Ghost Kitchen Food Cost: The Margin Math That Decides If You Survive
Ghost kitchen food cost is a harder problem than restaurant food cost, and not because the ingredients cost more. It’s because a delivery-only kitchen has nowhere to hide a weak plate. There’s no dining room, no $14 glass of wine, no dessert someone adds on impulse, no check average quietly carrying the one dish that runs too rich. Every order stands on its own — and the delivery app takes its cut off the top before you see a dollar.
That’s the whole game. In a restaurant, food cost is one number among several that a busy night can paper over. In a ghost kitchen, food cost is the margin, and it has to survive a commission you don’t control. Get it wrong on a single hero item and you can run at full volume, look busy on every screen, and still lose money on most of the tickets.
Why the delivery commission changes everything
A typical third-party delivery commission runs somewhere in the 15–30% range of the order total, depending on the app and the plan you’re on. That comes out before packaging, before food, before labor. So the money you actually have to build a plate with isn’t the menu price — it’s the menu price minus the app’s cut.
Here’s what that does to a single $16 order:
| Per order | Amount |
|---|---|
| Delivery menu price | $16.00 |
| Delivery commission (~25%) | −$4.00 |
| Net revenue you keep | $12.00 |
| Food cost | −$2.80 |
| Packaging | −$0.85 |
| Left for labor + overhead + profit | $8.35 |
The food itself is only $2.80 — about 17.5% of the menu price, which looks fantastic. But you never got the whole $16. Against the $12 you actually keep, that same food is 23%, and the commission alone ate a quarter of the ticket before the kitchen turned on. This is the number that fools new operators: they cost food against the sticker price, see a healthy percentage, and never account for the four dollars that left the building first.
Cost the plate the way delivery actually pays you
The fix isn’t complicated, it’s just honest. Cost every delivery item two ways:
- Food cost as a share of the price you keep (net of commission), not the menu price. That’s the number that tells you whether the plate is really healthy.
- Packaging as a real line item. The clamshell, the lid, the sauce cup, the bag, the cutlery — on a delivery order that’s often $0.60–$1.00 you’d never pay in a dining room. It belongs in the plate cost, every time.
Run one of your top sellers through the free food cost calculator and do exactly this: enter the net price after the app’s cut, and add a packaging line. The percentage you get back is your real ghost-kitchen food cost. It’s almost always higher than the one in your head.
The four levers still work — in this order
When a delivery item’s food cost comes back too high, you fix it the same way any operator does, cheapest lever first. Repricing is last, and on delivery apps it’s especially blunt because a higher price shows up right next to a competitor’s.
- Portion first. The 6-ounce protein that’s really going out at 7.5 is your biggest, quietest leak — and at delivery volume it repeats hundreds of times a week. This is the portion you eyeball, and it’s where most of the drift lives.
- Packaging second. This lever is unique to you. The right-sized container, buying the case instead of the sleeve, dropping the extra sauce cup nobody asked for — packaging is a food-cost lever in a ghost kitchen, and it’s usually softer than the menu.
- Supplier third. Same as any kitchen: verify the invoice, push back on the creep, know what you’re actually paying this week.
- Reformulate, then reprice — last. Tighten the recipe gently before you touch the delivery price, because on an app the price is the most visible thing you have.
Why “busy” doesn’t mean “profitable” here
A ghost kitchen can light up every delivery screen and still go under, for the same reason a packed restaurant can — volume magnifies whatever your unit economics are, and on delivery the economics are thinner and less forgiving to start. If each order clears a real margin, more orders is the whole point. If each order loses forty cents after commission, packaging, and an over-portioned protein, then a great night just means you lost money faster and told yourself a story about growth.
That’s the part that stings about delivery-only: the feedback is delayed and disguised. There’s no server telling you the kitchen’s slammed and something’s off, no dining room to read. Just tickets, and a deposit two weeks later that’s smaller than the volume implied. By the time the bank balance tells you, you’ve done it a few thousand times.
The honest catch
You can absolutely cost your menu by hand, and you should — one plate, net of commission, with packaging in, is a genuinely clarifying afternoon. Where it breaks down is the same place it breaks for everyone, only faster: prices move, portions drift, and you have a whole menu, not one item.
A ghost kitchen lives or dies on repetition. The same twenty items, hundreds of times a week, at whatever this week’s protein and packaging actually cost. A number you costed correctly in the spring is wrong by summer, and because there’s no dining room and no server, nothing tells you. Manual costing gives you a clean snapshot; it can’t keep the snapshot current against prices that change every delivery, across every item, forever.
That’s the whole reason Mise exists. We keep every item costed against the prices you’re actually paying — automatically, from your receipts — and flag the plate that drifted before a thin delivery margin turns negative. On delivery economics, catching a leak early isn’t housekeeping; it’s the difference between the deposit covering rent and not.
But you don’t need us to start. Take your best-selling delivery item, cost it net of the app’s cut, add the packaging, and see where it really lands — see what your menu actually costs →
Built by people who’ve worked the line, signed the leases, and stared at the books. We help independent food businesses know what every dish actually costs — and what to do about it.
