Ask an owner what their food cost is and you usually get a confident number. Ask when they last calculated it and the answer is often "when we set the menu."
That number came off a recipe sheet. It assumes a chef who portions to the gram, a kitchen that wastes nothing, no dish ever sent back, and no staff eating. Your actual food cost — the money that left your store divided by the money that came in — is almost always higher, and it is the only one your bank balance responds to.
The good news is that it takes about twenty minutes a month to get right.
Why the number matters more than any other single figure
Food cost is the largest controllable expense in most restaurants. Rent is fixed for years. Salaries move in steps. Food moves every single day, in both directions, and nobody notices until the quarter closes.
Put a number on it. A restaurant doing ₹10,00,000 in monthly food sales at 34% food cost spends ₹3,40,000 on ingredients. Pull that to 32% and you keep ₹20,000 more, every month, ₹2,40,000 a year — without one extra cover, one extra rupee of marketing spend, or one more hour of anyone's time.
Two percentage points is not an ambitious target. It is usually portioning on three dishes and one supplier conversation. Where that saving lands in the business as a whole is set out in restaurant profit margins.
The reverse is what catches people. Food cost drifting from 30% to 34% over eight months does not announce itself. Sales look fine. The floor is busy. The bank balance just stops growing, and by the time it is obvious you have lost a season's profit.
The formula
For a period — a month is the standard unit:
Food cost % = (Opening stock + Purchases − Closing stock) ÷ Food sales × 100
The top half of that is cost of goods sold: what you actually consumed, not what you bought. A ₹60,000 rice order on the 28th of the month is not a cost this month; most of it is still sitting in your store.
That is the whole reason the stock count exists, and the reason "purchases ÷ sales" — which is what most people do — bounces around uselessly month to month.
A worked example
A 45-cover restaurant in Pune, for the month of June:
| Line | Amount |
|---|---|
| Opening stock (1 June, counted) | ₹1,85,000 |
| Purchases during June | ₹3,20,000 |
| Closing stock (30 June, counted) | ₹1,72,000 |
| Cost of goods sold | ₹3,33,000 |
| Food sales (excluding GST, excluding beverages) | ₹9,80,000 |
| Food cost % | 33.9% |
Two details do most of the damage when they are wrong:
- Sales must exclude GST. Your cost figures are pre-tax, so your sales figure has to be too. Using the gross billed amount flatters your food cost by roughly three percentage points and makes the whole exercise pointless.
- Keep food and beverage separate. Beverage cost runs far lower than food cost. Blend them and you get an average that hides a food problem behind a healthy bar.
What is normal
| Format | Typical working range |
|---|---|
| Full-service restaurant | 28–35% |
| Café, bar or pub (drinks-led) | 22–30% on food, well under 30% overall |
| QSR and value formats | 30–38% |
| Buffet and thali | 35–42% |
| Fine dining | 30–38% |
Where these ranges come from. They are conventional operating ranges used across the industry, not findings from a published survey of Indian restaurants. Treat them as a sanity check on your own number rather than a target to hit, and take anything here touching tax treatment — what to include in sales, how GST affects your figures — to your accountant. This is general information, not financial or tax advice. Last reviewed: 31 July 2026.
These are working ranges, not laws. A buffet at 40% can be more profitable than a fine-dining room at 30% if the volume and the labour cost say so — which is why food cost belongs next to your other restaurant KPIs rather than on its own.
The comparison that actually earns money is against yourself. 33% this month against 31% in April is a real signal. 33% against a number from a US industry report is trivia.
Theoretical versus actual: where the money leaks
Two figures are worth having.
Theoretical food cost is what your dishes should cost: every recipe costed out, multiplied by how many of each you sold. Actual food cost is the stock-count number above.
The gap between them is your variance, and it is the most useful diagnostic in the kitchen. Theoretical 30%, actual 34% means four percentage points — about ₹40,000 a month at ₹10 lakh of sales — walked out of the building unsold.
Under two points is normal operating friction. Four or more points has a cause, and it is nearly always one of these five:
- Portioning drift. Nobody weighs anything after the first month. A biryani ladled 15% heavy on every plate is invisible on the floor and enormous at the till.
- Wastage nobody records. Prep trim, burnt batches, dishes returned, spoilage from a fridge that ran warm. If it is not written down at the moment it happens, it is not going to be remembered.
- Supplier price rises absorbed silently. Paneer went up ₹40 a kilo in March and nobody told the kitchen or the menu.
- Yield assumed rather than measured. Your recipe says 200g of chicken. You bought it bone-in. The costing is wrong by whatever the bone weighs.
- Theft. Real, less common than owners fear, and usually visible as a variance on exactly one category rather than across the board.
The order matters. Work through one to four before you start suspecting people — in most kitchens they account for the whole gap.
Five fixes worth doing this month
1. Count stock on the same day, every month
The last day of the month, after close, two people, one counting and one writing. It takes forty minutes in a mid-size kitchen once you have a sheet in a fixed order that follows how the store is physically laid out.
Inconsistent counts are worse than no counts, because they generate a number you will act on.
2. Cost your top ten dishes properly
Not the whole menu — the ten that make most of your sales. Cost every component including oil, spices and garnish, and use bought weight, not usable weight.
Most owners find at least one dish selling well below the margin they assumed, and it is very often a signature item.
3. Weigh the three biggest proteins for one week
Put a scale on the pass and portion to it for seven days. You are not doing this forever; you are finding out how far reality has drifted from the recipe. The chef who says portions are exact is usually right about their own and wrong about the other four people on the line.
4. Start a wastage log
A clipboard by the bin. Item, rough quantity, reason, initials. No blame attached in month one — the goal is data, and punishing the first honest entries guarantees an empty sheet by week two.
A month of that tells you more than any software will.
5. Re-quote your top five ingredients every quarter
Not to switch suppliers necessarily. Knowing the market rate is what keeps the price you already pay honest.
A weekly version for the ten dishes that matter
The monthly count tells you where you ended up. A weekly check on your highest-volume dishes tells you where you are heading.
Take the ten dishes that make the biggest share of your sales, know their costed price, and once a week compare units sold against the ingredient volume that should have supported them. It is rough, it takes fifteen minutes, and it catches a ₹40-a-kilo price rise in week one instead of month three.
Summary
- Food cost % = (opening stock + purchases − closing stock) ÷ food sales × 100, with sales taken excluding GST and beverages held separately.
- 28–35% is the usual working range for an Indian full-service restaurant, but your own trend beats any benchmark.
- The gap between your recipe costing and your stock-count number is where the money actually goes — portioning, wastage, silent price rises and yield assumptions, roughly in that order.
- Count stock the same day every month, with the same sheet, or the number cannot be trusted.
- Two percentage points on ₹10 lakh of monthly food sales is ₹2.4 lakh a year, earned without a single extra cover.
