Rostering a cafe to the day it actually has

A cafe roster copied forward each week reproduces last week's assumptions, including the ones that were wrong.

The result puts too many people on at eleven and too few at eight, on the same day.

Read the day in hours, not in totals

A daily takings figure hides the shape of the trade. Two days taking the same amount can need entirely different rosters.

Sales by hour, averaged over several weeks, show where the peak really sits. Most cafes find it earlier and sharper than they assumed.

Staff the peak, then trim the edges

A roster built from opening to close at a steady level pays for quiet hours and fails at the busy one. Building outward from the peak inverts that.

Staggered starts are the tool. Two people opening, four through the rush, back to two by eleven costs less and serves better.

Labour percentage is a result, not a target

Chasing a percentage encourages cutting the peak, which costs sales and makes the percentage worse the following week.

Our guide to labour cost on a roster covers reading it properly, as an outcome of decisions rather than a lever to pull.

Publish early enough to matter

A roster published on Sunday for a Monday start produces swaps, absences and resentment. Staff plan their lives around it.

Two weeks of notice costs the business nothing and buys reliability. A late roster is the most common reason a good casual takes another job.

Skill mix matters as much as headcount

Four people who cannot make coffee do not cover a morning. The roster needs the right capability at each hour, not simply bodies.

Marking who can run the machine, the pass and the floor turns rostering into a plan. It also shows where training is overdue.

Weather and the local calendar move the day

A rainy Tuesday in a strip of offices behaves differently from a sunny one. School terms, local events and public holidays move trade in predictable directions.

Noting the reason beside an unusual day stops the average from absorbing it. Otherwise one street festival distorts a Saturday forecast for months.

Compare the roster with the hours actually worked

Rostered hours and paid hours drift apart through late finishes, early starts and forgotten breaks. The gap is real money.

Our guide to checking a pay period covers catching it before anybody gets paid, which is the only cheap moment.

Breaks belong in the roster, not in the gaps

A break nobody rostered gets taken at the worst moment or not at all, and both cost the business.

Placing them deliberately, around the peak rather than through it, keeps the floor covered and the team upright.

One person should own the roster

A roster edited by several managers drifts into double bookings and gaps nobody owns.

One author, with everybody else requesting rather than editing, keeps it coherent and makes the changes traceable.

Look back at the week you actually had

A roster built and never reviewed repeats its errors. Ten minutes comparing planned cover against how the day felt is enough.

Most cafes find the same two hours consistently wrong, in the same direction, week after week.

Those two hours are usually the open and the last hour before close, where habit outlives the trade that justified it.

What to change first

• Sales by hour, averaged over weeks, replace a daily total
• The roster builds outward from the peak with staggered starts
• Labour percentage reads as a result rather than a target
• Two weeks of notice becomes the standard, not a courtesy
• Capability sits beside each name, so skill mix is visible
• Rostered hours meet paid hours before every pay run