Labour cost, and the rota that causes it.

Total labour divided by sales, same period. Total labour means gross wages plus employer PRSI or National Insurance plus holiday pay accrual plus any agency cost. Restaurants that use gross wages alone typically understate labour by a significant slice, which is exactly enough to turn a problem into an apparent non-problem.

UK hospitality labour commonly runs between 30% and 35% of revenue. Quick service and counter formats sit lower because the service model is lighter. Full service sits higher, and fine dining higher again, because the whole proposition is people. A number is only meaningful against your own format and your own last quarter.

This is the single most common finding when a restaurant maps labour against demand by hour. Owners schedule to a rush they remember rather than one they have measured, and the error is usually at the edges: staff on from five when the trade starts at six, or an extra pair of hands until close when the last hour does a fraction of the takings. Pull thirteen weeks of sales by hour and lay the rota over it. The gaps are normally obvious and normally at the start and end of shifts.

Labour percentage moves when sales move, which makes it a poor operational measure on a quiet week. Sales per labour hour is steadier and more actionable: it tells you what each scheduled hour produced. Track it by day part. You will usually find one shift carrying the week and one quietly losing money, and the percentage alone will never show you which.

There is a floor, and going under it costs more than it saves. Understaffing shows up as longer ticket times, worse reviews and staff turnover, and all three are more expensive than the wage you saved. Turnover in particular is brutal: recruiting and training a replacement costs weeks of productivity. Cut hours where demand genuinely is not there, and stop.