Restaurant guides

What should food cost percentage be in an Australian restaurant?

Michael Le · Sydney · Updated 9 August 2026

The target is 25% to 35% of food sales. Australian operators surveyed for the 2025 industry benchmarking report said the real average was 38% of turnover, with beverage at 26%. Most venues are running above the number they are aiming at, so if you are in the low thirties you are doing better than the field.

What Australian operators reported. R&CA 2025 Industry Benchmarking Report.
Cost lineShare of turnover
Wages, salaries and superannuation40.25%
Food cost38%
Beverage cost26%
What a healthy venue aims for. Zest Specialty Coffee Roasters.
Line itemTarget
Coffee cost of goods20% to 25% of coffee sales
Food cost of goods25% to 35% of food sales
Staff costs30% to 35% of sales
Rent10% of sales, 15% maximum
Utilities and overheads5% to 10% of sales
Net profit10% to 15%

Worth saying plainly: those two tables measure slightly different things. One is what operators reported across their whole turnover. The other is a target applied to food sales. Use the target to know where you are heading and the reported average to know whether you are normal. Do not use one to beat yourself up about the other.

Every dish on your menu was priced on a day that has already passed. Your suppliers did not check with you before they moved.

The four things that push food cost up

  1. Recipes that were never costed. Priced by checking the shop down the road and shaving a dollar. Nobody added the garnish, the oil, the container or the sauce that comes free.
  2. Portion drift. Nobody weighs anything after week three. A generous hand is a lovely quality in a person and an expensive one in a kitchen.
  3. Waste with no name. Prep that never sold, staff meals with no ceiling, stock that walks. It hides inside cost of goods and looks like inflation.
  4. Prices frozen while suppliers moved. The most common one. Your invoices went up four times last year. Your menu went up never.

How to find your real number

  1. Take opening stock, add purchases, subtract closing stock. That is your cost of goods for the period.
  2. Divide by sales for the same period. That is your percentage.
  3. Compare it to the target range, then to the reported average.
  4. Then cost your top ten sellers properly, line by line. Not your whole menu. Your top ten. They are most of your volume.

If your overall percentage looks fine but a top seller costs 55%, you do not have a food cost problem. You have one dish quietly funding itself out of the others.

What I got wrong

I kept dishes that lost money because regulars loved them. I told myself that was hospitality. It was avoidance with a nice story attached. You can absolutely keep a loved dish that runs hot on cost. You just have to know it, price the rest of the menu around it, and choose it on purpose.

Related guides

Sources. Reported industry figures are from the Restaurant & Catering Industry Association of Australia, 2025 Industry Benchmarking Report, published March 2026. Target ranges are from Zest Specialty Coffee Roasters, six key benchmarks every Australian cafe owner should know. Worked examples are arithmetic you can check yourself. Everything else is a first-person account of eleven years running one venue in Enmore, Sydney.

Venues measure cost of goods differently, so treat every published average as a signal to check your own numbers against, not a rule.

I ran Great Aunty Three in Enmore for eleven years. I know which of these numbers eat an owner alive, because they ate me.

If the room is full and the money is not, the leaks are the problem. That is what the Full Plate Method does, one client at a time, by application. If the venue runs fine and the room is the problem, that is Forkcast, my marketing studio for Sydney restaurants and cafes.

Want more of this?

The Note is my newsletter for operators rebuilding. Field updates, framework excerpts, and advance reading from No Plan B. New subscribers get the 8S Daily Tracker as a welcome gift. Sent only when there is something worth sending.