Restaurant Labor Cost Percentage: Formula and Fixes

Start with a comparable number

The arithmetic is simple. The definition is where useful comparisons are won or lost.

```text Labour cost percentage = (total labour cost for the period / sales for the same period) × 100 ```

Set the period first. A Monday-to-Sunday review can work if sales, labour cost, and hours all use that same Monday-to-Sunday window. If payroll is recorded on a different cycle, note how you have assigned it instead of comparing a partial payroll run with a full sales period. Keep the cut-off rules stable from one review to the next.

Then write down the numerator. Depending on the question you are asking, it may include hourly wages, salaried managers, overtime, employer payroll costs, benefits, bonuses, or other labour-related charges. There is no useful trend if one week includes payroll costs and the next week quietly excludes them. For diagnosis, it is often helpful to show a total labour figure and separate views for hourly, management, overtime, and payroll-related costs rather than hiding every category in one number.

Choose the denominator just as deliberately. You can work from gross or net sales, but label the basis and use it consistently. If discounts, refunds, comps, taxes, or other deductions are treated differently between periods, record that change beside the result. A percentage can move because the denominator changed, not because the operation became more or less productive.

A second measure makes the percentage more actionable:

```text Sales per labour hour = sales for the period / total labour hours for the period ```

Use hours worked when that is the measure you can verify. If you use scheduled hours for a planning view, label it as scheduled hours and do not mix it with worked hours in the same trend. Calculate both measures for the whole period and, where the data allows, for each daypart. A weekly percentage can conceal a quiet lunch, a busy dinner, a long close, or prep hours that do not belong to the same demand pattern.

A four-step labour triage framework

Use this sequence before changing a rota or removing a task. It separates a data problem from an operating problem and an operating problem from a demand problem.

1. Period: lock the accounting view.

Record the start date, end date, sales basis, labour scope, hours basis, and any unusual event. Note closures, holidays, private bookings, promotions, major menu changes, or missing clock records. These notes do not explain the number by themselves; they tell you whether the period is comparable with the one you are about to use as a reference.

2. Pattern: split the total before choosing a cause.

Break the result into dayparts such as prep, lunch, afternoon, dinner, and close. If useful, separate front of house, kitchen, management, overtime, and other defined labour groups. Look for direction rather than a magic threshold:

These are investigation prompts, not diagnoses. Confirm the cause in the underlying records before acting.

3. Service: set guardrails before the test.

Choose a small set of checks that describe the service your restaurant is meant to deliver. Depending on the operation, record ticket-time samples, order accuracy, guest complaints or comps, cleanliness and close completion, staff breaks, stockouts, or sales attachments. Use the measures you can collect consistently. Write down the current observation before the change so that a cheaper rota cannot look successful simply because the quality cost was left unrecorded.

4. Test: make one change that can be reviewed.

Choose one intervention, one owner, one primary metric, a review date, and a stop or reversal condition. A schedule adjustment, prep change, menu simplification, pricing change, and sales action can each affect the others. Changing all of them at once makes the result difficult to interpret. Start with the smallest change that addresses the pattern, review a comparable period, and record what happened even if you decide to restore the previous approach.

• A higher percentage with lower sales per labour hour suggests that you should investigate the relationship between hours and demand in the affected period.
• A higher percentage with stable sales per labour hour may point you towards labour rates, the scope of the numerator, sales mix, or a cost posted in that period.
• A stable percentage with worsening service checks can indicate that the operation is carrying too little coverage or the wrong skill mix for the service promise.
• A lower percentage with more errors, slower tickets, unfinished cleaning, missed breaks, or weaker selling is not automatically a win. The lower cost may be transferring work or lost sales elsewhere.

Example (illustrative)

The following figures are made up solely to show the arithmetic. They are not a benchmark, customer record, test result, or recommendation.

Assume one Monday-to-Sunday period using net sales, with the same labour scope and cut-off throughout:

| View | Sales | Labour cost | Hours worked | Labour cost percentage | Sales per labour hour | |---|---:|---:|---:|---:|---:| | Whole week | 60,000 currency units | 18,000 currency units | 1,200 | 30.0% | 50 currency units |

The calculations are:

Now split that same illustrative week:

| Daypart | Sales | Labour cost | Hours worked | Labour cost percentage | Sales per labour hour | |---|---:|---:|---:|---:|---:| | Lunch | 18,000 | 6,300 | 450 | 35.0% | 40 currency units | | Dinner | 42,000 | 11,700 | 750 | 27.9% | 56 currency units | | Total | 60,000 | 18,000 | 1,200 | 30.0% | 50 currency units |

The whole-week result does not tell the manager to remove a fixed number of people. In this illustration, lunch is the first place to investigate because its labour share is higher and its sales per labour hour is lower than dinner's. The next checks would be lunch demand against the rota, prep hours that serve lunch, actual clock records, and the service guardrails. If lunch service is already strained, the answer may be a sales, prep, or role-design investigation rather than a coverage cut.

• Labour cost percentage = 18,000 ÷ 60,000 × 100 = **30.0%**.
• Sales per labour hour = 60,000 ÷ 1,200 = **50 currency units per hour**.

Weekly labour-cost triage worksheet

Copy this structure into the weekly operating review. Rename the dayparts to match the restaurant; do not add rows simply to create more detail than the records can support.

Period and definitions.

| Field | Entry | |---|---| | Period start and end | ______________________________ | | Sales basis: gross or net | ______________________________ | | Labour cost included | ______________________________ | | Hours basis: worked or scheduled | ______________________________ | | Comparable-period notes | ______________________________ | | Missing, estimated, or late records | ______________________________ |

Daypart calculation.

| Daypart | Sales | Labour cost | Hours | Labour % | Sales per labour hour | Service observations | |---|---:|---:|---:|---:|---:|---| | Prep / open | ______ | ______ | ______ | ______ | ______ | __________________ | | Lunch | ______ | ______ | ______ | ______ | ______ | __________________ | | Afternoon | ______ | ______ | ______ | ______ | ______ | __________________ | | Dinner | ______ | ______ | ______ | ______ | ______ | __________________ | | Close / other | ______ | ______ | ______ | ______ | ______ | __________________ | | Total | ______ | ______ | ______ | ______ | ______ | __________________ |

For each row, calculate `labour % = labour cost ÷ sales × 100` and `sales per labour hour = sales ÷ hours`. If a daypart has no sales or no recorded hours, mark the result as unavailable instead of forcing a zero or an extreme percentage.

One-test record.

| Decision field | Entry | |---|---| | Observed pattern | ______________________________________________ | | Cause to verify | ______________________________________________ | | One change to test | ______________________________________________ | | Owner and date assigned | ______________________________________________ | | Primary metric | ______________________________________________ | | Service guardrails | ______________________________________________ | | Review date and comparable period | ______________________________________________ | | Stop or reversal condition | ______________________________________________ | | Decision after review | Keep / revise / reverse: ____________________ |

Decision matrix: choose the first fix

Use the matrix as a prompt, not as an automatic instruction. The owner, metric, review date, and stop condition should be filled in before the change starts.

| Observed pattern | Investigate first | Controlled first test | Owner, metric, and review | Stop or reverse if | |---|---|---|---|---| | Labour percentage is high in one daypart and sales per labour hour is lower there than in other dayparts | Forecast accuracy, actual clock records, prep demand, and coverage by role | Stagger a start, finish, or overlap for that daypart rather than reducing the whole week | Schedule owner; actual hours, sales per labour hour, and service checks; next comparable daypart review | Tickets, accuracy, breaks, cleanliness, or sales deteriorate; restore the prior coverage and investigate a different cause | | Labour percentage is high across the week while sales are soft | Sales pattern, menu mix, pricing, promotions, and whether the cost scope changed | Test one defined sales or menu action for a stated period, with labour held as comparable as possible | GM or commercial owner; sales, contribution measure used by the operation, labour percentage, and service checks; next comparable period | Service or demand signals worsen, or the action cannot be evaluated from clean records | | Hours or overtime rise while sales are broadly unchanged | Early clock-ins, late clock-outs, handoffs, approval rules, call-outs, and close routines | Change one handoff or approval routine and adjust the affected shift boundary | Rota or payroll owner; worked hours, overtime, close completion, and service checks; next comparable period | Work moves off the clock, required tasks are missed, or service deteriorates; reverse and redesign the handoff | | Prep hours are high and records show overproduction, spoilage, or repeated work | Prep list, forecast, par logic, shelf life, actual usage, and task ownership | Change the batch timing or quantity for one prep item or station | Kitchen lead; prep hours, discarded quantity, stockouts, and quality checks; next comparable prep cycle | Stockouts, quality problems, unsafe holding, or service delays appear; restore the previous par while checking the data | | Labour percentage is lower but service checks are worse | Coverage by role, skill mix, breaks, ticket flow, and work transferred to managers or the next shift | Add or reposition coverage only in the constrained period, or simplify one process before cutting further | GM; service checks, sales, worked hours, and labour percentage; next comparable period | Added coverage is not matched to a defined constraint, or the service problem persists; stop the change and reassess the process | | Labour percentage is stable but the wider margin is under pressure | Food cost, contribution by menu item, sales mix, fixed costs, and period accounting | Do not make a labour change first; run a separate food, menu, or sales review with its own record | Owner or finance lead; the relevant margin measure and labour percentage; next weekly review | The labour test is being used to explain a non-labour problem; cancel it and return to the wider diagnosis |

A test is complete only when the review record says what changed, what the data showed, what service looked like, and whether the next action is keep, revise, or reverse. “The number improved” is not enough if the period, sales basis, or labour scope changed at the same time.

Implementation steps for a weekly review

1. **Choose the cut-off.** Pick a recurring review time and use the same sales, payroll, clock, and inventory cut-offs as far as your records allow. 2. **Write the definitions at the top of the sheet.** State gross or net sales, included labour categories, and worked or scheduled hours. Treat any estimate as an estimate. 3. **Calculate the whole-period measures.** Record labour percentage and sales per labour hour before looking for a cause. 4. **Split by daypart and role where the data is reliable.** Compare like with like. Note a holiday, event, closure, menu change, or missing record instead of pretending it is a normal period. 5. **Check the service floor.** Select a few observable checks—such as ticket-time samples, accuracy, breaks, close completion, complaints, or stockouts—and record them beside the financial measures. 6. **Choose one test.** The first test might concern the rota, overtime control, prep architecture, menu complexity, pricing, or sales. Make the change narrow enough that someone can own it and the result can be read. 7. **Set the review and reversal rules in advance.** Use the next comparable period or a clearly stated date. If a service guardrail fails, stop or restore the previous approach while you investigate. 8. **Keep the learning.** A failed test is still useful when the definitions were stable and the reversal condition was respected. Record the cause that was ruled out and choose the next question rather than repeating the same change.

For a broader review of labour and food together, the [RestaurantMargin guides and plans library](https://restaurantmargin.com/books) includes a Prime Cost Recovery Plan. Use this worksheet to arrive with a defined period, a visible pattern, and one operating question rather than a request for a blanket reduction.

Limitations

• There is no single labour-cost percentage that fits every restaurant. Service model, opening hours, menu complexity, production method, management structure, local employment costs, and sales volume change what a sensible operating decision looks like.
• A percentage is sensitive to the sales denominator. Gross and net sales can produce different readings, and discounts, refunds, comps, taxes, and missing postings can change the comparison if the treatment is not documented.
• The numerator also matters. Excluding salaried management, overtime, payroll-related costs, or benefits may be useful for one diagnostic view but misleading for another. Show the scope rather than labelling a partial view as the full labour burden.
• Sales per labour hour is a productivity proxy, not a measure of hospitality, safety, training, food quality, or team sustainability. A higher value can accompany a busier period without proving that the underlying process improved.
• A single week can be distorted by weather, events, closures, unusual bookings, illness, menu changes, or recording errors. Do not turn an unusual period into a permanent staffing rule without checking comparable records.
• This worksheet does not forecast profit or cash, determine legal payroll compliance, or replace local accounting and employment advice. It is a decision record for operating experiments, not a complete financial statement.
• Any reduction in recorded labour cost is not automatically an improvement. Work can be shifted into unpaid time, management time, slower service, waste, lost sales, or future rework. Keep those possible trade-offs visible in the review.