How to Standardise Recipes Across Restaurant Locations

Every multi-site operator has eaten the same dish at two of their own restaurants and been served two different dishes. Same name on the menu, different portion, different garnish, different margin. Guests notice the inconsistency; finance notices that site A runs three points more food cost than site B on an identical menu and nobody can explain which site is right.

Recipe standardisation is the fix, and it is a system rather than a laminated folder. This guide covers the model that works across groups, how to build the central library, how to distribute and enforce it, and how to measure whether it is holding. It builds on standard recipe costs, yields and batch recipes, which covers getting one recipe right; this guide covers keeping it right in twelve kitchens at once.

Why do recipes drift between sites?

Drift is rarely rebellion. A supplier shorts one site and a chef substitutes; a strong sous chef improves a dish and the improvement never travels; a portion creeps up because the bowl at that site is bigger; a new opening copies recipes from memory under time pressure. Each change is small, local and usually well intentioned, and none of it is recorded, so a year later the group runs five unofficial versions of its best seller.

Market conditions no longer hide it either. With the Office for National Statistics reporting food and non-alcoholic beverage inflation of just 1.7 per cent in the 12 months to June 2026, the calmest annual rate since August 2024, a site whose food cost runs points above its siblings cannot blame prices. Site-to-site spread on an identical menu is a recipe control problem, and it is one of the most recoverable numbers in a group P&L.

What does standardised actually mean? The Master and Variant Model

Standardised does not mean identical everywhere with no exceptions; groups that demand that get quiet non-compliance instead. The workable model has three tiers.

The rule that makes the model work: a variant is legitimate the moment it is documented and costed, and illegitimate until then. This gives sites a fast, honest route to local flexibility, which is exactly what removes the incentive to drift silently.

How do you build the master recipe library?

Four steps, in order. First, audit what actually exists: collect every version of every dish currently served, including the unofficial ones, without blame, because the audit fails if sites hide their real recipes. Second, consolidate: for each dish, pick or compose the master, usually the best-margin version that guests demonstrably like, and retire the rest. Third, cost the master properly, at usable yields with batch sub-recipes, so the standard carries a true cost, not an invoice-price fiction. Fourth, version it: every master carries a version number and an effective date, so there is never a question about which recipe is current.

How do you distribute and enforce the standard?

Distribution fails on paper and email because the folder at site four is always one version behind. The standard has to live where the kitchen works: on the phones and tablets the team already uses, updated centrally so a change to the master appears at every site the same morning, with the photo of the correct plate next to the spec. Enforcement then has two honest layers: spot checks against the photo and spec during service, and the numbers, which do not negotiate. Central production adds a third layer for groups that batch centrally; our guide to central and commissary kitchen stock control covers how standardised batches travel to sites without losing their costing.

How do you measure whether standardisation is holding?

One number audits the whole system: theoretical versus actual variance, cut by site, on the same day, by the same method. If every site runs the same master recipes, their theoretical usage is directly comparable, and the sites where actual usage runs furthest ahead are the sites where the standard is not holding, dish by dish. Our guide to reducing inventory variance in multi-location restaurants sets out that discipline from the COO’s seat, and the multi-site consolidation playbook covers the reporting structure that makes site comparison routine rather than a quarterly project.

Worked example (illustrative only). A six-site group runs the same menu everywhere. Five sites show a theoretical versus actual gap of about one point; site three shows four points. Because the recipes are standardised, the comparison is legitimate, and the dish-level breakdown at site three points at two dishes, both containing the same batch sauce. The site was making the sauce from memory at double richness: an undocumented variant. Recosting and retraining closes the gap. Without a standard, that four-point site would simply look like an expensive market. Figures are illustrative to show the method, not benchmarks.

How does software make the model run?

The Master and Variant Model is workable on discipline alone up to perhaps two or three sites; beyond that, the version control and the recosting labour defeat manual systems. Recipe management software holds the master library centrally, pushes updates to every site instantly, carries approved variants as controlled versions, and reprices every recipe at every site when a scanned supplier invoice changes a price. Paired with restaurant enterprise management reporting, the site-by-site variance audit assembles itself weekly. StockTake Online runs both cloud-based from a phone or tablet, with no dedicated hardware, which matters precisely because the standard has to live at the pass, not in the office.

Start with one dish: take your best seller, run its master version through the free Food and Beverage Cost Calculators at proper yields, and compare that cost with what each site believes the dish costs today. If the spread surprises you, Book a Demo and we will show the master library, variants and the site variance audit running live.

Key takeaways

• Standardisation is a system: one costed master per dish, documented approved variants, and everything else treated as drift.
• Give sites a fast, legitimate route to local variants; it removes the incentive to drift silently.
• Build the library by auditing real recipes without blame, consolidating to the best-margin master, costing at true yields, and versioning.
• The standard must live on the devices at the pass, updated centrally, with a photo next to the spec.
• Site-by-site theoretical versus actual variance is the weekly audit; comparable recipes make site comparison legitimate.

Frequently asked questions

How do you standardise recipes across restaurant locations? Create one costed master recipe per dish with yields, batch sub-recipes and portion specs, distribute it centrally to every site, allow only documented approved variants, and audit compliance weekly through site-by-site theoretical versus actual variance.

Should sites be allowed to modify recipes? Yes, through a controlled route. Local supplier substitutions and market adaptations are legitimate when they are requested, costed and approved as documented variants. Undocumented modification is drift, and an easy approval route is the best defence against it.

How do you roll out a recipe change across multiple sites? Update the master centrally with a new version number and effective date, distribute it digitally so every site sees the same version the same day, brief the change with the plate photo, and watch the affected dish in the next variance cycle to confirm adoption.

How does recipe standardisation affect allergen compliance? It strengthens it. When allergens are held at ingredient level on one master recipe, every site publishes the same accurate allergen information, and an approved variant carries its own allergen rollup, so a local substitution can never silently change a dish’s allergen profile.

Why is food cost different at each of our locations? On an identical menu, persistent site-to-site food cost spread usually means the sites are not cooking the same recipes: portion drift, undocumented substitutions or batch preps made differently. Standardised, costed recipes make the spread diagnosable dish by dish.