Food cost control: the eleven leaks that show up on every audit
Your theoretical food cost is what your recipes say you should be spending. Your actual food cost is what your invoices say you did spend. The gap between them is not a rounding error, and on most sites it is worth more than any price rise you are considering.
About this guide to food cost control
Most operators can tell you their food cost percentage on request. Far fewer track their theoretical food cost alongside it, and fewer still watch the difference between the two. That difference is the single most useful number in the kitchen, because it is pure leakage. It is money you have already spent that never became a sale.
On a site turning over £900,000 a year on food, a three point gap between theoretical and actual is around £27,000. That is not a pricing problem and it will not be fixed by putting the burger up a pound. It is a control problem, and control problems have specific, findable causes.
Here are the eleven that show up most often, in roughly the order I would attack them.
First, measure the gap
Theoretical food cost is the sum of every dish's plate cost multiplied by the units sold in the period. Actual food cost is opening stock plus purchases minus closing stock. Divide each by net food sales. If the two are within one point of each other, your control is genuinely good. Two to three points is normal. Four or more and there is something structural going on.
1. Specifications that no longer describe what the kitchen cooks
This is the most common single cause and the least dramatic. A spec is written at launch, the dish evolves over eighteen months, and nobody updates the document. The chef de partie who developed the current version left last year. The spec says 180g, the section plates 210g, and every costing built on that spec is understating plate cost by roughly 17 percent on the protein.
Fixing it is dull and it works. Weigh the plate, three times, on three different days, cooked by three different people. Then update the spec to whatever you actually want it to be, and retrain to that.
2. Portion drift on the section
Related, but distinct. Specs can be current and portions can still drift, because portioning by eye under pressure is not a skill, it is a variable. A section that is 8 percent generous on chips across four thousand portions a month is a meaningful number, and nobody in the building will notice it happening.
Portion specs are also a labour decision: a plate with four separately prepped components drifts in four places rather than one. What the menu costs in time covers that side of it.
Scales on the section, portion pots for anything loose, and pre portioned proteins where the volume justifies the prep time. The kitchens with the tightest food cost are almost always the ones where the least is left to judgement at eight o'clock.
3. Yield loss costed at purchase weight
A £22 per kilo turbot is not £22 per usable kilo. Once it is filleted and trimmed you may be looking at £38 per usable kilo, and if your costing uses the invoice price you have understated that dish by 40 percent.
Run a yield test on every whole item you buy: weigh in, butcher, weigh the usable portion, and calculate the real cost per usable kilo. Do it once per item and record it. The recipe costing calculator has a wastage and yield field for exactly this.
| Item | Purchase price | Typical yield | True cost per usable kg |
|---|---|---|---|
| Whole turbot | £22.00 | 42% | £52.38 |
| Beef sirloin, whole | £18.50 | 78% | £23.72 |
| Whole chicken | £4.20 | 65% | £6.46 |
| Leeks, untrimmed | £1.90 | 55% | £3.45 |
Illustrative yields. Run your own tests, because they vary by supplier, season and butcher.
4. Waste that never gets written down
If your waste sheet shows less than one percent of purchases, the sheet is wrong. Waste that goes unrecorded is waste that cannot be managed, and the categories matter more than the total. Prep waste is a skills or specification issue. Spoilage is an ordering or rotation issue. Returns are a quality or consistency issue. Overproduction is a forecasting issue. Four different problems with four different fixes, and one number on a sheet tells you nothing about which one you have.
Separate bins, thirty seconds of recording, reviewed weekly. It is the highest return administrative task in a kitchen.
5. Staff food with no boundary
Feeding your team properly is not negotiable and it is not the problem. The problem is staff food that comes off the same shelf as service stock with no record, at which point it lands in your food cost as an unexplained variance rather than as a known cost.
Set a per head allowance, cook staff food from a defined list of ingredients, record it, and budget for it as a line. It stops being leakage the moment it becomes a number.
6. Comps, remakes and staff discount
Every comped dish costs you its full plate cost and returns nothing. Most sites do not track them by reason, which means they cannot see that four percent of covers are being remade because one dish has a consistency problem.
Record the reason code on every comp. If your EPOS will not do it, a clipboard by the pass will. Within a month you will know whether your comps are a service problem, a kitchen problem or a single dish that needs rebuilding.
7. Invoice prices that moved without anyone noticing
Suppliers put prices up. That is not a scandal. The problem is that on most sites the invoice is checked against the delivery, not against the previous price, so a 14 percent increase on rapeseed oil goes through unremarked and stays in the cost base permanently.
Track a basket of your top twenty lines by spend. Check them monthly against the previous month. Anything moving more than five percent gets a conversation with the supplier or a substitution. Twenty lines will typically cover 70 to 80 percent of your food spend.
Whether the answer is absorbing the rise or repricing the dish is a separate decision, and the pricing guide sets out how to work it out.
8. Deliveries checked in badly
Short deliveries, substituted grades, out of spec produce accepted because the person receiving it is also plating starters. Every one of those is money you paid for and did not receive.
One named person receiving, weighing anything sold by weight, and rejecting out of spec produce at the door rather than sighing about it later. The first time you reject a delivery, your supplier recalibrates what they send you. That effect is worth more than the individual rejection.
9. Stock counted inconsistently
If your stock count moves between a Sunday night and a Monday morning, or between two different people with two different ideas of what counts as a full container, your food cost figure will bounce around by points for reasons that have nothing to do with the kitchen.
Same day, same time, same person, same method, same sheet order. Consistency matters more than precision here, because you are looking for movement over time rather than an absolute truth.
10. Menu mix moving underneath you
This one is not leakage, and it catches out even well run kitchens. Your food cost can rise several points without a single price changing and without a gram of waste, simply because customers shifted toward your lower margin dishes.
That is a menu engineering problem rather than a control problem, and it is invisible unless you are tracking sales mix alongside cost. If your actual and theoretical food costs are in line but your GP has still fallen, this is almost always the reason.
11. Costing that is simply out of date
The last leak is the one that makes all the others invisible. A costing model built in January and consulted in October is not a model, it is a historical document. Every decision made against it, every price set from it, every dish kept or cut because of it, is being made on stale information.
Refresh volatile lines monthly. Refresh everything quarterly. It takes a couple of hours if the structure is already in place.
The order to do this in
If you attempt all eleven at once you will get through none of them. The sequence that works:
- Weeks 1 to 2. Measure the gap. Get theoretical and actual food cost onto the same page so you know the size of the problem.
- Weeks 3 to 4. Specifications and yields. This is the foundation, and nothing else can be trusted until it is right.
- Weeks 5 to 6. Waste recording and comp reason codes. Cheap, fast, and they tell you where to look next.
- Weeks 7 to 8. Purchasing controls. Top twenty basket, monthly price check, disciplined goods in.
- Ongoing. Monthly costing refresh and quarterly menu engineering analysis.
Setting the expectation honestly
How much this recovers depends entirely on where you are starting from, and the honest answer is that nobody can tell you the number before measuring the gap. What is predictable is the direction: every point of leakage you close is a point of gross profit you keep, and none of it requires you to put a price up or change a dish. Measure your own theoretical against actual first, and the size of the prize will be obvious.
How this page is sourced
The eleven causes below come from costing work on live menus rather than from a survey, and they are ordered by how often they turn out to be the expensive one. The yield percentages in the table are illustrative starting points, not measurements from your supply chain, which is why the page tells you to run your own tests.
No recovery figure is promised anywhere on this page, because the size of the gap is not knowable before it is measured. If you spot an error, tell me and it will be corrected with the review date updated.
Related
Recipe costing calculator, for building plate costs with yield loss included.
Food cost calculator, for food cost percentage, GP and matrix classification.
How to calculate food cost, for the method behind the numbers on this page.
The menu engineering guide, for the mix side of the problem.
Increasing restaurant sales, for growing the top line without giving the margin back.
Increasing spend per cover, for earning more from every guest you already have.
The same leaks section by section: menu profitability by section.
Want the gap measured properly?
A menu audit puts a number on every one of these, ranked by what each is costing you a year.