Labor cost percentage: how to calculate it, and what counts as normal

The formula

Labor cost percentage is total labor cost divided by total sales for the same period, expressed as a percentage. If you paid $14,000 in labor against $48,000 in sales, you are at 29.2%.

The arithmetic is the easy half. Almost every disagreement about this number comes from two questions: what went into the top, and what went into the bottom.

What belongs in the numerator

Wages alone understate what staffing actually costs you. A fully loaded labor figure includes everything you pay because someone worked:

• Hourly wages and salaries, including overtime at the rate you actually paid
• Employer payroll taxes
• Benefits, insurance and any employer retirement contribution
• Paid time off taken during the period
• Manager and owner salaries, if those people work shifts

What belongs in the denominator

Use net sales — gross sales less discounts, comps and refunds — and be consistent about it. Whether you include sales tax changes the answer by several points, so pick one convention and never mix it across weeks or locations.

The most common way to get a misleading number is to compare a gross-sales week against a net-sales week and conclude that labor moved. It did not; the denominator did.

What counts as normal

Ranges vary widely by service model, market and minimum wage, so treat these as orientation rather than targets. Your own trend matters far more than any benchmark:

• Counter-service cafés and coffee shops: roughly 25–35%
• Full-service restaurants: roughly 30–35%, higher where tipped wages are not offset
• Quick service: roughly 25–30%
• Small retail: usually well below food service, often 10–20%

Why weekly beats monthly

A monthly labor percentage averages away the thing you can act on. Overstaffing happens on particular shifts — a Thursday close scheduled two deep against a slow night, an opening pair where one person would do — and a month-long average hides it inside four weeks of noise.

A weekly number lines up with how schedules are actually written, which means a bad week is still close enough to remember and correct. Anchoring to complete Monday–Sunday weeks also keeps you from comparing a partial week against a full one, which is the other way this metric lies.

Where the number usually goes wrong

When labor percentage drifts up without anyone deciding to spend more, the cause is usually structural rather than dramatic:

• Scheduled hours are tracked instead of actual clocked hours, so early clock-ins and late clock-outs never show up
• Shift overlap at open and close is longer than the work requires
• Overtime accrues quietly at the end of the week rather than being redistributed earlier
• Sales dipped and the schedule did not move with it — the numerator held while the denominator fell

Reading it per location

With more than one site, the useful comparison is not each location against a benchmark but against each other. Two shops with the same menu and similar volume should land within a point or two; a persistent gap of five or more points is a scheduling difference, not a market difference.

With a single location, compare your own weeks, shifts and dayparts instead. The same logic applies — you are looking for the gap between a good week and a bad one, and for what the schedule did differently.