Labour cost

Why your restaurant’s labour cost percentage is wrong

Almost every multi-unit operator can tell you their labour percentage. Almost none can tell you what share of their hours it was calculated on. The gap between those two questions is where most restaurant reporting quietly falls apart.

The arithmetic everyone agrees on

Labour percentage is total labour cost divided by net sales. Spend $209,000 against $858,000 and you ran 24.4%. Nobody argues about the formula.

The argument nobody is having is about the numerator. Total labour cost means the total cost of total hours. If some of your hours have no cost attached, your numerator is short — and 24.4% is not your labour percentage. It is your labour percentage on the hours that happened to be counted.

How hours stop being counted

Your POS knows everyone who has clocked in. Your roster knows everyone you employ. These lists diverge for entirely mundane reasons:

• A starter is added at the till during a Friday rush and nobody circles back
• “Mike Sanchez” on the till is “Michael S” on the roster
• A leaver’s login is never retired and someone else uses it
• Terminal accounts, training logins and test users look exactly like people
• A transfer creates a second record instead of moving the first

Every one produces a person whose hours the till recorded and whose cost matched nothing. Hours worked. Wages paid. Absent from every report drawing labour from the roster.

What this looks like in real numbers

Their dashboard said labour was 24.4% of sales. With every hour costed at their average wage, the real figure was closer to 35%. That is not a rounding error — it is the difference between a business that works and one that does not, and every decision in between had been made on the first number.

Why the error always runs the same direction

Missing hours make labour look lower, never higher. That is what makes it dangerous: nobody investigates a good number. An operator whose labour lands at 24% against a 27% target does not open the report and start checking denominators. They move on.

Worse, the stores with the worst mapping look like the best performers and get held up as the example. I have watched a group award “most efficient location” to the store with the most unmapped staff. Twice.

Why it ruins store comparison, not just one number

One store with a coverage problem produces one wrong number. A group with uneven coverage produces a wrong ranking — and the ranking is what you act on.

If store A has 95% coverage and store B has 60%, B looks more efficient even if it is genuinely worse. You send the area manager to the wrong store, copy the wrong store’s practices, and consider closing the wrong site.

How to check yours in ten minutes

1. Pull total hours worked last month from your POS or timeclock — every clock-in, including people you do not recognise.
2. Pull the hours that appear in your labour cost report for the same period.
3. Divide the second by the first.

If the answer is not close to 100%, everything downstream is understated by roughly the inverse. At 68% coverage, multiply your labour percentage by about 1.47 to get near the truth. Do this before your next staffing decision.

The free labour blind spot auditor runs this arithmetic and shows what the missing cost is worth annually.

The fix is boring, one-off and retrospective

Mapping a POS employee to a roster record is a matching exercise. It takes an afternoon for most groups. And — the part operators do not expect — it applies backwards. Linking a person attaches their cost to every hour they have ever worked, not just from today. Your history corrects itself.

What good looks like afterwards

Coverage at or near 100%, and a system that reports coverage every time it reports labour. A labour percentage without a coverage percentage is not a measurement, it is a rumour.