A restaurant has too many variables for anyone to promise results — and Match Sous doesn't. What it does is bring into the open what today gets decided in the dark: where each dish sits against what your area charges, what was left when the month closed, and what changes when you adjust something. It's the method of people who ran the industry's largest groups, turned into continuous oversight. The decision is still yours — it just comes with a number behind it. Today the app opens on the menu, which is what's already in your hands — with the specialist at your side, to ask "if I raise this dish by R$ 20, what's the risk?"; the month, the days ahead and the group are part of the same method and come next.
The methodology running here was built by professionals who operated large restaurant groups and food-industry companies — the same analysis that today gets bought as an audit, costs a lot, happens once a year and ends in a document nobody opens again. Match Sous turns that method into a continuous process, inside the app, for any restaurant: from the ten-unit group to the diner, the bakery and the neighborhood café that never had access to this kind of analysis because it always cost more than it was worth.
Menu engineering, recipe costing, food cost, prime cost, vendor negotiation — criteria refined over years of real operations at groups and in the industry. What lived in a handful of people's heads became rules built into the product.
Classic consulting is a snapshot: once a year, expensive, out of date by the following month. Here the diagnosis is alive — it changes when your costs change, when your prices change, when the market around you changes.
The same criteria that hold in one restaurant hold across an entire group, location by location, comparable side by side — without depending on a consultant's calendar or a manager's spreadsheet.
You start with the menu because it's already in your hands — no system to integrate, no staff to train, no rollout to wait for. It works for a thirty-dish menu and for the diner's chalkboard with twelve. The other three fronts are part of the same method: the month that closes, the days ahead, and the whole group under one yardstick. They arrive later, in the right order: in the app they show in the menu as "coming soon", and you ask to be notified when they open.
Each dish compared with the real price distribution in your area, and crossed with the margin it leaves. With the operation specialist at your side: you ask "if I raise this dish by R$ 20, what's the risk?" and it answers with your numbers. The comparison leaves the screen as a PDF, spreadsheet or message — and a consultant can deliver it under their own brand.
See the comparisonWhat came in, what went to suppliers, what went to payroll — and what was left, down to the last line. From the second month on, every number gets its curve against your target line.
Coming soon — opens after the menu. You ask to be notified inside the app.Here the product stops just tallying what already happened: what's coming due, how much you usually bring in over that stretch and — if you tell it how much cash you have today — how long the money lasts.
Coming soon — opens after the menu. You ask to be notified inside the app.More than one restaurant under the same yardstick: who carries the result, who drains it, and the same dish location by location — each measured against its own market.
Coming soon — opens after the menu. You ask to be notified inside the app.Plot each dish's margin against how much it sells per week — the sales you enter, or the report from your POS — and every item on your menu lands in one of four quadrants. It's the reading a spreadsheet doesn't give you — and it's where the decision to protect, push, rework or cut shows up by name.
The menu's champion. The screen says: protect the price and leave the recipe alone.
Volume that carries the operation, but with cost pushing back. Go after ingredients before touching the price.
A good dish hiding on the menu. The screen suggests reworking its description or placement — not cutting it.
A natural candidate to cut — unless it's part of the restaurant's identity. The reading is yours; the verdict isn't blind.
The dividing line on each axis is the menu's own median — a classification relative to your restaurant, not an absolute yardstick. A dish sitting right on the line is marked as leaning one way, not as a verdict.
With sales volumes and cost per dish (entered by you or estimated by AI), the platform rolls the entire menu up into a single verdict: your restaurant comes out healthy, needs attention, or is critical. Nothing invented — it only covers dishes that already have data, and says clearly when something is missing.
A kiosk, a container or a food-hall stall doesn't compete with the sit-down restaurant down the street: it competes with whoever's twenty steps away. You declare the food court and the comparison shifts to your physical neighbors — every stall together, which is how the customer actually decides where to eat.
Type the dish and the area. Match Sous returns the statistical price distribution, the restaurants that make up the sample and where your price sits relative to them — if it's already on file.
You come in through the menu, with the operation specialist alongside from day one. The other fronts already exist and are in use in the pilot: they show in the app menu as "coming soon", open to everyone once the menu has proven its value, and inside the app you can ask to be notified. What's still being built is labeled as such.
Each dish against what your area really charges. What to protect, what to rework and what to cut — and the whole menu in a single verdict. With the recipe card estimated by AI when you don't have the cost at hand, the descriptions rewritten to sell the dishes that carry the margin, and the specialist to discuss each decision.
See the comparisonOnce the month closes, you see what came in, what went to suppliers, what went to payroll — and what was really left, down to the last line. From the second month on, the curve shows where the business is heading. And if you say how much cash you have today, how long the money lasts.
R$34k was left — 12.5% of what came in, above the 12% you said you wanted.
The cost your menu promises against what the month actually spent, with the gap in R$ and the dishes to start looking at. Your ingredient prices are stored with the date they took effect — which is what keeps recipe costs from quietly going stale.
The menu promised 26% in ingredients; the month spent 31% — the gap in reais, and the dishes to look at first.
Several restaurants under the same yardstick: who carries the result, who drains it, and the same dish location by location — each measured against its own market, because different prices across neighborhoods aren't a problem. The consolidated view knows which part of the result is yours.
The group closed positive, but one location eats up almost everything the other three left over.
Half of it already works: the products-sold report your point of sale exports goes in as it comes out — spreadsheet, PDF or a screenshot — and matches your menu's dishes on its own. What's missing is the last piece: the restaurant not needing even that. It's what we're building now, and it's what guarantees the numbers don't die in two weeks: data that depends on someone remembering to type it is data that stops being fed.
The figures above come from demo restaurants — they show the shape of the reading, not a promised result.
Nobody types dish by dish, and nobody needs to know costs by heart. On the menu, today, it does this:
What it doesn't do: it doesn't set prices, doesn't make up data and doesn't hide when something is an estimate. And the operation specialist, in Conversation, answers with your menu and your area in hand — and remembers what you've already decided.
São Paulo, Belo Horizonte, Salvador, Recife and Fortaleza — each restaurant compared only with its own city. Extra coverage in Tamboré/Alphaville, Pinheiros and Itaim.
Photos and PDFs take one to four minutes to read; pasted text, under a minute. The screen shows a clock and tells you when it's done.
A 300 g picanha against a 500 g rib on the same yardstick — for drinks, R$/100 ml.
You upload the menu photo or PDF, the spreadsheet, or paste the text. The model extracts each dish, description, portion size and price — no manual typing.
Your area's price database comes from collecting public menus and grows with what gets uploaded to the platform. In this version it's preloaded, and the How it works page says exactly which part is collected data and which is still a calibrated reference.
Dishes with different names but similar descriptions are recognized as the same item. Each gets a median, a range, a sample size and your price's relative position.
No screen saves a restaurant — there are too many variables, and anyone who runs one knows that better than any software. What Match Sous does is transparency and oversight: your operation's numbers in plain view (today the menu's, next the month's) and the industry's engineering — menu engineering first — working for you. The decision is still yours. It just gets easier to make.
Start with my menu