Hospitality Rostering + Demand Forecasting (AU)

When to use

Triggers:

• "Roster help" / "Shift planning for next week"
• "How many staff for [day/shift]?"
• "We were over/understaffed on [day]"
• "Wage cost is too high" / "Bottom-line hit last week"
• "Optimise the roster" / "Demand forecast for next week"
• Any AU hospitality operator planning shifts

Don't fire for:

• Payroll processing (different workflow)
• Industrial relations disputes / award interpretation edge cases (use Fair Work Commission directly or IR advice)
• Enterprise hospitality with existing workforce management software (the skill helps but their software is deeper)

Inputs needed

Before proposing a roster, gather:

1. Last 4 weeks of actual data — date, service period (breakfast/lunch/dinner/bar), covers / revenue / average spend, hours worked by role
2. This week's bookings — reservations on the books, any private functions, any known walk-ins (recurring regulars)
3. Team roster data — who works, hours they prefer, hours they can't work, any leave approved, junior-rate staff (under 21) vs adult-rate
4. Award context — which modern award applies (Restaurant Industry Award 2020 for most restaurants, Hospitality Industry (General) Award 2020 for pubs/hotels/cafes, Hair & Beauty Industry Award for salons)
5. Local context for next week — weather forecast, public holidays, school terms, local events (concerts, sports finals, festivals), competitor closures (Mondays are often competitor-closed)
6. Target labour-cost percentage — industry benchmark: cafe 30–35%, casual dining 28–32%, fine dining 32–38%, pub (food-heavy) 25–30%, hotel bar 20–25%

The three-step rostering workflow

Step 1 — Demand forecast per shift

For each service period next week, estimate:

Base demand — last 4 weeks' average for the same day-of-week and service period, adjusted for trend (growing, flat, declining).

Adjusters (each multiplies base):

• Weather — hot day (+15% cafe/bar, -5% indoor-only venues); rain (-10% outdoor; +5% pubs)
• Public holiday — Long weekend Friday / Monday (+25–40% depending on venue), day-of (venue-dependent — some venues close)
• School holidays — family-friendly venues (+15–25%), date-night venues (-10%)
• Local event — concert/stadium proximity (+40–60% pre-show / post-show shifts)
• Competitor dynamic — if neighbouring venue is closed for renovation, +15–25%; if a new competitor opened, -10–20% for 3 months then normalise

Apply adjusters, arrive at forecast covers / revenue.

Conservative AND optimistic bounds — forecast ± 20%. Roster to the midpoint, have on-call staff ready for the upside.

Step 2 — Staffing requirement per role

Convert forecast to staff hours per role. Example ratios (adjust for your venue):

Front of house:

• Floor staff (waiters): 1 per 20 covers per hour, breakfast slightly lighter, dinner tighter
• Baristas: 1 per 40 coffees per hour at peak
• Bar staff: 1 per 40 drinks per hour
• Floor manager / supervisor: 1 per shift for venues >40 covers at peak

Back of house:

• Kitchen: 1 chef per 15–20 covers per hour at peak for casual; 1 per 10 for fine dining
• Kitchen hand / dishie: 1 per 40 covers per hour
• Prep staff: 1–2 for the 2–3 hours before service, scaling to venue size

Management:

• Duty manager always present for alcohol-licensed venues
• Junior supervisor acceptable for non-licensed lunch shifts

Step 3 — Schedule within constraints

Now fit people into shifts honouring:

Fair Work Award constraints (common traps)

• Minimum engagement — most awards require 3-hour minimum shifts for casuals; shorter shifts are a breach
• Break entitlements — 30-minute unpaid meal break after 5 continuous hours; 10-minute paid break in shifts 4–6 hours
• Maximum daily hours — 10 ordinary hours per day; anything more is overtime
• 11-hour rest between shifts — roster an 11pm close + 6am open on the same person → award breach
• Weekend penalty rates — Sat 25–50% loading; Sun typically 50–75%; adjust cost model
• Public holiday rates — 225–250% of ordinary rate; usually opt to close marginal shifts
• Junior rates — under 21 staff have scaled rates; different by age band

Team preferences — known unavailability, requested days off, consistent shift patterns (people perform better on routine)

Skill coverage — not all staff are interchangeable. Senior bar staff per shift; at least one RSA-certified manager; at least one first-aid-certified staff member.

Output — roster proposal

Week of [date] — [Venue] — roster proposal

Forecast summary:
- Mon lunch: 30 covers (forecast) / 35 capacity — light
- Mon dinner: 80 covers / 100 capacity — normal
- Tue lunch: 45 covers / 60 capacity — normal
- Wed lunch: 60 (+15% — local business lunch event at Town Hall)
- ...
- Sat dinner: 150 (+35% — long weekend + forecast 28C sunny)
- Sun lunch: 120 (normal strong)

Labour cost forecast:
- Total scheduled hours: 342
- Weighted average rate: $32.40/hr (including penalty rates for weekend)
- Total labour cost: $11,100
- Forecast revenue: $36,500
- Labour cost %: 30.4% — target was 30% — within range

Watch items:
- Sat night: tight — no buffer if one person calls in sick. Have [Name] on standby.
- Sun lunch: Possibly under-rostered if weather holds. [Name] available to come in for the peak 12–2pm if called.
- Wed lunch: Over-rostered if the Town Hall event is quieter than expected; [Name] to be the first cut if covers are soft by 11:30.

Compliance:
- All shifts ≥ 3 hours ✓
- No <11hr rest gap ✓
- Mon public holiday — venue closed (opt to close marginal trade)
- Junior staff scheduled within their age rate bands ✓

Common rostering anti-patterns (AU hospitality)

• Hero-ing — rostering your best staff every Saturday night every week → burnout and turnover
• Thin-middle — under-rostering the less-busy shifts because "they're not busy" → a bad experience for every customer who happens to be there, and less room to train juniors
• Night-owls-on-breakfast — rostering night staff on early shifts too frequently → performance drop
• Perfect-for-normal, broken-for-abnormal — a roster that exactly matches forecast falls over when a group of 8 walks in at 8pm. Add a small buffer.
• Over-reliance on casuals — if >60% of your roster is casual, you have no continuity or training investment. Balance with a core part-time spine.
• Unpaid breaks as cost-reduction — if staff work through their break, that's worked time. No-break "savings" show up later as a Fair Work complaint.

Post-shift review

Compare actuals to forecast weekly:

• Covers actual vs forecast: variance %
• Revenue actual vs forecast: variance %
• Labour cost actual vs budget
• Breakage / incidents / complaints
• Team feedback: under-resourced? over-resourced? for which shifts?

Adjust forecast adjusters and staffing ratios based on the review. Over 4–8 weeks, this tightens materially.

Integration with existing tools

The skill outputs formats compatible with:

• Deputy, Tanda, RosterElf — common AU hospitality rostering software; skill outputs CSV they accept
• Humanforce, Square Staff, Lightspeed Restaurant — larger-venue options
• Plain spreadsheet — for venues not yet on dedicated software

The skill doesn't replace these. It produces the rostering logic; the software posts the roster to staff and handles punch-in/out, payroll integration.

What this skill does NOT do

• Guarantee Fair Work compliance. It flags common traps; the licensed manager or a Fair Work adviser must sign off on each roster, especially for complex awards or non-standard arrangements.
• Handle payroll. STP2 lodgement, super, PAYG — all separate workflows.
• Resolve IR disputes. If a team member has a grievance about their rostering, that goes through your formal dispute process.
• Replace a competent duty manager. Good managers read the room in real time — a roster is a starting point, not the final word.

Tier access

Pro. Rostering directly affects gross margin weekly; the skill's impact compounds across many weeks. Base-tier members get the three-step framework and demand adjusters; Pro adds the integration-ready outputs and the post-shift review cadence.

Related skills

• ai-roi-measurement-for-sme — measure whether improved rostering actually moved labour cost % and service quality
• spreadsheet-analyst — roster data analysis pairs here
• customer-retention-churn-signals — repeat hospitality customers are the ones served well during busy periods; both skills compound

References

• Fair Work — Restaurant Industry Award 2020
• Fair Work — Hospitality Industry (General) Award 2020
• Fair Work — Breaks and meal breaks