The one number to start with, and what it actually means
Food cost percentage is the cost of the food you used divided by the food sales you made, over the same period. The word doing the work is "used". It is not what you bought that month, because you may have bought three months of rice. It is opening stock, plus purchases, minus closing stock. That is why the number is useless without a stock count, and why a kitchen that does not count cannot tell you whether it is making money.
The second number is prime cost: food plus labour, as a percentage of sales. This matters because food cost can look respectable while wages quietly take the month. A Kenyan restaurant with a long opening day and a large floor team can run a perfectly healthy food cost and still lose money every week, and no amount of staring at the food figure will show it.
The gap worth hunting is between theoretical and actual. Theoretical cost is what your recipes say the dishes you sold should have consumed. Actual cost is what the stock count says actually left the store. If you sold 200 plates of pilau and the recipe says each takes 180g of rice, theory says 36kg left the store. If the count says 43kg left, you have found seven kilos and a question.
That question has only a few possible answers: the recipe is wrong, the portion is heavier than the recipe, something was wasted, something was given away, or something walked. Each has a different fix, which is why measuring the gap is more useful than guessing at causes. See what causes stock losses for the retail version of the same logic.
None of this requires software to start. It requires a scale, a count sheet and the discipline to use both on the same day each week. Software earns its place when the counting becomes the bottleneck, not before.
Want to see it running in your own business first?
Sign UpFinding the leak, in order of how much it usually pays
Work down this list. Each step is worth doing even if you stop there.
- 1
Count the high-value items weekly
Not everything. Meat, cooking oil, cheese, beverages and anything imported. Twenty items usually carry eighty per cent of the value, and counting twenty things weekly is a job somebody can actually keep doing, where counting four hundred monthly is a job everybody quietly stops doing.
- 2
Work out food cost percentage for one month
Opening stock, plus purchases from your invoices, minus closing stock, divided by food sales for the same dates. The answer is less important than having it: you now have a baseline, and next month tells you a direction.
- 3
Cost your ten best-selling recipes
Weigh what actually goes on the plate, not what the recipe card says. Multiply by current prices. This is tedious once and cheap forever, and it is the only way to get a theoretical cost to compare against.
- 4
Compare theory with the count
Sales mix times recipe cost gives you what should have been used. The count gives you what was used. The difference, as a percentage of sales, is your controllable loss. Under 2% is tight. Over 5% is a real problem with a real cause.
- 5
Check receiving before you blame the kitchen
Weigh deliveries against the invoice, in front of the driver, for two weeks. Short deliveries are common, rarely malicious, and almost never caught by a kitchen that signs first and counts later. More than one owner has found the entire gap here and apologised to a chef.
- 6
Then look at portioning and waste
Scales on the line for the five costliest items, and a waste log for two weeks: what went in the bin, why, and roughly how much. Both are covered in detail in reducing kitchen waste.
- 7
Add labour and look at prime cost
Food plus labour against sales, weekly. This is where you find out whether the quiet Tuesday shift has three people on it because it needs three people, or because it always has.
Where cost control goes wrong
Using purchases instead of usage
Buying 50kg of rice in a month you used 20kg makes the food cost look terrible, and the month you eat the stockpile makes it look brilliant. Neither is true. Usage needs opening and closing stock, which needs a count.
Counting everything, monthly
A full monthly count is a long evening that produces one data point too late to act on. Twenty high-value items weekly beats four hundred items monthly, every time.
Costing recipes from the recipe card
The card says 150g. The line serves what the line serves. Cost what is actually plated, or the theoretical number you compare against is fiction and the gap you find is your own recipe error.
Blaming theft first
Theft is real and worth guarding against, but short deliveries, heavy portions, trim waste and recipe drift together usually dwarf it. Starting with an accusation costs you staff and rarely finds the money.
Raising prices to fix a cost problem
If the gap is a 6% controllable loss, a price rise hands that loss a bigger base to grow on. Fix the leak, then decide whether the price is right.
Measuring once
A single food cost figure is a number. A run of twelve weekly figures is information, because it shows direction and it shows which week something changed.
A Nairobi café finds seven kilos
A café in Westlands sells a chicken and rice plate as its biggest line. The owner costs the recipe properly for the first time: 180g of rice and 140g of chicken, weighed off the line rather than read off a card. At the week’s prices that is a food cost the owner is comfortable with.
They sold 200 plates that week. Theory says 36kg of rice and 28kg of chicken should have left the store. The count says 43kg of rice and 29kg of chicken. The chicken is within a kilo, which is normal. The rice is seven kilos out, which is not.
Chicken being tight and rice being loose rules out theft, which rarely takes the cheap thing and leaves the expensive one. Two days of watching the line gives the answer: the rice is served by eye with a sufuria lid, and on a busy evening the eye is generous. A scoop costing a few hundred shillings fixes most of it.
The money involved is not dramatic in any single week. Over a year, on one dish, it is the kind of figure that funds a new fridge. The point of the exercise was never the seven kilos; it was learning that the gap can be measured at all.
In a busy kitchen, unrecorded orders, comped meals and stock leaving the back door quietly eat the margin you worked all service to earn.
Veira ties every order to the bill and the stockroom, so covers, voids and portions all reconcile at close.
When the counting becomes the bottleneck
Everything above works with a scale, a notebook and a routine, and a small kitchen should start there. What stops working is the arithmetic: once you are costing thirty recipes against a weekly count and a sales mix, the comparison takes an evening that nobody has, and the exercise quietly dies.
Veira keeps the sales mix because it is the till, holds the recipes against the menu items, and records the counts, so theoretical against actual is a report rather than an evening with a calculator. Veira for restaurants covers how the counter and the kitchen connect, and the menu pricing calculator will cost a dish for you before you commit to it.
It also files the eTIMS invoice on every sale, which matters here only because it means the sales figure you are dividing by is the same figure KRA has: eTIMS for restaurants.
Frequently asked questions
What is a good food cost percentage for a restaurant in Kenya?
How do I calculate food cost percentage?
What is prime cost and why does it matter more?
How often should a restaurant count stock?
What is the difference between theoretical and actual food cost?
Should I fix food cost or raise prices first?
Can I do this without a POS system?
Restaurant cost control is not a one-off audit; it is a weekly habit that gets faster each time. Start with twenty items and one month of food cost, then add recipe costing when the baseline stops surprising you. When the arithmetic becomes the obstacle rather than the counting, see how Veira handles restaurant stock and sales together.
Need help, or want to talk it through first? Chat with us on WhatsApp