Key takeaways
- Forecast covers for each day part first, then turn covers into labour hours using your own productivity standards, then build shifts.
- Use bookings on the book, the same week last year, booking pace, walk-in share, weather and events to forecast.
- Check every rota against a labour budget before publishing, and compare actual covers with the forecast every week.
- Demand is uncertain: 33.7% of accommodation and food businesses expected November 2026 turnover to fall[1], so plan core and flex shifts.
Why start from covers
Many rotas are built from last week's rota with a few names changed. That works until demand moves, and then the venue is either overstaffed on a wet Tuesday or running short on a sunny Saturday. Starting from forecast covers ties staffing to the thing that actually drives the work. It also gives you a record of why the rota looks the way it does, which helps when the new rules on shift notice and cancellations arrive; see shift notice and cancellation pay.
Uncertainty makes the method more valuable, not less. For November 2026, 33.7% of UK accommodation and food businesses expected turnover to fall, 41.5% expected it to stay the same and 10.8% expected a rise[1]. When demand is hard to call, a forecast you review weekly beats a habit.
- Expect a fall 33.7%
- Expect no change 41.5%
- Expect a rise 10.8%
- Other responses 14%
Source: Office for National Statistics, 2026[1]
Show the data
| Item | Value |
|---|---|
| Expect a fall | 33.7% |
| Expect no change | 41.5% |
| Expect a rise | 10.8% |
| Other responses | 14% |
Step 1: forecast covers by day part
Split each day into the periods that need different staffing, for example breakfast, lunch, afternoon, early evening and late evening. Forecast covers for each one, using several inputs rather than one.
- Bookings already on the book for that day part.
- Booking pace: how many bookings you usually have at this point before the day, compared with the final number. Example: if a Saturday is normally 70% booked a week out and this one is 50%, expect a quieter night or a late surge of walk-ins.
- Walk-in share: the proportion of covers that arrive without booking, which varies by day part and season.
- The same week last year, adjusted for any change in trading hours or capacity.
- Weather for terrace-led or walk-in-led venues, checked three to five days out.
- Events: football, concerts, school holidays, local markets, Christmas parties.
Calls and messages are an early signal too. If your phone line records enquiries, a rise in booking calls for a weekend is often visible before the bookings themselves. A system such as Ava, which logs every call it handles, gives you that record without anyone writing it down.
Step 2: turn covers into labour hours
Set simple productivity standards for each role, from your own experience: how many covers a server can look after in an hour on your floor, how many plates a chef can send from each section, how many drinks a bartender can make. Write them down and refine them over time.
| Role | Example standard | Fixed minimum |
|---|---|---|
| Server | One server per 15 covers per hour at peak | One server whenever open |
| Bartender | One bartender per 40 drinks per hour | One bartender whenever the bar is open |
| Chef on the line | One chef per 25 covers per hour | Two chefs whenever the kitchen is open |
| Kitchen porter | One per 60 covers per hour | One from first prep to close |
| Host or supervisor | One on shifts above 40 covers | Duty manager always on site |
Example: Saturday dinner is forecast at 140 covers, with the peak hour at 7.30pm carrying 60. At the Example standards, that hour needs four servers, three chefs on the line, one kitchen porter and a host. Early evening at 30 covers an hour needs two servers and two chefs.
Step 3: build shifts around the curve
Draw the demand curve
Plot forecast covers hour by hour for each day. The peaks and troughs show where staggered starts and finishes make sense.
Lay down the fixed minimums
Opening and closing crew, duty manager, the minimum kitchen. These shifts are core and should be published furthest ahead.
Add staggered peak shifts
Start extra servers and chefs 30 minutes before the peak builds and finish them as it falls, rather than everyone working the full evening.
Place breaks in the quiet times
Plan rest breaks into the troughs, so the law is met without leaving the floor short. See breaks and rest under the Working Time Regulations.
Mark flex shifts
Identify which shifts would be dropped or added if the forecast moves, and give them to staff who want variable hours.
Step 4: check against the labour budget
Before publishing, cost the rota and compare it with forecast sales. Most venues track labour as a percentage of sales and set a target range. Include employer on-costs, holiday accrual and any managers on salary, or the percentage will look better than it is.
Example: a week's rota costs £6,800 including on-costs, against forecast sales of £21,000, a labour cost of about 32%. The target range is 28% to 31%. The manager moves two midweek lunch shifts to flex and trims one late close, bringing the rota to £6,450, about 31%.
Wage floors keep rising: the National Living Wage is £12.71 an hour from April 2026 and £13.21 from April 2027[2]. Re-cost the standard rota every April so the budget reflects the new rates. For targets, see labour cost percentage targets.
Step 5: review forecast against actual
Every week, compare forecast covers with actual covers by day part, and staffed hours with the hours the standards say you needed. Note why any day was more than a tenth out. Within a couple of months you will see your own patterns: perhaps Thursday walk-ins are always underestimated, or wet Sundays are busier than expected because the walkers come in.
Setting labour standards you trust
Standards copied from another venue rarely fit. A server in a compact bistro with a short menu can look after more covers than one in a large dining room with tasting menus and wine pairings. Build your own in three weeks. For each busy service, note covers per hour and the staff on, then ask the duty manager and head chef to rate the service: comfortable, stretched or overstaffed. The hours rated comfortable give you a realistic standard; the stretched ones show where it breaks.
Review the standards whenever something changes: a new menu, a refit, a new till or ordering system, a terrace opening for the summer. Small changes to the menu can move kitchen standards a long way, especially if a dish needs more work at the pass.
Common rota mistakes
- Everyone starting at 5pm for a peak that does not build until 7pm.
- Rostering for the busiest Saturday of the year every Saturday.
- No one rostered for set-up, so the first hour of service starts behind.
- Breaks left to chance, so they are skipped on busy nights and taken at the worst moment on quiet ones.
- Forecasting from bookings alone in a venue where half the trade walks in.
- Ignoring the kitchen porter and glass collector, whose absence slows everyone else down.
- Publishing late because the forecast is not ready, which is a sign the process needs simplifying.
Plan the calendar, not just the week
Some swings are known months ahead. Not seasonally adjusted, UK accommodation and food jobs rose by about 91,000 between March and June 2025[3], as venues staffed up for spring and summer. NIQ describes Christmas and New Year as the sector's crucial trading period[4], and July 2026 was managed pubs' best month of the year, boosted by the men's football World Cup[5].
- Mark school holidays, bank holidays, big fixtures and local events on a 12-month calendar.
- Start recruiting seasonal staff six to eight weeks before you need them.
- Agree holiday blackouts for the busiest weeks with the team early in the year.
- Plan quieter weeks for training, deep cleans and staff holidays.
Rota planning checklist
- Covers forecast by day part, using bookings, pace, walk-ins, last year, weather and events.
- Written labour standards and fixed minimums for each role.
- Shifts staggered around the demand curve, with breaks in the troughs.
- Core and flex shifts marked; rota published two weeks ahead.
- Rota costed against forecast sales before publishing.
- Weekly forecast versus actual review.
- Twelve-month calendar of known peaks and troughs.
For tools that do much of this automatically, see rota and scheduling software compared.
Frequently asked questions
How do you forecast restaurant covers?
Combine bookings already on the book, booking pace compared with normal, the usual walk-in share, the same week last year, the weather outlook and local events, for each day part.
How many staff do I need per cover?
Set your own standards from experience, such as covers per server per hour at peak, plus fixed minimums. Example standards are a starting point, not a rule, because layout, menu and service style differ.
How far ahead should a rota be published?
Two weeks ahead is good practice, with core shifts fixed and flex shifts marked, so staff can plan and late changes are rare.
What labour cost percentage should a restaurant aim for?
It depends on the concept and pricing. Set a target range from your own accounts, include on-costs and salaried staff, and check every rota against it before publishing.
References
Office for National Statistics (2026). Business insights and impact on the UK economy, wave 166 dataset (opens in a new tab) ↑
GOV.UK (2026). National Minimum Wage and National Living Wage rates (opens in a new tab) ↑
Office for National Statistics (2025). JOBS02: Workforce jobs by industry (September 2026) (opens in a new tab) ↑
NIQ (CGA by NIQ) (2026). Hospitality loses 19 licensed premises a week ahead of crunch Christmas (Hospitality Market Monitor Q3 2026) (opens in a new tab) ↑
NIQ (CGA by NIQ) with RSM (2026). Bright September gives hospitality groups their best growth of 2026 (NIQ RSM Hospitality Business Tracker) (opens in a new tab) ↑