At one restaurant, instinct works. At thirty, only data does

Watch a good chef-owner for a week and you will see a management system that appears in no report. She knows her walk-in by feel — what came in, what is left, what tonight will consume — before any list is printed. She sees the new commis over-portioning from across the kitchen. Her numbers are late, approximate, occasionally wrong, and it barely matters: the numbers are not how she manages. Presence is.

Now look at the group F&B director with thirty sites. No walk-ins to feel, no pans to smell — thirty dashboards. Every judgement — which site is drifting, which manager needs support, which menu is dying, where the next franc of margin hides — arrives through a number somebody else produced. The moment an operation outgrows one person’s attention, the quality of the numbers becomes the quality of the management. There is no second channel.

That is why the performance gap between food service groups has moved. Purchasing conditions converge for groups of similar size; culinary technique circulates freely; everyone owns the same ovens. What separates groups today is data literacy — and unlike a chef or a supplier contract, it cannot be poached or copied by Friday.

What data literacy actually means in an F&B group

The phrase sounds like it belongs in a bank, so let me ground it in kitchen reality. Data literacy is not dashboards, and it is emphatically not more reports. It is the state in which the numbers deserve the decisions built on them.

It means trusting your recipe costs — because the recipes are maintained, with real yields and current ingredient prices, not costed once at menu launch and left to rot while the market moved. It means knowing your allergen declaration is correct today — because it flows from one source through every menu, label and tray card, instead of living in four parallel spreadsheets with four last-modified dates. It means reading a variance report and acting on it — because period cut-offs, transfers and stock valuation follow one convention everywhere, so a variance describes reality rather than the bookkeeping.

Groups in that state make decisions in days. Groups outside it hold meetings about whose number is right.

The uncomfortable middle: your ERP knows invoices, not recipes

Here is where most groups actually stand. They are not data-poor — they run serious ERP systems, SAP or Odoo or their peers, and finance has excellent visibility of one thing: money. Invoices, payments, general ledger, supplier accounts. All true, all auditable.

But the ERP does not know what a recipe is. It knows you bought 400 kilos of veal at a certain price; it does not know what the veal became, at what yield, in which dishes, at which margin, on which sites. It records the invoice and loses the transformation — and food service is the transformation. Between the invoice finance sees and the plate the kitchen sends, there is a chain of recipes, yields, portions and transfers that lives, in most groups, in spreadsheets and heads. That gap — between what finance sees and what the kitchen does — is precisely where margin dies, quietly, one unmeasured yield and one uncosted recipe change at a time.

The answer is not to ask the ERP to become a kitchen system; it was never built for that, and forcing it produces monsters. The answer is a culinary data layer that speaks the kitchen’s language — recipes, yields, allergens, menus, transfers — and reconciles with the ERP’s language of invoices and accounts. That is where we have deliberately positioned CalcMenu for decades: alongside SAP and Odoo, never in their place. Finance keeps its system of record for money; the kitchen finally gets a system of record for food; and the two agree with each other instead of arguing through a spreadsheet.

Allergens and nutrition: no longer differentiators — baseline

A decade ago, impeccable allergen data was a selling point. Today it is the legal floor. EU Regulation 1169/2011 and its national implementations, written allergen disclosure spreading through US state law, calorie labeling regimes from London to Quezon City — the direction is uniform: the composition of what you serve is becoming a regulated, published number.

For a multi-site group this changes the nature of the risk. A wrong cost estimate loses money; a wrong allergen declaration harms a guest and puts individuals in courtrooms. And the root cause of wrong declarations at scale is almost never ignorance — it is fragmentation: the recipe changed in the kitchen, and the change never reached the menu, the label, the tray card, the delivery platform. A group without a single source of truth for recipe composition is not carrying a cost risk. It is carrying a legal one, distributed across every site, every service, every day.

Built, not bought

The temptation, at group level, is to buy the endpoint: a beautiful dashboard layer over the existing chaos. It fails every time, because a dashboard on top of inconsistent data is fiction with better typography.

Data literacy is built as a discipline, in an unglamorous sequence. Systems that capture the work where it happens — recipes, stock, transfers, checks — rather than reconstructing it at month-end. Ownership: one named owner for recipe data, for allergen data, for costing conventions, so “the database” is somebody’s job and not everybody’s assumption. And one source of truth, from which every menu, label, cost report and declaration is derived, so that a change made once is true everywhere.

None of that can be purchased as a module and switched on. But every part of it can be started this quarter — and the groups that started three years ago are now managing while their competitors are still narrating.

At one restaurant, instinct works. At thirty, only data does — and only if the data deserves it.