Food cost percentage is cost of sales divided by food revenue, multiplied by 100. The part that trips most operations up is the numerator: cost of sales is not what you bought in the period. It is opening stock plus purchases minus closing stock, adjusted for transfers, staff meals and complimentary covers.
Food cost % = Cost of sales ÷ Food revenue × 100
Dividing purchases by sales is the single most common error in F&B control. In any period where stock levels move — and they always move — it produces a number that is confidently wrong. A big delivery on the last day of the month makes a good month look terrible; running the freezers down makes a bad month look excellent.
On this page
The calculation, step by step
- Opening stock. The valuation of everything in your stores, kitchens and outlets at the start of the period — which must equal the previous period's closing stock exactly.
- Purchases. Everything received in the period, valued at invoice cost excluding recoverable tax. Received, not ordered, and not paid.
- Closing stock. The valuation at the end of the period, counted on the same cut-off as the sales figures.
- Transfers. Stock moved between outlets, departments or bars — out of one cost centre, into another.
- Deductions. Staff meals, complimentary covers and marketing entertainment, valued at cost and moved to the line where they belong.
- Revenue. Net food sales for the same period, excluding sales tax or VAT, so cost and revenue are measured on the same basis.
Cost and revenue must be on the same tax basis. Costing stock inclusive of tax while reporting revenue net of it is a quiet, systematic error that overstates food cost by several points and survives for years because nobody re-checks the basis.
A worked example
A single restaurant outlet, one month. Figures are illustrative and shown without currency, so they apply anywhere.
| Line | Amount |
|---|---|
| Opening stock | 42,000 |
| Purchases received | 168,000 |
| Closing stock | (38,000) |
| Transfers out to bar | (4,500) |
| Transfers in from banqueting | 2,000 |
| Staff meals at cost | (9,000) |
| Complimentary covers at cost | (2,500) |
| Cost of sales | 158,000 |
| Net food revenue | 480,000 |
| Food cost percentage | 32.9% |
| Gross profit | 322,000 (67.1%) |
Note what the naive calculation would have produced: 168,000 ÷ 480,000 = 35.0%. Same month, same trading, a two-point error — and two points on this revenue is roughly 10,000 of misreported profit.
Transfers, staff meals and complimentary
These three adjustments separate a control system from a spreadsheet.
Transfers. When the kitchen sends lemons and cream to the bar, that cost belongs to the bar. Untracked transfers make the kitchen look wasteful and the bar look brilliant, and they make outlet-level accountability impossible. Every transfer needs a docket, valued at cost.
Staff meals. Consumed, not sold. Leaving them in cost of sales inflates the percentage and hides the true trading position. Value at cost and transfer to payroll or employee benefits — where, incidentally, they are also easier to manage as a per-head budget.
Complimentary and entertainment. A comped table for a VIP or a journalist is a marketing cost, not a kitchen failure. Move it to the marketing or guest-recovery line so both numbers stay honest.
Theoretical versus actual: the real number
The percentage alone tells you where you landed. It does not tell you why, and it cannot tell you what is recoverable. For that you need both numbers:
- Theoretical cost — what the food should have cost: every recipe costed, multiplied by the number of each dish sold in the period.
- Actual cost — what it did cost, from the stock calculation above.
The gap is the variance, and the variance is the whole point. A restaurant running 32% actual against a 28% theoretical does not have a 32% problem — it has a four-point leak, and every one of those points is recoverable without changing a single menu price.
This is why recipe costing is not administrative overhead. Without a theoretical cost you have a number, but no diagnosis.
Where the variance comes from
| Source | What it looks like | Where to check |
|---|---|---|
| Yield | Recipes costed on raw weight when the usable yield is lower after trimming and cooking | Yield test the ten highest-volume proteins |
| Portioning | Plates served heavier than the recipe, inconsistently between shifts | Scales at the pass; spot-weigh during service |
| Waste & spoilage | Over-production, poor rotation, expired stock | Waste log; FIFO discipline; par level review |
| Receiving | Short deliveries signed for, price increases accepted silently, wrong specification | Weigh and count at the dock; check invoice against order and quoted price |
| Pricing drift | Supplier increases never reflected in recipe cost or menu price | Re-cost the menu quarterly, not annually |
| Shrinkage | Unexplained loss after every other cause is eliminated | Store access control, requisition discipline, count integrity |
| Sales mix | Percentage moves without anything going wrong — high-cost dishes simply sold more | Mix analysis before assuming a control failure |
Work them in that order. Yield and portioning explain more variance in most kitchens than theft does, and they are far cheaper to fix.
Benchmarks
Benchmarks orient you; they do not set your target. A steakhouse and a pizzeria cannot run the same percentage and should not try.
| Context | Commonly cited range |
|---|---|
| Restaurants, food cost | 28–35%, with the full-service average around 32% |
| Beverage cost | Typically 15–25%, with total beverage programmes often 18–24% |
| Upscale hotel F&B | A common starting point is around 35% food and 25% beverage; at a 70/30 sales mix this blends to roughly 32% overall |
| Prime cost (food plus labour) | Around 55–65% of revenue |
Remote operations sit differently again. A bush lodge carrying freight, spoilage risk and a fortnightly resupply cycle cannot be judged against a city restaurant with daily deliveries — the logistics are part of the cost structure, not evidence of poor control.
Period discipline
Most "unexplainable" swings are period errors, not control failures:
- Count on the same cut-off as sales. If the stock take is Sunday morning and the sales period closed Saturday night, Sunday's trading is in one number and not the other.
- Freeze receiving during the count. Deliveries arriving mid-count are either counted twice or missed entirely.
- Count in the base unit. Half-cases, opened bottles and part-used containers need a consistent convention, applied by everyone.
- Two people count. One counts, one records. Self-verified counts are the least reliable number in the building.
- Cut-off invoices properly. Goods received on the last day but invoiced next month still belong in this period's cost.
- Never adjust a count to make the percentage look right. It converts a one-month problem into a permanent one, and it will surface at year-end audit.
Percentage is not profit
Percentage is a ratio, and you cannot bank a ratio. A dish at 25% cost selling at 40 contributes 30 in cash. A dish at 38% cost selling at 120 contributes 74. The "worse" percentage is more than twice the money.
Read percentage and contribution margin together. Driving percentage down in isolation pushes a menu toward cheap, low-cash items and can reduce total gross profit while the headline number improves — which is why the F&B report should always show cash gross profit alongside the percentage.
Common errors
- Purchases ÷ sales. The classic. Ignores stock movement entirely.
- Mismatched tax basis between cost and revenue.
- Staff meals left in cost of sales, permanently inflating the percentage.
- Transfers untracked, making outlet accountability impossible.
- Recipes costed once and never updated as supplier prices move.
- No theoretical cost, so variance can never be diagnosed.
- Chasing the percentage instead of the cash margin.
- Reacting to one month. Read the trend across three periods before changing anything structural.
Frequently asked questions
How do you calculate food cost percentage?
Cost of sales ÷ food revenue × 100, where cost of sales is opening stock + purchases − closing stock, adjusted for transfers, staff meals and complimentary covers.
What is a good food cost percentage?
Restaurants commonly benchmark at 28–35% with a full-service average near 32%. Upscale hotels often start from roughly 35% food and 25% beverage, blending to about 32% at a 70/30 mix. Your correct target depends on your menu, mix and market.
What is the difference between theoretical and actual food cost?
Theoretical is what the food should have cost from recipes and dishes sold; actual is what it did cost from the stock calculation. The variance between them is where recoverable money sits.
Should staff meals be included in food cost?
No. Value them at cost and transfer them to payroll or employee benefits. Leaving them in cost of sales inflates the percentage and obscures trading performance.
Why is my food cost percentage different every month?
Usually period discipline — stock takes not aligned to the sales cut-off, deliveries counted or invoiced in the wrong period, untracked transfers — or a genuine shift in sales mix. Fix the discipline before diagnosing a control problem.