Australian food and beverage businesses need roughly A$100 to A$125 of sales for every labour hour they buy, a band Skai Solutions derives from the ABS finding that labour cost 30.4 per cent of industry sales in 2024-25.
Almost every sales per labour hour target an operator finds in search is American, and the American wage structure has no casual loading, no weekend penalty rates and no superannuation guarantee. This page builds the Australian and New Zealand numbers from published figures instead, shows the arithmetic, and gives you the shift level method rather than a weekly average.
There is no official Australian sales per labour hour figure, because the Australian Bureau of Statistics does not publish labour hours for the industry. What it does publish is both halves of the ratio that sits underneath it, and those are measured rather than surveyed.
From the ABS Australian Industry release for 2024-25, published 18 September 2026, food and beverage services took A$134,077 million in sales and service income and carried A$40,816 million in total labour costs. Labour is therefore 30.4 per cent of industry sales.
Turn that into an hourly figure using the award the sector actually pays. A Level 2 employee under the Restaurant Industry Award MA000119 is on A$27.08 an hour ordinary time, or A$33.85 an hour as a casual. Add the 12 per cent superannuation guarantee and the all in hourly cost is A$30.33 and A$37.91. Divide each by 0.304 and you get the sales each labour hour has to produce for the venue to sit on the industry ratio.
| Who is on the floor | All in hourly cost | Sales per labour hour needed |
|---|---|---|
| Level 2, full time or part time | A$30.33 | about A$100 |
| Level 5 supervisor, full time | A$35.06 | about A$115 |
| Level 2, casual | A$37.91 | about A$125 |
The benchmark, and this is our arithmetic on published figures rather than a published benchmark: roughly A$100 to A$125 of sales for every labour hour you buy. A venue running mostly casuals sits at the top of that band, one running mostly permanents at the bottom. It is the same relationship our page on restaurant labour cost percentage describes from the other direction, expressed in a unit a duty manager can use on a Tuesday.
One sense check from the same ABS release: sales per person employed across the industry were A$121,833 in 2024-25, up from A$119,234. That is per head per year rather than per hour, but it moves the way you would expect against the figures above.
Because the wage share of revenue is higher there, not because wages are.
The Restaurant Association of New Zealand's 2025 Hospitality Report puts wage costs at about 40 per cent of hospitality revenue. The Hospitality New Zealand and Restaurant Association Remuneration Survey 2025 puts the average hourly wage in the sector at NZ$27.84. Divide one by the other and the New Zealand figure lands at about NZ$70 of sales per labour hour.
Handle that number carefully, and we would rather say so than round it off. The Australian figure comes from a statistical agency census of the whole industry. The New Zealand one comes from two industry surveys on different bases, and New Zealand has no hospitality award and no statutory weekend penalty rates, so the cost of an hour moves differently across a week. Treat A$100 to A$125 as a benchmark and NZ$70 as an indication. Our New Zealand hospitality wages page has the underlying rates.
The gap is still the useful part. A New Zealand operator who reads an Australian or American sales per labour hour target and applies it will conclude the venue is failing when it is merely operating in a market with a different wage share.
Two numbers, one division, and the trap is in the second number.
Do it by shift, not by week. A single weekly figure of A$108 can hide a lunch running at A$64 and a dinner at A$140, and the lunch is the decision you actually have to make.
Working the other way is just as useful. If you need A$110 per hour and you are rostering 210 hours a week, the roster is asking the venue for A$23,100 a week, or A$1.2 million a year. If that is not what the venue takes, the roster is wrong, not the team.
Work in this order, because the first two are free and the last one is the one everybody reaches for first.
Check the number before you act on it. Salaried managers and unpaid owner hours are the two things most often missing. Adding them can move a figure by 15 per cent, which is the difference between a rostering problem and no problem.
Fix the shape of the shift, not its size. Most venues are not overstaffed across a shift, they are overstaffed for the first 90 minutes of it. Staggered starts at 30 minute intervals against your actual sales by half hour will usually find hours without anyone losing a shift.
Then look at the denominator from the other side. Sales per labour hour rises if sales rise, and that is the better lever when the dining room is already tight. Our pages on table turnover and prime cost cover that half.
Cut hours last. A roster cut to hit a ratio usually reappears as overtime, a worse service, or a resignation, and what turnover costs is larger than the saving.
More than it looks, because the phone is the one task in the venue that consumes labour hours without ever appearing on a roster.
Nobody is rostered to answer the phone. It is answered by whoever is nearest, mid task, and the hour it takes is charged to whatever that person was supposed to be doing. At A$110 of sales per labour hour, a person spending 40 minutes a shift on the phone is 0.67 of an hour the venue paid for and did not get service from: about A$74 of capacity, per shift, before you count the calls that rang out.
That is the number worth putting against the cost of answering the phone another way. Our page on answering service costs prices the market, and an AI receptionist takes the task off the roster entirely rather than moving it to a quieter person. Setup at Skai Solutions takes five days.
The honest version: if your phone rings four times a shift, this is a rounding error and you should spend your attention on the roster instead. The arithmetic only turns interesting past about 20 calls a day.
The companion measure on the customer side rather than the roster side is average spend per head.
Sales per labour hour is the measure menu engineering cannot see, because contribution margin counts food cost and ignores the minutes a dish takes. Read them together: menu engineering and its four blind spots.
Is sales per labour hour better than labour cost percentage? It is more actionable and less stable. Labour cost percentage is the right number for a profit and loss review; sales per labour hour is the right number for writing next week's roster, because it is denominated in the thing you are actually deciding.
Should I include the kitchen? Yes, if you want a venue level figure. Split front of house and kitchen separately as well, because they move for different reasons and a combined number hides both.
Does it work for a cafe? Yes, and the band is usually lower, because average spend per head is lower. Build your own baseline from three good weeks rather than importing the restaurant band. Our page on how many staff a cafe needs covers the staffing side.
What about public holidays? The ratio breaks on a public holiday, because the hourly cost can be 225 per cent of ordinary. Either exclude those days or set a separate target for them. See public holiday surcharges.