Restaurateurs are told they need more data. Almost none do. The data is already there, rung up one ticket at a time by the till, sitting in an export nobody has an evening to read. This guide describes what is in that file and what a careful reading returns, so you know what to expect from a margin report, whoever builds it for you.
What is in the export
A till export is unglamorous: one line per sale, with the item, the quantity, the price, the time, and the address when you run more than one. Three to six months of it is enough to see patterns instead of anecdotes.
Pulling it takes about five minutes, from the back office of any of the common systems. It requires no technician and changes nothing in how you work. It is, by a wide margin, the cheapest raw material in your business.
The four answers a margin reading returns
- Margin by dish, not popularity. Your best seller and your best earner are rarely the same plate. Popularity crossed with recipe cost sorts the menu into what deserves a push, a price adjustment, or a quiet retirement.
- Hours that pay. Sales by hour and by service show where staff hours are covered by revenue and where they are not.
- The honest gap between addresses. Same menu, different results: the export shows whether the difference is traffic, mix or pricing.
- Two or three priced decisions. The point of the reading is not a dashboard, it is a short list: this price, this dish, this service, with the numbers attached.
From one report to a habit
A single reading is a photograph. The useful version is a tool that redoes the analysis every time you drop in a fresh export: the menu changes, seasons turn, and last quarter's decisions get checked against what actually happened.
Built properly, that tool lives in your accounts and asks one gesture a month of you. No new system, no integration project, no dependence on whoever built it.
What to be suspicious of
Be wary of any margin analysis that arrives without your recipe costs: sales alone flatter the popular and hide the expensive. Be wary of averages across addresses, which bury exactly the gap you need to see. And refuse projections dressed as findings: a reading of your history is a fact, a promise about next quarter is a guess wearing its clothes.
A serious report separates the two, tells you what it could not see, and puts a number on every recommendation it makes.