counting published shifts only. Set that beside the week’s sales and you can see whether the hours went where the money came from.

Restaurant peak hours: where the rush sits inside each day

Restaurant peak hours are the few hours each day when most of the day’s sales land, and you find them in your register’s sales-by-hour report, not by looking at the clock. A busy day is not busy all day. It has a rush, a lull and a tail.

Pull sales by hour for your two busiest days and group the hours into blocks. Then staff each block, not the whole day. People start as the rush builds and go home as it fades. A mid shift from 4 to 9 puts someone on the floor for the whole dinner rush without paying them to stand through the afternoon lull.

For example, let’s say that restaurant’s $23,000 Saturday splits like this:

Almost half of that Saturday’s sales land between 5 and 8. Under a tenth land between 2 and 5.

Everyone clocked in from 11 until close means the 2 to 5 lull is fully staffed while almost nothing sells. Start people as the rush builds and send them home as it fades, and the same hours cover more of the rush. The hours that ranked your days now shape the shifts inside them.

Is January a slow month for restaurants?

Yes, January is a slow month for restaurants. In U.S. Census Bureau sales for food services and drinking places, the average January day sold 11 to 12 percent less than the average day of the year in 2023, 2024 and 2025. It was the slowest month per day in all three years.

Count sales per day, not per month. February reported a smaller monthly total than January in 2023 and in 2025. That is only because February is shorter. Per day, February outsold January in every one of those years.

These are sales dollars, not guest counts, and dollars drift up through a year as prices rise. The drop still shows against the month right before it. The average January day in 2025 sold about 8 percent less than the average December day before it.

Are restaurants slow during September?

Nationally, September is close to an average month. Its sales per day sat within 3 percent of the year’s average day in 2023, 2024 and 2025. It fell from August in 2024 and 2025, and it rose in 2023. Whether your September slows depends on what drives your own traffic, such as school starting, summer visitors leaving or a college coming back. Check your own September against your own August.

For example, let’s say that restaurant’s January weeks sell $90,000 instead of $100,000. At 10 staff hours per $1,000, a January week needs 900 hours. Running January on a normal 1,000-hour week adds 100 hours.

At $20 an hour that is $2,000 a week, and $8,000 over four January weeks, in wages for guests who stayed home. Build January weeks to 900 hours and that $8,000 stays in the business. Rank your months once a year, the same way you ranked your days.

The busiest restaurant days of the year, and how to staff them

The busiest restaurant days of the year are the holidays people celebrate with a meal out, such as Mother’s Day. The National Restaurant Association says Mother’s Day has historically been one of the most popular holidays to celebrate at restaurants. In its 2024 survey, 43 percent of consumers planned to use a restaurant to celebrate it.

Big days come in three kinds, and each is planned from a different comparison. Always compare a big day with the same day last year, not the same calendar date. A holiday that moves lands on a different date every year, and a fixed date lands on a different weekday.

For example, let’s say last year’s Mother’s Day sold $32,000, twice a normal Sunday. At 10 staff hours per $1,000 it needs 320 hours, not the 160 a normal Sunday gets. Staff it like a normal Sunday, and the biggest day of that restaurant’s spring runs on half the people it needs.

In XShift AI you save that whole week once as a schedule template. It keeps every shift’s day, times, location, headcount and role counts, and no names. Next year you lay it down again from the new start date and fill it with this year’s people. Last year’s holiday week comes back in one step, with every shift and every role count already in place.

Set a different headcount for every day in XShift AI

XShift AI stores a different headcount for every day and every shift, so a Monday dinner can need five people and a Saturday dinner eleven, and nobody has to remember the difference. That is restaurant staffing by day of the week, written down where the schedule can read it.

In a spreadsheet or an outdated restaurant scheduling tool, the difference between Monday and Saturday lives in the manager’s head, and it gets copied wrong every week.

That number can live in three places. The first is a sentence typed to the AI Copilot, which changes the shifts you already have. The second is role counts on those shifts, so the headcount is a mix of jobs rather than one number. The third is a standing staffing rule for each location, which sets the counts by day and by hour.

For example, let’s say the restaurant from the week above sets up its busy and slow days like this:

Type it once, and every Monday and Tuesday 5 PM shift in January needs four people until you change it.

Every location keeps its own staffing rules. A downtown location that lives on weekday lunches and a suburban location that lives on weekend dinners each carry their own busy days. One location never borrows another’s counts. That is what restaurant scheduling software has to do for a group: the seven numbers from each restaurant’s own sales now live inside its schedule.

Let XShift AI fill every shift to its count

A headcount is half the week. The other half is putting the right people on each shift. By hand, every check in the card below is done from memory, for every name, on every shift, every week.

XShift AI’s schedule generator fills every shift in the dates you pick up to the headcount you set. It works role by role and in date order, and it runs sixteen checks on a name before that name goes on a shift. You choose Fair mode, which gives the next shift to the people with the fewest hours so far, or Max mode, which gives it to the people with the most. The ten rules come from Autopilot, and each one runs once you write it.

For example, let’s say the Saturday 5 PM shift needs 6 servers. Anyone who does not work at this restaurant, is on approved time off, has marked Saturdays unavailable, or would break an overtime rule you wrote is passed over. The next server who clears every check gets the place. Busy nights get their full count and slow nights get theirs in the same run, and nobody was phoned.

What you get back

What comes back is the wages that were going to your slowest days, and people on your busiest days, where the sales pay for them. In the example restaurant from this guide:

Next January is built to January’s sales with one typed sentence. Mother’s Day comes back from last year’s week in one step. The Saturday rush has its full count, and the Monday lull has only the people Monday pays for.

Starting is short. Tell the AI Copilot your location, your people and your repeating shifts in plain English, and it creates them in one step after you confirm. Or bring your team in by pasting a list, uploading a spreadsheet file or uploading a photo of your staff list, and check the preview before anyone is added. Get fully set up in minutes.

In that example, every week the days stay flat is another $1,600 spent on nights that did not need it. Book a demo and watch a week get built to its busy days.

Frequently asked questions