Staffing, Hiring and Training

Labor Forecasting From Sales and Covers, Hour by Hour

Forecast restaurant labor by hour from last year's sales, the reservation book, events and weather, then convert covers into shifts with labor standards.

AssistCall research teamPublished Reviewed by Faruque Amin, VP of Sales and Marketing7 min read

On this page
  1. Why labor forecasting pays for itself
  2. The four inputs to a covers forecast
  3. Turn covers into labor with standards
  4. The forecasting routine
  5. Example: forecasting a Friday dinner
  6. Measure forecast accuracy
  7. Forecasting takeout, delivery and private events
  8. When software helps
  9. Common forecasting errors
  10. Frequently asked questions
  11. References
Staffing, Hiring and Training: Labor Forecasting From Sales and Covers, Hour by HourArticle coverSTAFFING, HIRING AND TRAININGLabor Forecasting FromSales and Covers, Hourby HourAva for RestaurantsInsightsStaffing, Hiring and Training: Labor Forecasting From Sales and Covers, Hour by HourArticle coverSTAFFING, HIRING AND TRAININGLabor ForecastingFrom Sales andCovers, Hour by HourAva for Restaurants

Key takeaways

  • Forecast covers by hour first; labor follows from covers using your own productivity standards.
  • Combine four inputs: same day last year, recent trend, reservation pace and known events or weather.
  • Separate fixed positions from variable ones so the forecast moves only the labor that should move.
  • Measure forecast error every week; a forecast you never check never improves.

Why labor forecasting pays for itself

Labor is one of the two largest lines on a restaurant P&L. The National Restaurant Association estimates that food and labor each take roughly 33 cents of every sales dollar[1], and average hourly earnings of restaurant employees have risen 41% since February 2020, faster than the 35% rise in wholesale food prices over the same period[2]. When each hour of labor costs more, scheduling an extra server for a quiet hour, or one too few for a rush, matters more than it used to.

Restaurant cost growth since February 2020
Restaurant cost growth since February 2020Comparison. Average hourly earnings of restaurant employees: 41%; Average wholesale food prices: 35%.41%Average hourly earnings of restaurantemployeesAverage hourly earnings of restaurant employees: 41%35%Average wholesale food pricesAverage wholesale food prices: 35%Restaurant cost growth since February 2020Comparison. Average hourly earnings of restaurant employees: 41%; Average wholesale food prices: 35%.41%Average hourly earnings ofrestaurant employeesAverage hourly earnings of restaurant employees: 41%35%Average wholesale food pricesAverage wholesale food prices: 35%

Source: National Restaurant Association, 2026[2]

Show the data
ItemValue
Average hourly earnings of restaurant employees41%
Average wholesale food prices35%

Average hourly earnings for all employees in food services and drinking places reached $22.25 in September 2026, up from $20.69 two years earlier[3]. A labor forecast turns those rising costs into a plan: it tells you how many people each station needs in each hour, before you write the schedule. Done well, it improves service and margin at the same time, because labor sits where the guests are.

The four inputs to a covers forecast

Forecast covers, or orders for counter service, rather than sales. Covers translate directly into work: seats to greet, plates to cook, glasses to wash. Sales per cover changes with menu prices and promotions, so a sales-only forecast can mislead you about how busy the floor will be.

  1. Same day last year: covers by hour for the matching weekday, not the matching date. The second Friday of March last year is a better guide than March 10 if March 10 was a Tuesday.
  2. Recent trend: compare the last four to six weeks with the same weeks last year. For example, if you are running 6% ahead, apply that factor to the baseline.
  3. Reservation pace: compare covers already booked with the number booked at the same lead time last year. Pace tells you about this specific night in a way the trend cannot.
  4. Known events and weather: a home game, a convention, a school holiday, a road closure, rain on a patio night. Adjust specific hours, not the whole day.

Seasonality also shows up in the workforce itself. Not seasonally adjusted, US restaurant and bar employment peaked at 12.64 million in June 2026, about 637,000 more than in January[4]. If your local market swings the same way, your forecast should anticipate the hiring needed ahead of the busy months, not only the shifts.

Turn covers into labor with standards

A labor standard is the amount of work one person can do well in an hour at each station. Set your own from your history: look at your smoothest busy shifts and count covers per server hour, tickets per line cook hour and covers per dishwasher hour. Standards differ by concept. A fine-dining server may carry far fewer covers an hour than a casual-dining server, and that is by design.

Fixed, step-variable and variable restaurant positions
Position typeHow it is scheduledExamples
FixedScheduled whenever the restaurant is open, regardless of volumeOpening manager, one host, one dishwasher, one cook per open station
Step-variableAdded in whole people when covers cross a thresholdSecond host above a set number of covers per hour, a runner on peak nights
VariableScaled directly to the forecast using the labor standardServers, additional line cooks, bussers, bartenders on service bar

Splitting positions this way stops the forecast from doing silly things. Your opening manager is not optional on a slow Monday, and a second host is not needed until the door gets busy. Only the variable positions should move hour by hour with the covers.

The forecasting routine

  1. Pull the baseline

    Export covers by hour for the same weekday last year from the POS or reservation system, for each meal period.

  2. Apply the trend

    Multiply the baseline by the ratio of the last four to six weeks this year to the same weeks last year.

  3. Check reservation pace

    Compare booked covers to last year's at the same lead time. If pace is well ahead or behind, adjust the peak hours more than the shoulders.

  4. Adjust for events and weather

    Add or remove covers in the specific hours affected. Keep a notes column so you learn how much a home game or a rainy evening really changes things.

  5. Convert to labor hours

    Divide variable covers by the labor standard for each position, round to whole people per hour, and add fixed and step-variable positions.

  6. Compare to the labor target

    Price the labor hours at average wage rates and compare with forecast sales. If labor is above target, look first at shoulder hours and staggered starts.

  7. Hand it to the schedule

    Give the hourly grid to whoever writes the schedule, so shifts follow the curve.

Example: forecasting a Friday dinner

Example (fictional): a 100-seat bistro is forecasting the Friday dinner two weeks out. Last year's matching Friday had 180 covers. The last six weeks are running 5% ahead of the same weeks last year, which gives 189. Reservation pace is ahead too: 120 covers are already booked, against 100 at the same lead time last year, so the manager adds most of the extra to the 7 to 8 p.m. peak and forecasts 196. A local festival ends at 9 p.m., so she adds ten late covers, for 206.

Example hourly covers forecast and server needs (fictional)
HourLast year coversForecast coversServers needed at 12 covers per server hour
5-6 p.m.22232
6-7 p.m.45484
7-8 p.m.58666
8-9 p.m.40444
9-10 p.m.15252

The grid tells the manager to start two servers at 4:30, add two at 5:45 and two more at 6:45, cut two at 8:30 and keep two to close, rather than bringing six in at 5. The kitchen applies the same logic using tickets per cook hour.

Measure forecast accuracy

Every Monday, compare forecast covers with actual covers for each meal period and each peak hour. Record the difference as a percentage. Over a few weeks you will see whether you are consistently high or low, and on which days. Most restaurants find that the peak hour on weekends is the hardest to forecast and the most expensive to get wrong.

  • Track error by meal period and by peak hour, not only by day.
  • Write down the reason for any large miss: weather, a private party, a competitor's opening, a road closure.
  • Adjust the event and weather notes so the next forecast learns from the miss.
  • Compare scheduled hours with actual hours too; a good forecast is wasted if managers do not cut on time.

A simple forecast checked every week beats a sophisticated one nobody looks at. Start with a spreadsheet and move to software once the routine is a habit.

Forecasting takeout, delivery and private events

Off-premises orders need their own line in the forecast. They do not use seats, but they load the line, the packing station and often the phone. Forecast them by hour from the same inputs as dine-in: last year's matching weekday, the recent trend and any promotions you have planned. If third-party orders arrive in waves around the dinner peak, schedule a packer or expo for that window rather than asking a line cook to bag orders between tickets.

Private events and large parties should be forecast separately from regular covers, because they follow the event calendar rather than the weekday pattern. Add their labor as a block, with the setup, service and breakdown hours they really need, and keep them out of next year's baseline so a one-off wedding rehearsal does not inflate the forecast for an ordinary Saturday.

When software helps

Operators expect technology to help here: 47.6% expect technology, automation, data and AI to improve labor productivity[5]. Many scheduling and POS tools now produce hourly forecasts automatically from your sales history. They save time, but they still need your judgment on events, weather and reservation pace, and they still need the weekly accuracy check. Our comparison of restaurant scheduling software explains what to look for, and our guide to writing the schedule shows how to turn the forecast into shifts.

Common forecasting errors

  • Using last week instead of last year, which ignores seasonality.
  • Forecasting by date instead of by weekday.
  • Treating every position as variable, so the forecast cuts a host or dishwasher the restaurant cannot run without.
  • Ignoring reservation pace, which is often the best signal you have for a specific night.
  • Not separating private events from regular covers, which distorts next year's baseline.
  • Forgetting to forecast takeout and delivery orders, which load the kitchen and packing station even when the dining room is quiet.

Frequently asked questions

How do restaurants forecast labor?

Forecast covers by hour from last year's matching weekday, the recent trend, reservation pace and known events, then convert covers into people using labor standards for each station.

What is a labor standard in a restaurant?

The amount of work one person can do well in an hour at a station, such as covers per server hour or tickets per cook hour. Set it from your own best busy shifts.

Should I forecast sales or covers?

Covers or orders, because they translate directly into work. Use sales to check the labor cost against your target.

How accurate should a restaurant forecast be?

Track your own error each week by meal period and peak hour, and aim to reduce it over time. Large misses should always have a recorded reason.

References

  1. National Restaurant Association (2026). Restaurants remain resilient despite challenging business conditions (opens in a new tab) ↑

  2. National Restaurant Association (2026). Restaurants remain resilient despite challenging business conditions (opens in a new tab) ↑

  3. US Bureau of Labor Statistics (2026). Current Employment Statistics: average hourly earnings, food services and drinking places (CES7072200003) (opens in a new tab) ↑

  4. US Bureau of Labor Statistics (2026). Current Employment Statistics: food services and drinking places employment, not seasonally adjusted (CEU7072200001) (opens in a new tab) ↑

  5. National Restaurant Association (2026). Research Insight: The Future of Restaurant Technology (opens in a new tab) ↑

Book a demo

See Ava handle your calls.

Book a short demo. We’ll walk through your calls and workflow, and agree what Ava should take care of.

Book a demo