The Margin Hiding in Your Sub-Recipes

Ask most operators their food cost and they’ll give you a number with confidence. Ask them to show you how the sauce on their bestselling dish is costed and you’ll usually get a pause. That pause is where the money is.

The plate gets costed. The dish has a recipe card, a sell price, a target margin, and everyone feels covered. But a plate is built from components, and the components have components, and almost nobody costs all the way down. The margin you’re missing isn’t on the menu. It’s one or two levels beneath it, in the sub-recipes that feed everything and get re-costed by nobody.

The problem with costing the plate only

Here’s how it goes wrong. A dish card says: protein, sauce, garnish, starch. The protein has a clear price per portion, so that’s costed properly. But the sauce is listed as “80g house sauce,” and house sauce is priced at a number someone worked out eighteen months ago, when butter was cheaper, cream was cheaper, and the recipe used slightly less of both because the chef at the time was tighter with it.

So the plate cost on the card is fiction. Not wildly wrong — just quietly, persistently wrong, on every single cover, forever. And because the error lives in the sub-recipe rather than the plate, it’s invisible to anyone looking at the menu. You’d have to open up “house sauce,” re-cost it at today’s prices and today’s actual yield, to see it. Almost nobody does that on a schedule. They cost it once, at creation, and treat the number as permanent in a world where nothing about ingredient prices is permanent.

Multiply a small per-portion error across a sub-recipe that touches six dishes and thousands of covers a month, and you’re not talking about rounding. You’re talking about a serious, recurring leak that never appears as a line item anywhere, because it’s smeared invisibly across the whole menu.

Yield is where the real losses live

Ingredient price is the obvious half. Yield is the half that quietly does more damage, and sub-recipes are where yield gets lost.

A recipe card might say a sauce needs a kilo of an ingredient. But between the raw kilo and the finished sauce there’s trim, there’s reduction, there’s what sticks to the pot, there’s the batch that gets pushed too far and thrown out. The usable yield is lower than the raw weight, sometimes dramatically, and if you cost the sub-recipe on raw input weight rather than finished usable output, every downstream dish inherits a cost that’s understated. You think the sauce costs X per portion. It actually costs meaningfully more, because you paid for input you never served.

This gets worse with anything reduced, rendered, cooked down, or portioned by hand. A stock that reduces by half, a confit that loses weight, a dough that has offcuts: the finished, usable, plated cost of these is a different number from the shopping-list cost, and the gap is pure margin erosion if you never measure it. The kitchens that actually make their targets are the ones that cost sub-recipes on tested, finished yield, not on the back of the invoice.

Cross-utilisation cuts both ways

Sub-recipes are also where the smartest operators claw margin back, which is the flip side of the same coin. A well-designed menu shares sub-recipes deliberately: one base sauce, one pickle, one confit, one dough, feeding several dishes. That’s not just a kitchen-efficiency trick. It’s a costing advantage. Shared components turn over faster, so less is wasted and the effective cost per portion drops. A batch made for volume is cheaper per unit than the same thing made in six small lots.

But it cuts the other way when the sub-recipe is wrong. Because a shared component is shared, an error in it is leveraged across everything it touches. Cost your base sauce optimistically and you haven’t made one dish slightly wrong. You’ve made a third of your menu slightly wrong, all at once, in the same direction. The sub-recipe is a lever, and a lever moves both ways depending on whether you got it right.

What a proper costing discipline looks like

The operators who don’t leak margin here aren’t doing anything clever. They’re just doing the boring thing consistently. They cost sub-recipes as their own cards, on finished yield, at current prices. They re-cost the sub-recipes that carry the most volume on a regular cycle rather than at birth, because those are the ones where a small drift does the most damage. And they treat the sub-recipe layer as the foundation of the whole costing model, not an afterthought below the “real” recipe.

Then, when an ingredient price moves, and in this region ingredient prices move, sometimes hard, sometimes because a supply route hiccuped, they can see exactly which sub-recipes are affected and exactly which dishes inherit the change, and they can respond: re-engineer the recipe, adjust the portion, reprice the dish, or swap the ingredient. The operator who only costed at the plate level can’t do this. They feel the margin slipping in the monthly numbers but can’t point to where, because the leak is one level below anywhere they’ve ever looked.

The reframe

Food costing isn’t a single number you check monthly. It’s a structure with layers, and the layer everyone ignores is the one doing the most quiet damage. The plate cost is just the top of a stack, and it’s only as honest as the sub-recipes holding it up.

I’ve sat with kitchens that were sure they were running a healthy food cost, walked through what their three highest-volume sub-recipes actually cost at today’s prices and real yield, and watched the number they were proud of quietly come undone. The food didn’t change. The market moved underneath a costing model that was frozen in the past, one level down where nobody was looking.

So if you want to know where your missing margin went, don’t start with the menu. Open up the components. The plate is where you sell. The sub-recipe is where you win or lose, and it’s often quietly losing in a place you haven’t thought to check.