You costed every dish. The menu works out at 28% food cost, which is where you wanted it. At the end of the month you take a stocktake, do the arithmetic, and the actual figure is 34%. On £48,000 of food sales that gap is £2,880 in one month. Nothing was stolen. Nobody did anything obviously wrong. The money left in six-pound increments over thirty days, and you have no record of any of it.
Theoretical and actual are different numbers, and the gap is the whole job
Theoretical food cost is what your recipes say the food should have cost given what you sold. Take every dish sold in the period, multiply by its costed recipe, add it up.
Actual food cost is what you really consumed. Opening stock, plus purchases, minus closing stock, gives cost of goods used. Divide by food sales excluding VAT and you have your actual percentage.
Both are needed. Theoretical alone tells you the menu is well designed and nothing about the operation. Actual alone tells you there is a problem and nothing about where it is. The variance between them is the number to manage, and it should be calculated every month at minimum, weekly in a business over a certain size.
A variance of a point or two is normal. Five or six points is an operational failure with a specific cause, and the causes are a short list: portioning, waste, theft, incorrect recipes, purchase price movement not reflected in the costings, and staff food.
The stocktake is the measurement, so it has to be done properly
An inaccurate stocktake produces an inaccurate variance, which sends you looking for a problem that is not there.
Count on the same day of the week, at the same point in the delivery cycle, ideally after service and before the first delivery. Count everything, including the walk-in, the dry store, the bar back-up and the things in the cellar that nobody thinks of as stock. Use the same unit of measure and the same prices each time, updated from actual invoices rather than from a price list agreed in the spring. Two people counting is slower and considerably more accurate than one.
Invoices have to reconcile too. A delivery signed for without checking is a delivery you have paid for and may not have received. Short deliveries, substituted items at a higher price, and a weight on the docket that does not match the weight in the box are all routine, and none of them are visible a week later. Weighing deliveries of expensive proteins is five minutes that pays for itself.
Portion control is the single largest source of drift
A recipe says 180g of protein. Under pressure at half past eight, without a scale, a chef puts on what looks right. Across a service that averages 195g, and 15g on 40 covers is 600g a night, which at £18 a kilo is £10.80, which is £3,900 a year from one dish.
Portion control is therefore not pedantry, it is the main mechanism. Scales on the section, portioned and labelled during prep rather than at service, standard ladles and scoops, marked glassware for wine, and a photograph of the finished plate in the kitchen so that everyone is building the same dish. Spot check by weighing a plate at random once a shift.
The same discipline applies in reverse. A portion that is too small generates complaints and comps, and a comped dish is 100% food cost with no revenue against it. Track comps and voids as a line in their own right and look at who is authorising them.
Yield is the other half of the arithmetic. Costing a whole fish, a side of beef or a crate of leaves at its invoice price understates the true cost per portion, sometimes by half. Do a yield test on anything you break down yourself: weigh it in, weigh the usable product out, and cost the recipe from the yielded figure. This is where a menu that looks like it works on paper turns out not to.
Waste is only manageable if it is written down
Nobody wants to record waste. It is an admission, it takes time, and the bin is right there. Which is exactly why the waste log is the highest-value piece of paper in the kitchen.
Log every item that leaves the kitchen without being sold, with the reason. Over a fortnight a pattern emerges that no amount of intuition would have produced:
- Prep waste from over-production on a quiet day, which is a forecasting problem.
- Spoilage from over-ordering against minimum order values, which is a purchasing problem.
- Cooking errors and remakes, which is a training or equipment problem.
- Returns and comps, which is a consistency problem.
- Date expiry in the walk-in, which is a stock rotation problem.
Each has a different fix and they are routinely treated as one problem called waste. Add a physical stock rotation discipline behind it: date labelling, first in first out, and a weekly walk-in check where somebody moves things forward rather than around.
There is now a regulatory reason to look at this as well. Under the Simpler Recycling reforms in England, most businesses have been required to separate food waste for collection, with smaller businesses phased in later. Separated food waste is weighed by the collector, which means for the first time many kitchens have an external measurement of how much they throw away.
Staff food should be a policy, not a habit. A defined staff meal, costed and prepared from a planned list, is a legitimate and worthwhile cost. Unlimited grazing from the section is an unmeasured one, and it appears in your variance as a mystery.
Supplier price rises eat a month before you notice
A menu costed in January is a historical document by June. Proteins, dairy, oils and anything energy intensive have all moved substantially in recent years, and suppliers pass changes on by email or simply on the invoice.
The defence is a short watch list. Pick the ten ingredients that make up the majority of your spend, record their unit price from actual invoices every month, and recost the affected dishes whenever one moves more than a few percent. A rise on an ingredient in your bestselling dish is worth more attention than a rise on something you sell twice a week.
When you need to respond, you have four options and only one of them is raising prices: reformulate the dish, change the supplier or specification, reduce the portion of the expensive component while improving something cheaper, or reprice. Doing nothing is a fifth option and it is the one most kitchens choose by default, which is how a 28% menu becomes a 34% reality.
Where this shows up in the game
Restaurants are part of Business Baron's full game, and a café or a pub is free to run. The game calculates gross margin from actual stock movements rather than list prices, so waste shows up where it belongs. The gap between what the recipes say and what the bin says is a gap in your margin, and that is where you will find it.