What should labor cost be?

Where it comes from. The National Restaurant Association’s Restaurant Operations Data Abstract, 2025 edition, reporting 2024 figures from more than nine hundred restaurants. Salaries and wages including benefits: a median of 36.5% of sales at full-service restaurants. Write the year beside the number whenever you use it. A bare figure with no year cannot be checked, and this one moves.

The gap is the finding. The operators who finished 2024 in profit ran wages at 34.2% of sales. The ones who finished in the red ran 42.9%. Eight and a half points of sales between the two groups. On a restaurant doing $1.2 million a year, that is a little over a hundred thousand dollars, which is the difference between a year you would do again and a year you would not. And here is the good part: it is the one line on the whole statement that moves every single week, which means it is the one you can still do something about.

The number the internet gives you. Search the question and you get a range, twenty-five to thirty-five percent, often with thirty called the industry standard. Notice where that sits. Below the measured median. Below what the profitable half ran. Nobody will show you the survey behind it, because there is not one you can read. The Association’s number has a year, a sample and a page. Use that one.

Three things the Association does not publish. Worth knowing, because all three get quoted in its name.

• A margin of three to five percent. It is not in the Association’s 2026 State of the Industry report; we pulled the document and searched it. The published figure for 2024 is 2.8% of sales before income tax, down from 4% in 2019.
• A full-service prime cost. The Association publishes one for limited service, sixty-five cents of every sales dollar, and nothing for full service. What circulates is a food median and a wage median from different restaurants, added up. Medians do not add. Why, and what to do instead.
• A split by menu segment. Full service and limited service, that is all. A table with a fine dining row and a casual dining row under the Association’s name has two rows that are theirs and the rest are somebody’s guess wearing their badge.

So what should yours be? Nobody who has not seen your restaurant can tell you, and the honest version of this article ends there. What the numbers above are good for is one thing: if you are north of 42%, you are where the restaurants that lost money were, and that is worth a hard week of looking.

The comparison that pays is smaller and closer. Your own last eight Tuesdays against this Tuesday. Same room, same menu, same market. If labor moved three points, something changed and you can go and find it. If you are at 38% and the median is 36.5%, you have learned nothing, because the median restaurant is not your restaurant.

How to see it every morning. Hours worked last night by job, at what each hour costs you, over last night’s net sales. Then the same line for the same weekday over the last eight weeks. That is the whole report. Once it is in front of you every morning, the Tuesday meeting stops asking what labor should be and starts asking why it moved, which is a much better meeting. That is the line Syphor writes: last night’s labor by job against your own normal for that weekday, with the hours behind it. Never against a median, and never against another business.