Guides🌐 GlobalJuly 27, 2026·8 min read

How to make a staff roster that doesn't blow your labour budget

Most rosters are copied from last week, which was copied from the week before, which someone wrote in 2023 based on a hunch. Meanwhile your sales data knows exactly when you're busy. Here's how to build a roster from demand — and keep labour under control without leaving the floor short.

W

Wemu Team

Product

ATO · STP PHASE 2Pay event submission — 14 Apr 2026Employees: 6GROSS PAY (disaggregated)Ordinary hours$8,420.00Overtime$612.50Paid leave (annual)$980.00Allowances (travel)$140.00Bonus / commission$450.00PAYG WITHHELD$1,842.30SUPER GUARANTEE (11.5%)$1,221.45Submitted to ATO · Ref 2026-EVT-0842All 6 employees · Phase 2 compliant · delivered 14 Apr 14:32 AEDT

Labour is usually the largest controllable cost in a service business, and the roster is where it's decided — days before the money is spent. Yet most rosters are built from memory and copied forward, which means they encode last year's trading patterns and every staffing mistake made since.

Start with demand, not availability

The order matters. Build the shape of the week from what your sales data says, then fill it with the people available — not the other way around.

  1. 1Pull sales by hour for the last 8–12 weeks from your POS, split by day of week
  2. 2Mark the peaks and the dead zones. Most venues discover at least one shift that's staffed out of tradition rather than trade
  3. 3Set a coverage rule for each band — for example, one staff member per $X of hourly revenue, or per Y covers, or per Z bookings
  4. 4Now lay in your people, matching your strongest staff to your busiest bands
  5. 5Check the total against your labour cost target before you publish, not after payroll runs

The overlap trick

Most understaffing complaints come from changeover, not headcount. Staggering starts by 30–60 minutes so shifts overlap through the peak usually fixes the "we're slammed at 12:30" problem without adding hours.

Know your labour cost percentage

Labour cost percentage is total wage cost (including on-costs like superannuation, payroll tax, and casual loading where they apply) divided by revenue for the same period.

Business typeTypical labour costNotes
Café25–35%Rises fast if you over-staff quiet mid-afternoons
Full-service restaurant30–40%Kitchen plus floor; prime cost target is 60–65% with food
Salon / spa35–50%Commission structures push this higher by design
Retail10–20%Lower ratios; the trap is under-staffing peak conversion hours
Fitness studio30–45%Instructor pay per class dominates
Directional. Track your own trend weekly — the direction is more useful than the benchmark.

Set the target before you build the roster, then check the projected cost as you go. Discovering you're four points over after payroll has run is a lesson you pay for twice.

The rules that keep a roster fair

  • Publish at least two weeks ahead — short notice is the single biggest driver of resentment and turnover in hourly teams
  • Rotate the undesirable shifts rather than assigning them to whoever complains least
  • Respect availability you've agreed to, and hold people to the availability they gave you
  • Set a swap policy in writing: who can swap with whom, and who approves it
  • Honour minimum rest between shifts and any award or contract rules that apply where you operate
  • Never roster to a number that only works if nobody calls in sick — build one flex person into peak days

Swaps, sickness, and the Saturday-morning phone tree

The roster you publish is not the roster you run. What separates a calm operation from a chaotic one is how changes are handled after publication.

The worst version is a group chat where a manager asks "can anyone cover?" and eight people ignore it. The better version is a system where a staff member posts the shift, everyone eligible sees it, the first to accept takes it, and the roster updates itself — with the manager approving rather than brokering. That's exactly what the Wemu Team app does: staff see their schedule, request swaps as open offers or direct trades, and accepted swaps update the roster immediately.

ATO · STP PHASE 2Pay event submission — 14 Apr 2026Employees: 6GROSS PAY (disaggregated)Ordinary hours$8,420.00Overtime$612.50Paid leave (annual)$980.00Allowances (travel)$140.00Bonus / commission$450.00PAYG WITHHELD$1,842.30SUPER GUARANTEE (11.5%)$1,221.45Submitted to ATO · Ref 2026-EVT-0842All 6 employees · Phase 2 compliant · delivered 14 Apr 14:32 AEDT
Staff see their shifts, clock in, and resolve swaps themselves — the roster updates without a phone tree.

Close the loop: scheduled vs actual

A roster is a plan. Clock-ins are what happened. The gap between them is where labour budgets quietly die — the 15-minute early starts, the shifts that ran 40 minutes long, the overtime nobody approved.

  • Compare scheduled hours against clocked hours every week, per person
  • Flag overtime as it accrues, not when payroll runs
  • Watch for consistent early clock-ins — usually a rostering problem (the shift genuinely needs to start earlier) rather than a discipline problem
  • Track labour as a percentage of same-day revenue so a quiet day's over-staffing is visible immediately

When scheduling, clock-ins, and sales live in one system, that comparison is a report rather than a reconciliation project — and the next roster gets built from what actually happened instead of what was planned.

Roster from real demand

Wemu builds schedules alongside your sales data, gives staff their own app for shifts and swaps, and compares scheduled hours against actual clock-ins with overtime flagged. 7-day free trial.

Start free trial

Frequently asked questions

How far in advance should I publish the roster?+
Two weeks is the practical minimum for hourly teams, and some jurisdictions legislate a notice period. Beyond compliance it's a retention issue — unpredictable schedules are one of the most cited reasons hospitality and retail staff quit.
What is a good labour cost percentage?+
It varies by model: cafés typically 25–35%, full-service restaurants 30–40%, salons 35–50% because commission is built in, and retail 10–20%. Track your own weekly trend rather than chasing someone else's benchmark.
How do I handle shift swaps without chaos?+
Write the policy down — who may swap with whom, and who approves — then use a system where staff post and claim shifts themselves with manager approval. Group-chat brokering doesn't scale past about six people.
Should I roster by hours or by expected revenue?+
By expected revenue or covers, then convert to hours. Rostering purely by hours encodes habit; rostering to demand adapts to how your trade has actually changed over the last quarter.
How do I stop unapproved overtime?+
Make it visible in real time rather than at payroll. Track clock-ins against scheduled hours daily, flag anyone approaching overtime mid-week, and require approval before the shift runs long — not after.

Keep reading