Staff scheduling that stops losing money: labour cost, the rota, and sales per labour hour

In most owner-operated service businesses, labour is the largest cost you can actually control, and it is controlled in one document: the rota. Most rotas are written from habit and fairness-by-memory. Here is how to write one from the numbers, without turning your team into a spreadsheet.

The two numbers that judge every rota

Revenue tells you nothing about the rota until you divide by it. Two ratios do the judging:

• Labour cost as a share of sales. Total wage cost (including your own hours at a real rate) divided by revenue, per week. It is the number that quietly decides whether a busy week was a good week.
• Sales per labour hour. Takings divided by scheduled staff hours, by daypart if you can. This is the sharper tool: it shows WHERE the rota loses money, not just that it does.

Neither needs software to start. Your till knows takings by hour; your rota knows hours by person. One spreadsheet afternoon gives you a baseline most competitors never compute. The point of the weekly KPI habit is exactly this kind of number.

Schedule to the demand pattern, not the week

Almost every service business has a demand shape that repeats: the cafe's 8am spike and 3pm trough, the salon's Saturday wall, the restaurant's Friday double-peak, the gym's before-work and after-work waves. The rota that loses money is the one that staffs the AVERAGE instead of the SHAPE.

The fix is mechanical, not clever:

• Pull takings by hour and day for the last eight typical weeks.
• Mark the peaks that need full strength and the troughs that need a skeleton.
• Staff the shape: split shifts where labour law and decency allow, short overlap at handover, one flexible mid shift instead of two fixed ones.
• Re-check monthly, because the shape drifts with seasons and you will not feel the drift from behind the counter.

The quiet leaks inside a reasonable-looking rota

• The owner's invisible hours. If you cover every gap yourself, the rota "works" while your unpaid overtime subsidises it. Cost your own hours at a manager's wage, or every number you compute is fiction. This is owner dependence wearing an apron.
• Overtime creep. Overtime that happens every week is not overtime, it is an under-planned rota paying premium rates for predictable hours.
• Full strength during setup and close. Opening and closing rarely need the same crew as service. An hour of overstaffing at both ends of every day is a full shift a week.
• Fairness by memory. When shift allocation lives in your head, it drifts toward whoever complains best, and the resentment that follows costs you people, which is far more expensive than the hours. Publish the rota early and keep the rules visible.

What good looks like

A rota built this way usually finds meaningful savings without cutting a single person's total hours, because the waste was in the PLACEMENT of hours, not the amount. And the same numbers that trim cost also spot the opposite problem: a peak you consistently understaff is revenue you are turning away, which no cost report will ever show you.

Two cautions to keep it honest. First, do not chase the ratio into the ground: a labour percentage that looks heroic usually means service is quietly degrading, and you will pay for it in repeat customers before the quarter is out. Second, scheduling to demand only works if people can rely on their hours; volatility you create for the business becomes volatility your staff carry, and they will leave over it, which is the subject of hiring and keeping good people.

Run your own numbers against the break-even calculator to see what an hour of overstaffing actually costs your specific business per year. It is usually more than the raise your best employee is hoping for.