A cafe roster that does not quietly cost you a week's margin

Most owners learn what last week's cafe roster cost only when the pay run lands. By then the week is over and the money is spent.

No single shift feels like over-rostering. The cost builds a few paid hours at a time. It needs a price before publication, and a check against the hours worked afterwards.

Cost the cafe roster before you publish it

Three numbers belong on the roster itself, visible while the shifts are still moving:

1. Forecast sales for each day, from the same day in recent weeks, adjusted for anything you already know is coming.
2. The real cost of each rostered hour: the rate for that day and time, plus super, workers compensation and payroll tax if you pay it.
3. Rostered wages as a percentage of forecast sales, for each day and for the week.

Each day's forecast comes from that weekday across the last four weeks, never last week alone. One wet Saturday should not set next Saturday's roster.

Then track the actual percentage daily rather than waiting for the month's figures. Wednesday's roster can still correct a bad Tuesday. A monthly report cannot.

Compare against the right benchmark

Shiftly's guide to hospitality labour costs points out that the 28 to 32% benchmark most venues quote is a wages only figure. By its calculation, adding super, workers compensation and payroll tax makes the loaded figure about 14.5% higher.

The comparison has to match one week against a similar one. A roster costed with on-costs, measured against a wages only benchmark, looks worse than it is. A roster costed on base wages against a loaded benchmark looks better than it is, and that is the direction that loses money.

Where paid hours leak

Everyone starting at the same time

Staff start together because last week's roster had them start together. Stagger starts in half hour steps as sales build, using your point of sale's sales by hour.

The last hour

The final hour of trade usually has more people than work. Trim the close before you cut anyone from the rush.

Rounded clock-outs

Eight minutes added to a clock-out does not stand out on any single timesheet. Across 60 shifts a week, it is eight paid hours nobody rostered.

The extra person, just in case

One extra person for six hours on Saturday and six on Sunday is twelve hours a week. They only pay for themselves when the rush arrives. Those hours take that day's rate before anybody adds them.

Where the rush is a real risk rather than a habit, a short list helps. Staff who have said they will take extra hours at short notice. Calling one of them in costs the hours you need. Rostering the extra person in advance costs the hours you might need.

What that does to a week's profit

Consider a cafe with $25,000 of sales in a week and a 10% net margin, so $2,500 of profit. Assume a loaded cost of $40 an hour on weekdays and an average of $50 an hour for the weekend shifts. These are illustrations, not benchmarks.

• Eight rounded hours at $40: $320.
• Twelve just in case weekend hours at $50: $600.
• Two staff kept half an hour past need at close, six days, at $40: $240.
• Three staff started 45 minutes before the trade needs them, six days, at $40: $540.

That is $1,700 a week, or 68% of the $2,500 profit, from four habits. Across a year it is $88,400.

Check the roster against the hours worked

Before approving each pay run, put three columns side by side for every shift: rostered hours, clocked hours, and the difference. A shift that runs over every week needs a different roster, not another approval.

The rostered hours stay on file after the week ends. Without them the timesheet is the only record of the week's intended cost, and nothing exists to compare.