How to reduce food cost in a café without changing the menu
Most cafés know their food cost as a single percentage of revenue. That number hides everything worth acting on.
Ask a café owner their food cost and you usually get one number — thirty per cent, thirty-five, sometimes "about a third". It is a real number and it is nearly useless, because it is an average across dishes with wildly different economics. A café at 32% overall can easily contain a signature item at 55% that sells forty a day, and a drink at 12% that sells four.
Reducing food cost is mostly a measurement problem before it is a purchasing problem. Here is the sequence that works, in order.
Step 1: Cost your ten best-selling dishes properly
Not the whole menu. Ten. In most cafés the top ten items are the large majority of covers, and costing them accurately gets you nearly all of the benefit of costing everything.
For each one, write the recipe as quantities, not descriptions: 180ml milk, 18g coffee, 1 paper cup, 1 lid, 1 sleeve. Then price each line from your last purchase invoice, not from memory. Include:
- Packaging, which is routinely forgotten and is material on takeaway-heavy menus.
- Anything given away with the dish — sauces, a side salad, a mint.
- GST paid on ingredients if you are on the concessional restaurant rate without input tax credit, because that tax is a real cost to you and does not come back.
- A realistic wastage allowance on anything perishable.
That last one matters. A dish using an ingredient you throw away a fifth of costs 25% more than its recipe suggests.
Step 2: Put cost next to sales volume
Now cross the cost figures with how often each item sells. Four quadrants, and each one has a different action:
| Sells a lot | Sells rarely | |
|---|---|---|
| High margin | Protect it. Do not touch the recipe or the price. | Promote it. Move it up the menu, suggest it at the counter. |
| Low margin | Fix it — this is where the money is. Reprice, resize or re-source. | Cut it. It costs prep, stock and menu space for nothing. |
The high-volume low-margin box is where nearly all recoverable money sits. A ₹12 improvement on an item selling fifty a day is ₹18,000 a month. The same ₹12 on an item selling twice a day is ₹720 and not worth the effort.
Step 3: Fix the expensive items in the right order
Four levers, roughly in order of how well they hold up:
- Portion discipline. The cheapest fix and the most reliable. Weigh the protein and the cheese for a week. Most kitchens are over-portioning by ten to twenty per cent without knowing it, and a scale on the line costs almost nothing.
- Re-sourcing. Same specification, different supplier or pack size. Worth doing on your top three ingredients by spend, not on everything.
- Recipe change. A cheaper component or a smaller quantity of an expensive one. Effective, but it changes the dish — test it before rolling it out.
- Price increase. Works, and guests notice. Save it for items where the cost has genuinely moved, and change several prices at once rather than nudging one item repeatedly.
Step 4: Close the gap between theoretical and actual
Once recipes are costed, your system can calculate what you *should* have used: sales multiplied by recipe. Compare that with what you actually bought and counted, and the difference is waste, theft, over-portioning or miscounted stock. Chase the biggest gap, not all of them.
This is the single most valuable number in food cost control and almost nobody small tracks it, because it needs recipes attached to menu items and stock that moves when you sell. That is exactly what recipe-linked inventory does — selling a dish deducts its components, so the theoretical figure maintains itself. KhaoPiyo does this on its inventory module; the principle applies whatever you use, including a spreadsheet.
Step 5: Attack waste where it actually happens
In a small café, waste concentrates in a few predictable places:
- Prep for a busier day than you got. Prep against last week's same weekday, not against a hopeful average.
- Perishables ordered in supplier pack sizes. If a case is more than you use before it turns, split the order or change the item.
- Remakes. Every returned dish is double cost. If one dish gets remade often, the problem is a recipe or a station, not the guest.
- Staff meals with no rules. Legitimate and worth having — but they should be a defined meal, not open access to the line.
- Cancelled orders after prep started. Track these. If the number is high, the problem is usually order timing, not customers changing their minds.
What good looks like
You do not need a target percentage — those vary enormously by format and are more useful for comparing yourself to yourself than to anyone else. What you need is: recipes costed for your top items, per-item margin visible without doing maths, the gap between theoretical and actual usage tracked monthly, and one specific item being worked on at any given time.
A café that does those four things will find several per cent of margin in the first quarter, almost always from two or three dishes nobody suspected.