Restaurant Labor Forecasting: How to Schedule Staff from Expected Sales

A restaurant schedule should not begin with the question, “Who is available next week?”

It should begin with a more operational question:

How much business do we expect, when will it arrive, and how many labor hours will it take to serve it properly?

Labor forecasting connects expected demand to staffing. Done well, it helps a restaurant avoid two expensive mistakes: paying for too many hours when business is quiet, and running too few people when demand is high enough to damage service, ticket times, quality and sales.

This guide shows how to build a practical labor forecast using sales history, dayparts, labor productivity and manager judgment.

Labor forecasting is not the same as writing the schedule

The forecast comes first.

The schedule comes second.

A labor forecast estimates how much staffing the restaurant is likely to need. The schedule assigns actual employees to those required hours.

That distinction matters because a schedule can look complete on paper while still being wrong for the business level.

Start with expected demand

The labor plan should be built from a realistic demand forecast.

Depending on the concept, useful demand measures may include:

• net sales,
• guest count,
• transactions,
• covers,
• orders,
• delivery volume,
• production units.

Sales is often the easiest starting point, but some departments may be better forecast from transactions or covers.

Use comparable historical days

Do not forecast Friday from a random Tuesday.

Start with comparable periods:

• same weekday in recent weeks,
• same daypart,
• similar season,
• similar promotion status,
• similar local-event conditions.

If recent Fridays produced $8,100, $8,450, $7,900 and $8,300 in sales, a forecast around that range may be more useful than simply copying the most recent day.

Adjust history for what you already know

Historical averages are a starting point, not the final answer.

Adjust for known factors such as:

• large bookings,
• local events,
• holidays,
• school calendars,
• marketing campaigns,
• menu promotions,
• expected weather where it materially affects traffic,
• delivery-platform campaigns,
• temporary opening-hour changes.

The point is not to predict perfectly. The point is to be deliberately less wrong.

Forecast by daypart, not just by day

A restaurant may forecast $10,000 in daily sales and still schedule badly if it does not understand when those sales will occur.

Break the forecast into useful operating periods:

• opening and prep,
• breakfast,
• lunch,
• afternoon lull,
• dinner,
• late night,
• closing.

Different concepts will need different dayparts.

Example: the same sales with different labor needs

Imagine two days both forecast at $8,000.

Day A produces a steady flow from noon to 9:00 p.m.

Day B produces 60% of sales between 7:00 and 9:00 p.m.

The same daily revenue can require very different staffing patterns.

Labor should follow the shape of demand, not only the total.

Choose a labor productivity measure

A simple way to translate sales into labor is sales per labor hour.

The formula is:

Sales per labor hour = Net sales ÷ Total labor hours

Example:

• Net sales: $7,200
• Total labor hours: 240

$7,200 ÷ 240 = $30 sales per labor hour

Our Restaurant Labor Productivity guide explains this metric in more detail.

Convert the forecast into labor hours

If the restaurant has a realistic productivity target, reverse the formula:

Required labor hours = Forecast sales ÷ Target sales per labor hour

Illustrative example:

• Forecast sales: $9,000
• Target productivity: $30 sales per labor hour

$9,000 ÷ $30 = 300 labor hours

This does not mean 300 hours should be scheduled mechanically. It gives the manager a planning reference.

Do not use one productivity target for every hour of the day

Opening, prep and closing work often happens before or after meaningful sales occur.

If the restaurant applies one sales-per-labor-hour target to every period, it may under-resource essential non-selling work.

Separate labor into useful groups:

• fixed opening labor,
• variable service labor,
• production labor,
• closing labor.

This creates a more realistic model.

Separate fixed labor from variable labor

Some labor does not fall much when sales fall.

A restaurant may still need:

• one opening manager,
• one minimum kitchen setup team,
• one cashier or host,
• one closing manager,
• basic cleaning coverage.

That is effectively fixed labor for the operating period.

Other labor can flex more directly with demand.

Build a base crew before adding volume-driven labor

A practical staffing model can start with the minimum safe and workable crew.

Then add positions as forecast volume rises.

Illustrative example:

Forecast dinner sales	Kitchen staffing example
Up to $2,000	2 line cooks + 1 prep/support
$2,000–$3,500	3 line cooks + 1 prep/support
$3,500–$5,000	4 line cooks + 1 support
Above $5,000	4 line cooks + expediter + support

This table is illustrative only. The staffing thresholds must be built from your own kitchen layout, menu complexity, equipment and service model.

Build position-level staffing rules

Total labor hours are useful, but managers also need to know where those hours belong.

Create staffing rules for:

• servers,
• cashiers,
• hosts,
• line cooks,
• prep cooks,
• dishwashers,
• bar staff,
• delivery packing,
• supervisors.

A front-of-house role may be driven by covers. A prep role may be driven by production quantity. A dishwasher may be driven more by transaction volume and operating hours.

Use the best demand driver for each role

Examples:

Role	Possible planning driver
Server	Covers or tables
Cashier	Transactions per hour
Line cook	Sales or entrées per hour
Prep cook	Production plan
Dishwasher	Transaction / cover volume
Delivery packer	Delivery orders per hour

A restaurant does not need a complex algorithm. It needs a staffing rule that reflects how work is actually created.

Use hourly demand where the peak matters

If your POS provides hourly sales, use it.

An hourly view can reveal that:

• sales rise sharply at 12:30,
• delivery peaks before dine-in,
• bar demand continues after kitchen demand falls,
• closing workload starts before sales end.

This helps managers stagger start and end times instead of scheduling everyone for identical shifts.

Staggered shifts are often more efficient than full-block shifts

Suppose lunch begins building at 11:30, peaks from 12:30 to 2:00, and falls sharply after 2:30.

Instead of scheduling five people from 10:00 to 6:00, the restaurant might use:

• two openers,
• two employees starting before the rush,
• one short peak shift,
• fewer people through the afternoon lull.

The exact design depends on local employment rules and the restaurant’s staffing model, but the operating principle is simple: match coverage to workload.

Do not solve every forecast with split shifts

A schedule that is financially efficient but impossible for employees to live with will eventually create turnover, call-outs and hiring costs.

Labor forecasting should balance:

• business need,
• employee availability,
• shift quality,
• rest periods,
• local labor requirements,
• retention.

The cheapest theoretical schedule is not always the lowest-cost operating schedule.

Use forecast labor cost as a second check

After estimating hours, translate them into labor cost.

A simple formula is:

Forecast labor cost = Scheduled hours × Expected hourly cost

For a more accurate model, include the labor-cost components your restaurant actually manages, such as payroll taxes or other employer costs where applicable.

Calculate forecast labor percentage

One management check is:

Labor cost % = Forecast labor cost ÷ Forecast sales × 100

Illustrative example:

• Forecast labor cost: $2,250
• Forecast sales: $9,000

$2,250 ÷ $9,000 × 100 = 25%

This percentage should not be used alone. A schedule can hit the percentage and still be badly deployed by hour or position.

See our Restaurant Labor Cost Control guide for the broader financial view.

Forecast labor by department

One total labor percentage can hide a problem.

Break the plan into areas such as:

• front of house,
• back of house,
• management,
• delivery / dispatch,
• production or bakery where relevant.

If total labor looks acceptable but kitchen labor is consistently too high, the restaurant needs a different solution than if front-of-house coverage is the issue.

Account for skill mix

Ten labor hours are not always equivalent to ten other labor hours.

A shift with inexperienced employees may need more supervision and coverage than a shift with a highly trained team.

Consider:

• who can open,
• who can close,
• who can run each station,
• who can train,
• who can manage cash,
• who can supervise food safety or other required responsibilities.

A schedule is not valid simply because the headcount is correct.

Account for training hours separately

Training shifts should be visible in the labor plan.

If a new hire shadows an experienced employee, those hours may not create normal productive capacity yet.

Mark them as training so management can distinguish:

• productive labor,
• training investment,
• temporary overstaffing by design.

Our Restaurant Staff Training Checklist provides a structured onboarding framework.

Forecast prep labor from the production plan

Prep labor often needs a different model from service labor.

Use the production plan to estimate:

• batch quantities,
• prep minutes per batch,
• equipment constraints,
• skill requirements.

For example, if tomorrow’s demand requires 10 batches and each batch takes 18 minutes of active labor:

10 × 18 minutes = 180 minutes = 3 labor hours

Add setup, sanitation and realistic handling time rather than assuming pure production time is the whole workload.

Forecast closing labor too

Managers sometimes cut too deeply after the final rush and leave the closing team with an unrealistic workload.

Estimate closing work:

• cleaning,
• restocking,
• cash close,
• waste handling,
• equipment shutdown,
• handover notes.

The correct end time is when the work is reasonably finished, not simply when sales stop.

Build a scheduling matrix

A useful matrix connects demand bands to staffing levels.

Forecast band	FOH	BOH	Manager	Support
Low	Minimum coverage	Base line crew	1	Shared
Normal	Standard coverage	Full core stations	1	As needed
High	Additional sections	Extra line position	1–2	Dedicated
Peak / event	Peak plan	Peak line + expeditor	2	Dedicated

The matrix above is illustrative. Build your own from observed operating performance.

Use actual results to calibrate the model

A labor forecast improves only if the restaurant compares forecast with reality.

After each period, review:

• forecast sales,
• actual sales,
• scheduled hours,
• actual hours,
• sales per labor hour,
• service results,
• manager notes.

If the forecast missed, ask whether the problem was demand prediction or staffing productivity.

Measure forecast accuracy

A simple sales forecast variance is:

Forecast variance = Actual sales − Forecast sales

And percentage variance:

Forecast variance % = (Actual sales − Forecast sales) ÷ Forecast sales × 100

Example:

• Forecast: $8,000
• Actual: $8,600

($8,600 − $8,000) ÷ $8,000 × 100 = +7.5%

Repeated large misses mean the forecasting method needs work.

Compare scheduled hours with actual hours

The schedule may plan 260 hours, but employees may actually clock 278.

That 18-hour difference deserves explanation.

Possible causes include:

• employees staying late,
• early clock-ins,
• unexpected demand,
• poor shift handover,
• slow closing,
• manager approval not followed.

The restaurant should manage actual labor, not only scheduled labor.

Do not cut labor blindly when sales miss forecast

If sales are below forecast at 5:00 p.m., the correct response is not always to immediately send employees home.

Check:

• reservation book,
• delivery queue,
• expected peak,
• weather or event conditions,
• prep workload,
• closing requirements.

Use live conditions, not panic.

Use trigger points for in-shift labor decisions

Managers can define simple triggers.

Examples:

• If sales are 15% below forecast by 2:00 p.m. and no large bookings remain, review one early-out.
• If delivery volume exceeds the normal band for 30 minutes, add packing support.
• If ticket times exceed the operating threshold, do not cut kitchen labor until service stabilizes.

These examples are illustrative. The value comes from having a consistent decision process.

Protect the positions that prevent bottlenecks

Not every position should be cut in the same order.

A dishwasher, expediter or host may appear indirect, but removing that position at the wrong time can slow the entire operation.

Labor decisions should consider system bottlenecks, not just wage rate.

Use service metrics as guardrails

A labor model that improves percentage while damaging operations is not working.

Monitor guardrails such as:

• ticket time,
• guest complaints,
• order accuracy,
• delivery delays,
• table turn time where relevant,
• employee overtime,
• manager intervention frequency.

If labor productivity improves while these measures deteriorate, the restaurant may be understaffed.

Look for consistent overstaffing patterns

One quiet shift does not prove overstaffing.

Repeated patterns do.

Examples:

• too many employees scheduled before demand begins,
• too many full-length shifts through the afternoon lull,
• closing team stays scheduled long after work is complete,
• the same low-volume weekday receives weekend staffing.

These are structural schedule problems.

Look for consistent understaffing patterns

Understaffing often appears as:

• ticket-time spikes,
• missed breaks,
• manager constantly jumping onto stations,
• closing tasks rolling into the next day,
• avoidable overtime,
• increased mistakes,
• employee burnout and turnover.

If these appear consistently during the same periods, the schedule may need more capacity even if the labor percentage looks attractive.

Review schedule efficiency by daypart

Instead of reviewing only the whole day, calculate productivity by operating period.

Example:

Daypart	Sales	Labor hours	Sales/labor hour
Lunch	$3,600	110	$32.73
Afternoon	$900	55	$16.36
Dinner	$5,500	155	$35.48

The figures are illustrative.

The afternoon period may deserve investigation even though total daily productivity looks acceptable.

Connect the schedule to daily management

Forecasting is not complete when the schedule is posted.

Managers should review actual demand during the shift and record meaningful exceptions.

Our Restaurant Manager Logbook guide provides a practical handover system for unusual staffing and operating conditions.

Use the pre-shift briefing to explain staffing changes

If the forecast requires a different deployment, communicate it before service.

For example:

• one server covers a larger section early,
• an extra cook begins before the peak,
• a manager supports delivery packing during the rush.

The Restaurant Pre-Shift Meeting guide shows how to communicate these priorities in a short briefing.

Build a weekly forecasting routine

A practical weekly process is:

1. Pull recent comparable sales and transaction history.
2. Review bookings, events and promotions.
3. Forecast sales by day and daypart.
4. Translate demand into labor-hour requirements.
5. Apply minimum staffing and skill requirements.
6. Build position-level coverage.
7. Check forecast labor cost and percentage.
8. Assign actual employees.
9. Review overtime and availability.
10. Publish the schedule.
11. During the week, compare actual demand and adjust responsibly.

Our Restaurant Staff Scheduling guide covers the schedule-building side after the forecast is complete.

Build a simple forecast sheet

A basic worksheet can include:

Field	Example
Date	Friday
Forecast sales	$9,000
Forecast covers	420
Target sales/labor hour	$30
Initial labor hours	300
Training hours	8
Scheduled productive hours	292
Forecast labor cost	$2,250
Forecast labor %	25%

All figures are illustrative.

Use ranges instead of false precision

Restaurant demand is uncertain.

Instead of pretending Friday sales will be exactly $9,000, management can use:

• low case: $8,200,
• base case: $9,000,
• high case: $10,200.

Then identify which staffing decisions change across those scenarios.

This makes the schedule more resilient.

Know which shifts can flex

Some shifts have natural flexibility.

Examples:

• a support employee can start 30 minutes later if demand is weak,
• a trained cross-functional employee can move between packing and counter service,
• a manager can delay administrative work to support a rush.

Build flexibility intentionally rather than improvising every day.

Cross-training increases scheduling flexibility

Cross-trained employees give the restaurant more options when demand shifts unexpectedly.

But cross-training should follow competency in the primary role, not replace it.

Use training records so managers know who can safely and consistently cover each station.

Do not chase labor percentage at the expense of revenue

A restaurant can reduce labor percentage by cutting coverage, but if service slows and guests leave, the business may save wages while losing sales.

Labor should be managed as a productivity system, not simply minimized.

The goal is:

the right people, in the right positions, at the right times, for the business level expected.

Use actual results to keep improving

At the end of each week, compare:

• forecast versus actual sales,
• scheduled versus actual hours,
• forecast versus actual labor cost,
• sales per labor hour,
• service exceptions,
• overtime,
• manager comments.

Then update the staffing rules.

A labor model should become more accurate as the restaurant learns from its own data.

The forecast is a management tool, not a promise

No forecast will predict every rush, cancellation, weather change or call-out.

The purpose is not perfect prediction.

The purpose is to create a disciplined starting point for the schedule and a clear framework for in-shift decisions.

When sales history, daypart demand, labor productivity, skill mix and manager judgment are combined, labor scheduling becomes much more deliberate—and much less dependent on guesswork.

Measure forecast error separately from labor execution

A bad labor result can come from a weak demand forecast or from poor staffing execution after the forecast was reasonable. Track both.

Absolute forecast error % = |Actual sales − Forecast sales| ÷ Forecast sales × 100

Then compare scheduled hours with actual hours. This helps management decide whether to improve the forecasting model, the schedule, or in-shift labor control.

Build a weekly forecast-accuracy dashboard

Measure	Why it matters
Sales forecast error %	Quality of demand prediction
Scheduled vs actual hours	Execution against the labor plan
Sales per labor hour	Productivity outcome
Service guardrails	Whether productivity hurt the guest experience

Review the dashboard by daypart where enough data exists to make the comparison useful.

Frequently asked questions

How much history should a restaurant use for labor forecasting?

Use enough comparable history to reduce random noise, then adjust for current conditions. Recent same-weekday and same-daypart data is usually more useful than an old annual average.

Should labor forecasts use sales or transactions?

Use the demand driver that best explains workload for each role. Sales may work for total planning while transactions, covers, delivery orders or production quantities may work better for specific positions.

What should happen when the forecast is wrong during the shift?

Use pre-defined operating triggers and manager judgment. Protect critical stations and service before changing labor simply to force the percentage back to plan.

Operational note: The formulas, thresholds and examples in this guide are illustrative. Scheduling, breaks, overtime, predictive-scheduling requirements and other employment rules vary by jurisdiction and should be reflected in your local labor practices and professional advice where appropriate.