The Complete Guide to Menu Costing and Margin Control

Menu costing is the discipline of knowing, to the cent, what every plate you serve costs to produce, and then using that number to set prices, shape the menu, and protect the business when the cost of doing business moves. It sounds simple. In practice, most independent operators either skip it, do it once and never update it, or keep a spreadsheet that quietly drifts out of date the moment a supplier changes an invoice. The result is a menu where some items make money, some lose money, and nobody can say which is which until the monthly numbers come in and the answer is already baked in. By then the only levers left are blunt ones: a price increase across the board, or a quiet cut in portion size that guests notice before the owner does.

This guide is our attempt to lay the whole subject out in one place. We build recipe costing software, so we spend our days talking to chefs, owners, and operations managers about where their numbers go wrong. Those conversations shaped the six themes below: how to cost a single recipe properly, why sub-recipes matter more than people think, how food cost percentage and target margin turn into a price, how to find the items dragging the menu down, how to respond when ingredient prices climb, and how to run a monthly control loop that survives real kitchen life. Each theme links out to a deeper article on the specific problem. Read it straight through if you are setting up costing from scratch, or jump to the part that hurts right now.

Costing a Single Recipe the Right Way

The foundation of everything else is a recipe card that reflects reality. That means starting from the price you actually pay (from a recent invoice, not a memory of what chicken thighs cost last spring), converting the purchase unit into the unit you use in the kitchen, and applying a yield so that trim, bones, peels, and evaporation are paid for by the dish that causes them. A recipe costed from list prices and as-purchased weights will almost always look cheaper than it is, and that gap is where margin disappears without anyone noticing. The card does not need to be beautiful. It needs to be current, complete, and built on the same conventions as every other card in the kitchen.

The mechanics are straightforward once you break them into steps. For each ingredient, record the pack size and price, calculate a cost per base unit (per ounce, per gram, per each), adjust for yield to get an edible portion cost, and multiply by the quantity the recipe actually calls for. Add up the lines, then divide by the number of portions the batch produces. That per-portion figure is your plate cost. The trap is in the conversions: a case of tomatoes priced per pound, a recipe that calls for cups of diced tomato, and a yield loss from coring and skin all have to line up or the number is fiction. Most costing errors we see are not math errors. They are unit errors, where a tablespoon was treated as an ounce or a case was treated as a pound.

There is a second layer that many cards skip: the small stuff. Salt, oil, garnish, a lemon wedge, the paper the sandwich is wrapped in, the ramekin of sauce on the side. Individually these round to nothing, but across a hundred covers a night they add up to real money. A common approach is to include a modest allowance per plate for these incidentals rather than costing every grain of salt. What matters is consistency: pick a method, apply it to every recipe, and revisit the allowance when you see the actual purchase data. Portion control belongs in this conversation too, because a plate cost is only true if the line cooks are plating the quantity the card says. A recipe that calls for six ounces of protein and is routinely served at seven is under-costed by design, and no amount of spreadsheet precision fixes that.

A quick word on labor. Plate cost in the strict sense covers food and packaging only; labor is tracked separately as part of prime cost. Some operators fold a labor allocation into their recipe cards, especially for items with heavy prep like hand-rolled pasta or laminated dough. There is nothing wrong with that as long as it is explicit and consistent, because mixing labor into some cards and not others makes the whole menu impossible to compare. Our recommendation for most independents is to keep the card to food cost, then look at labor intensity as a separate lens when engineering the menu. It keeps the numbers clean and makes it obvious when a dish is expensive because of ingredients versus expensive because of hands.

Sub-Recipes, Batches, and Compounding Accuracy

Most menus are built from shared components: a house stock, a mother sauce, a spice blend, a pickled onion, a compound butter, a dressing. If you cost those components inline inside every dish that uses them, you end up with the same stock costed five different ways by five different people on five different days. Sub-recipes solve this. You cost the stock once, as its own recipe with its own yield, and then reference it as an ingredient everywhere it appears. When the price of onions goes up, you update one card and every dish that uses the stock updates with it. This is not a nice-to-have for large operations only; a taqueria with three salsas and a cafe with two house dressings benefit just as much.

The accuracy gain is bigger than it looks. A sauce that reduces by half has a very different cost per ounce than the sum of its raw ingredients divided by the starting volume. A brine that is mostly water but takes a full day of cooler space still costs something in the ingredients that flavor it. When each component carries its own yield and its own cost per unit, the dishes that consume them inherit that precision automatically. Without sub-recipes, every menu item carries its own private set of assumptions, and the errors compound quietly across the whole menu. One dish might assume the stock costs a few cents an ounce; another might assume twice that. Neither is checked against the other until someone notices the food cost report does not add up.

Sub-recipes also change how a kitchen thinks about batch size. If your demi-glace card says it produces 96 ounces and the actual batch reliably yields closer to 80, every dish using it is under-costed. Measuring real batch yields a few times a year and updating the cards is one of the highest-return tasks in the whole costing process. It takes a scale, a notebook, and a willingness to weigh the finished batch instead of trusting the recipe header. It is also the kind of task that a spreadsheet makes painful and a purpose-built tool makes routine, which is why nested recipes are usually the first feature people notice when they move off a workbook.

Nesting can go several levels deep. A finished dish might use a sauce, which uses a stock, which uses a mirepoix that is itself prepped in bulk. Each level should carry its own yield and its own portion unit. The discipline to maintain is simple: every component that is made in a batch and used in more than one place gets its own card. Components made fresh for a single dish can stay inline. Draw that line once, write it down, and hold to it, because the alternative is a kitchen where half the shared components are tracked and the other half are guessed.

Food Cost Percentage, Target Margin, and the Price

Food cost percentage is the plate cost divided by the menu price. It is the most widely used number in restaurant finance, and also one of the most misunderstood. A low percentage is not automatically good and a high percentage is not automatically bad. A steak with a higher food cost percentage can put far more dollars of margin on the table than a soup with a very low one. The percentage is a diagnostic, a way to compare items and periods on a common scale, not a goal in itself. Operators who chase the lowest possible percentage often end up with a menu full of cheap items that nobody wants to order twice.

Pricing from plate cost works in two directions. Working forward, you take the plate cost and divide by your target food cost percentage to get a floor price, then adjust for what the market and the concept will bear. Working backward, you take the price the menu can support and ask whether the plate cost leaves enough contribution margin to cover labor, occupancy, and profit. Most operators use a blend: percentage targets for consistency, contribution margin in dollars for the items that anchor the menu. Many operators use a target somewhere in the high twenties to mid thirties as a starting point, but the right number depends heavily on concept, format, service model, and what else the business has to cover. A counter-service bakery and a full-service steakhouse should not be using the same target.

Whichever direction you work, the price you land on has to be tested against reality: what nearby competitors charge, what your guests expect at your format, and how the item sits alongside the rest of the menu. Rounding, psychological pricing, and the placement of high-margin items all matter, but they are second-order. The first-order question is always the same: does this price, at the volume we actually sell, leave enough margin after the plate cost is honestly counted? If the answer is no, no amount of menu design fixes it. And if the answer is yes only because the plate cost is stale, the menu is carrying a problem that will surface later.

One more distinction worth holding onto: theoretical versus actual. The food cost percentage you calculate from recipe cards is theoretical, meaning it assumes perfect portioning, zero waste, and prices exactly as recorded. The percentage you get from inventory counts and purchases is actual. Pricing decisions should be made from theoretical cost, because that is the cost you control by design. The difference between the two is an operations problem, not a pricing problem, and confusing them leads to raising prices to cover waste that should have been fixed on the line. Keep the two numbers side by side, understand why they differ, and resist the urge to price your way out of an execution gap.

Menu Engineering: Finding the Winners and the Losers

Once every item has an honest plate cost and a price, the menu becomes a dataset. Combine plate cost, price, and sales mix from the point of sale, and you can sort items into four rough groups: high margin and high volume (protect these), high margin and low volume (promote them), low margin and high volume (re-cost, reprice, or re-portion), and low margin and low volume (cut or rework). This is the core of menu engineering, and it only works if the plate costs are current. Stale costs push items into the wrong quadrant and lead to confident decisions built on the wrong numbers. We have seen operators promote an item as a star that was actually losing money on every order, purely because the card had not been updated since the recipe was written.

The margin losers are rarely the obvious ones. A signature dish that everyone assumes is a profit engine may carry a plate cost that crept up over eighteen months of supplier increases. A side that sells constantly may lose money on every order because the portion grew, the packaging changed, or a sub-recipe was never re-costed. Finding these items is a matter of looking at contribution margin per item and per period, not just percentage, and cross-checking against actual usage when something looks too good or too bad to be true. The losers tend to hide in the middle of the menu, in the items that nobody thinks about because they are neither new nor famous.

Fixing a loser has more options than raising the price. Trim the portion slightly, swap one expensive component for a comparable one, adjust the plating so the costly element is featured but smaller, move it to a section of the menu where it sells less, or bundle it in a way that lifts the check. Sometimes the right answer is to keep a low-margin item because it drives traffic that buys high-margin drinks and desserts. The point of the analysis is to make that a deliberate choice, not an accident you discover at year end. Every item on the menu should be able to answer the question of why it is there.

Sales mix is the ingredient most people leave out. A menu engineering exercise done from plate cost and price alone treats every item as if it sells equally, which is never true. Pull item-level sales from the point of sale for a representative period (a full month is a reasonable minimum, and a quarter is better for seasonal menus) and weight the analysis accordingly. An item that sells forty times a night matters far more to the blended number than one that sells four, even if the per-plate margin on the rare item is spectacular. This is also why menu engineering should be repeated, not done once: the mix shifts with the seasons, the weather, and whatever the neighborhood is talking about.

Rising Ingredient Prices and Re-Costing Cadence

Ingredient prices do not move politely. Proteins, dairy, eggs, cooking oil, and produce all have their own cycles, and a supplier price change can arrive mid-month with no warning beyond a line on an invoice. The only real defense is a costing system that is easy to update, so that when the price of a key ingredient jumps, you can see within minutes which dishes are affected, by how much, and what the new food cost percentage looks like at the current menu price. Operators who cost annually are always reacting to last year's problem. Operators who cost continuously get to choose their response instead of having it forced on them.

A practical cadence looks something like this. Update the prices of your top movers (the ingredients that account for the largest share of spend) every time an invoice shows a change. Review the full ingredient list monthly. Re-measure yields on the most expensive sub-recipes a few times a year. Reprint or republish the menu when the cumulative drift crosses a threshold you set in advance, rather than waiting for a crisis. The goal is a small amount of steady work instead of a painful annual reconstruction of every card. Ten minutes a week with the invoices beats a lost weekend every January.

When prices rise, the menu response should be proportional and deliberate. Small, frequent price adjustments are usually easier for guests to accept than one large jump. Substitutions, portion tweaks, and a rotating special that absorbs the most volatile ingredient can buy time. Some operators build a small buffer into their target margin on volatile items specifically so that a modest price increase does not force an immediate menu change. None of this works without current plate costs, which is why re-costing discipline is the real hedge against inflation. The kitchens that weather a price spike best are not the ones with the cheapest suppliers; they are the ones that knew about the spike first and had options ready.

It also helps to know which ingredients actually move the needle. In most kitchens a relatively short list of ingredients accounts for the bulk of food spend, usually the primary proteins, dairy, oil, and a few high-use produce items. Tracking those closely and treating the long tail more loosely is a reasonable trade of precision for effort. When one of the key items spikes, run the numbers for every dish that uses it before deciding whether to absorb the increase, adjust the portion, or move the price. Doing that in an afternoon rather than a week is the whole point of keeping the system current.

Systems: From Spreadsheets to a Monthly Control Loop

Nearly every operator starts with a spreadsheet, and for a short menu that is fine. The trouble begins when the menu grows, when sub-recipes multiply, when more than one person needs to edit, and when the sheet has to be reconciled against purchases and sales. Formulas break, versions fork, and the person who built the workbook leaves. The spreadsheet did not fail because spreadsheets are bad; it failed because recipe costing is a relational problem (ingredients feed sub-recipes, sub-recipes feed dishes, dishes feed the menu) and a grid is a poor fit for relationships. Every price change has to be chased through every tab by hand, and eventually someone stops chasing.

Whatever tool you use, the monthly control loop is the same. Start with opening inventory, add purchases, subtract closing inventory to get actual cost of goods sold. Divide by food sales to get actual food cost percentage. Compare that to the theoretical food cost you would expect from plate costs multiplied by the sales mix. The gap between actual and theoretical is where waste, theft, over-portioning, and unrecorded comps live. A small, stable gap is healthy. A growing gap is the earliest warning you will get that something in the kitchen has changed. Running this loop monthly, on the same day, with the same method, is worth more than any single clever analysis.

Blended food cost across the whole menu is the number that ties all of this together, and it is the one most operators actually manage against. Individual items can and should vary widely; a bar snack and a dry-aged ribeye have no business sharing a target. What has to hold is the weighted average across what you actually sell. Managing it well means keeping plate costs current, keeping the sales mix in view, and adjusting prices, portions, and menu placement so the blend stays where the business needs it. That is the loop. Tools like ours exist to make it routine rather than heroic, but the discipline matters more than the software.

If you are deciding whether to move off the spreadsheet, the honest test is not menu size but pain: how long it takes to update a price, how many people can safely edit, whether you trust the numbers enough to reprint a menu from them, and whether you can answer 'what does this dish cost today' without opening three files. When those answers get uncomfortable, a dedicated tool pays for itself quickly. When they are still fine, keep the spreadsheet and put the effort into keeping it current instead. The worst outcome is not the wrong tool; it is a tool of any kind that nobody updates.

More guides on this topic

Further reading from the PlateCostr blog, each answering one specific question in depth.

Menu costing is not a project you finish. It is a habit with a handful of moving parts: honest recipe cards, sub-recipes that carry their own yields, a pricing method that respects both percentage and dollars, a regular look at which items earn their place, a re-costing rhythm that keeps pace with suppliers, and a monthly reconciliation that catches drift early. Operators who get this right do not necessarily have lower food cost than their neighbors. What they have is fewer surprises, faster reactions when prices move, and a menu where every item is there on purpose.

If you are starting from nothing, do not try to cost the whole menu in a weekend. Start with the ten items you sell the most, get those cards right down to the yield and the garnish, and build outward from there. Add sub-recipes as soon as you notice a component appearing in more than one dish. Set a monthly date for the inventory count and the actual-versus-theoretical comparison, and keep it. The articles linked throughout this guide go deeper on each step. The order does not matter much; the consistency does.