Run the kitchen on numbers, not on memory
Recipe specifications that stay true, menu costing that includes yield and waste, and training that changes what staff actually do at 8pm on a Saturday.
Where margin is lost
The four leaks every kitchen has
None of these show up as a single large error. They accumulate quietly until the food cost is simply higher than it should be.
Uncontrolled portions
The single largest variable in food cost. An uncontrolled portion of a £2.80 ingredient varies easily by a factor of two, and nobody notices until the month-end number is wrong.
Ignored yield
A whole fish yields 30–40% usable. A cauliflower yields 40–45%. Costing on purchase price rather than usable yield overstates margin and hides the real dish cost.
Specifications that drift
A recipe that only exists in the head of the person who wrote it is cooked differently every shift. That variance is cost, and it compounds.
Waste nobody weighs
Trim, spoilage, remakes and over-portioning. You cannot control what you do not weigh, and a weekly log by category is the highest-return hour of admin in a kitchen.
Turnover you cannot see on payroll
Two experienced chefs leaving a quarter costs recruitment, weeks of reduced productivity and a higher error rate — none of which appears as a labour line.
Menus nobody removes
A dish at 45% food cost that 2% of guests order is a liability, not an asset. Removing the bottom of a menu usually raises overall profitability.
The cycle
How a menu stays profitable
Not a one-off project. Prices, mixes and preferences all move, so this is a quarterly loop rather than an annual audit.
-
1
Cost the top 20 honestly
By sales volume and by revenue margin, with usable yield and measured portions. Not purchase price — usable cost.
-
2
Classify each dish
Popular and profitable, popular and not, or neither. Volume disguises the problem with popular-but-thin dishes.
-
3
Act on one quadrant at a time
Re-portion a plowhorse. Reposition a puzzle. Re-spec a dog. One change at a time, then re-cost after thirty days.
The manual trap
Why costing keeps being wrong
It is not that anyone is careless. A monthly manual process cannot catch every change as it happens, so the costed price always describes last month.
Popularity measured in revenue
A £60 dish sold twenty times is not popular, it is high-value. Measure in item mix or you will conclude your most popular items are your most expensive ones.
Profitability read backwards
Lower food cost percentage is better. Reversing this inverts every conclusion and produces the opposite advice.
A four-week window
Seasonal variance produces permanent misclassification, and a dish that is merely seasonal looks like a failure.
Supplier changes nobody told costing about
A substitution on a delivery note changes the cost of a dish without changing a single document on file.
Put this into practice
Upload a menu you already use. Review the allergen matrix MenuOps generates from it — that is the fastest way to know whether it is worth your time.