Stock count accuracy: six ways a count goes wrong

A stock count, or stocktaking as it is often called, looks like the most objective number a kitchen produces. Someone walks the shelves, writes down what is there, multiplies by the prices and adds it up. The result goes straight into the food cost, and nobody questions it, because it was counted. Yet a large share of the strange food cost figures kitchens chase every month come from the count itself, not from anything that happened to the food.

Stock count accuracy fails in predictable ways. A case is recorded as one item. A half-used tub is guessed. A fridge is forgotten. A tray of sauce is not counted because it is not on the sheet. A delivery arrives halfway through. And the whole thing is done at midnight by one person who wants to go home. Each of these moves the result without a single portion having been wasted.

This guide sets out the six most common ways a restaurant stock count goes wrong, how much each can move the number, and how to prevent each one. Every figure in it is an example.

The stock count feeds the cost of goods sold calculation directly: opening stock plus purchases minus closing stock. Any error in the closing count is an error of the same size in the month's food cost, and it reverses next month, when that count becomes the opening stock.

An example. A kitchen with food sales of 40,000 a month overstates its closing stock by 400: a few cases counted twice, a tub of something expensive guessed generously. Its food cost for the month looks a full point better than it was. Next month, the inflated figure becomes opening stock, and the kitchen looks a point worse. Two months of misleading numbers, from one sloppy count. Our guide to cost of goods sold explains the calculation, and our guide to reading food cost variance explains why the count is always the first thing to check.

The most common error is in the stock units. A supplier sells cream in cases of six litres; the count sheet lists cream by the litre; the counter writes "2" meaning two cases. Twelve litres become two. Or the reverse: tins counted individually against a sheet priced per case.

Units cause the largest single errors, because they multiply. A unit mistake on an expensive line can move a count by more than all the other errors together.

Prevent it with these habits:

Kitchens are full of things that are partly used: open bags of flour, half tubs of crème fraîche, a third of a tin of olive oil, a bottle of vinegar with a little left. Each has to be recorded as a number, and the temptation is to guess, or to round to a whole unit.

Guesses on cheap items do little harm. Guesses on expensive ones do a lot. A tub of a premium ingredient called "half" when it is a quarter, repeated across a few items, is a meaningful overstatement.

Weigh part-used containers of anything expensive. A scale reading to the gram, a note of each container's empty weight, and a simple subtraction give a far better figure than an estimate. For cheaper items, a consistent rule, such as rounding down to the nearest quarter, keeps the errors small and in one direction. Our guide to liquor inventory shows the same weighing approach for open bottles.

Stock lives in more places than the count sheet remembers. The walk-in and the dry store are counted; the prep fridge on the line, the freezer at the back, the shelf above the pass, the bar's fridge of lemons and cream, the function room's store, stock waiting to be returned and the delivery left by the back door are often not.

Each missed location understates closing stock, making the month look more expensive than it was. A forgotten freezer drawer of portioned fish can be worth more than everything else that went wrong in the count.

Prevent it by listing every location on the count sheet, in the order the counter walks, and counting every location every time, even if it is usually empty. Write "nil" against an empty location rather than leaving it blank, so you know it was checked.

Sauces, stocks, dressings, marinated proteins, portioned fish, doughs and other prepared food sit in the fridges at every count. They contain ingredients that were bought and are still in the building. If they are not on the count sheet, they are not counted, and the ingredients in them are treated as used.

In a kitchen that preps heavily, prepared food can be a large share of what is in the fridges. Leaving it out understates stock significantly, and doing so in one month but not the next creates a swing with no cause in the kitchen.

Add prepared items to the count sheet, valued at the cost of their ingredients from their recipe costing. A tray of braised meat is counted by portion or by weight, and valued at what its portions cost to make. Our guide to recipe costing shows how to cost a batch.

A count is a snapshot, and it is only accurate if nothing moves while it is taken. In practice, deliveries arrive, the bar borrows cream, the kitchen moves stock from the store to the line, and service sometimes continues. Each movement during the count risks counting the same stock twice or not at all.

Timing also matters between counts. A count taken on a Sunday night and the next taken on a Monday morning cover different amounts of trading, and a delivery received the morning after one count but entered before it puts stock in the wrong period.

Prevent it with a few rules:

The last common failure is the human one. The count is done at the end of a long shift, by one person, in a hurry, with nobody checking. Mistakes of every kind above become more likely, and nobody catches them.

A count finished at midnight by a tired chef who wants to go home will skip the awkward shelf, guess the half tubs, round in whatever direction is quickest and miss the freezer at the back. None of this is laziness; it is what happens when an important task is given the worst time slot of the week.

Prevent it by giving the count a proper slot: a quieter time, ideally two people, and enough time to do it carefully. Two people, one counting and calling out, the other writing, are quicker and much more accurate than one, and they catch each other's mistakes.

A count can be perfect and its value still wrong, if the prices on the sheet are out of date or in the wrong unit. A case price applied to a single item, last year's price on a line whose cost has risen, or a price for a different pack size all distort the stock value.

Keep the count sheet's prices current, from the latest checked invoices, and in the same unit as the count. A count valued at stale prices understates or overstates stock in a way that is easy to miss, because every number on the sheet looks reasonable.

To see how the errors combine, take an invented kitchen whose closing stock is worth about 6,000 and whose food sales are 40,000 a month. One month's count contains a typical mixture of mistakes:

The errors partly cancel, but not much. Closing stock comes out about 360 too low, which makes the month's food cost nearly a point worse than it really was, and next month a point better. Each error took seconds to make. Together they produce a food cost conversation about nothing.

Counting is a skill, and it is rarely taught. Most people learn it by being handed a sheet and told to get on with it.

Train every counter on the sheet itself: what each column means, what unit each line is counted in, how to weigh a part-used container and how to record a location that is empty. Count alongside a new counter the first few times, and compare your figures with theirs. The differences show exactly what they have not yet understood.

Keep the same people counting where you can. A small team of trained counters, rotating between them, produces more consistent figures than whoever happens to be on shift.

A group with several venues has every problem above, multiplied, and one more: stock that moves between venues. A case sent from one kitchen to another, or prepared food from a central kitchen, can be counted in both places or neither if transfers are not recorded on both sides.

Use the same count sheet layout, units and counting day in every venue, so the figures can be compared and combined. Record every transfer with a note signed by both venues, and check at each count that transfers sent and transfers received agree. Our guide to running a central kitchen covers transfers in more detail.

Most of these errors are prevented by the count sheet itself, before anybody starts counting.

Our guide to running a restaurant stock count properly covers the rest of the routine around the sheet.

Before a count goes into the food cost, give it a quick check. Compare the total value with the last few counts; a large jump either way is worth a second look before it is trusted. Then check the lines with the largest value and the lines that changed most since the last count, and recount any that look odd.

This takes ten minutes and saves hours later. A count error caught the morning after is a correction; one caught a month later, after the food cost has been discussed and acted on, is a problem.

A full count once a month leaves long gaps. Spot checks on a few expensive lines each week, counted carefully and compared with what was sold, show whether the full counts can be trusted. If the spot checks agree with the trend and the full count does not, look at the full count first.

You can measure how accurate your counts are, and it is worth doing once. Pick twenty lines, have one person count them, then have a second person recount them independently, and compare. The differences show where your counting habits are weakest: units, part-used containers or particular locations.

Repeat it every few months, or whenever someone new starts counting. Inventory variance that disappears when you recount was never a kitchen problem.

A few simple tools remove whole categories of error, and none of them is expensive.

Counting apps and inventory software can go further, with barcode scanning and automatic valuation, but they help only if the basics are right: every location on the list, every line with a unit, and a count done at a sensible time by people who know the sheet. Our guide to moving from a spreadsheet to inventory software covers when that step is worth it.

The stock count, the count sheet and stock values stay in your own sheet today: servearound does not track stock or counts. Inventory tracking, for the stock behind the menu, is coming.

servearound holds the sales side that a count is compared with. The menu report in Analytics shows units sold for every dish over any range of dates, which is what a spot check or a theoretical usage figure is set against. If you would like to see it, get in touch.

Units: a case recorded as a single item or the other way round. Writing the counting unit on every line of the count sheet prevents most of it.

Weigh part-used containers of expensive items and subtract the container's empty weight. For cheap items, use a consistent rounding rule, such as rounding down to the nearest quarter.

Yes, valued at the cost of its ingredients. Leaving it out understates stock and makes the period look more expensive than it was.

At the same point every period, at a quieter time, with no deliveries or service under way. Before opening or after close, the same day each week or month, works for most kitchens.

Two where possible: one counts and calls out, the other writes. It is faster, more accurate, and each catches the other's mistakes.

Have two people count the same twenty lines independently and compare. The differences show where counting habits are weakest.

The difference between the stock a kitchen should have used, from its sales and recipes, and what the count says it used. Much variance that looks like waste turns out to be counting error.

Whenever a product, a supplier, a pack size or a price changes, and with a full review every few months.

Both, ideally. Count the twenty or thirty most valuable lines weekly, so problems show within days, and count everything monthly for the accounts. A weekly count of the main lines is short enough to do carefully.

Trained counters, ideally two together, rotating among a small team. Avoid leaving the count to whoever happens to be on last.

Recount the lines with the largest value and the largest change since the last count before using it. Check units, part-used containers and missed locations first, because they explain most errors.

Only if the delivery has been recorded first, with its time. Counting a delivery that is not yet in purchases, or recording one that has not been counted, puts stock in the wrong period.

Stock count accuracy fails in six familiar ways: units that do not match the sheet, guessed part-used containers, missed locations, uncounted prepared food, stock moving during the count, and a tired count done alone at the end of a shift. A seventh, the right count at the wrong price, distorts the value even when the counting is perfect.

Prevent them with a count sheet built for the walk, with every location, the counting unit and a current price on each line, prepared items included and empty weights for expensive containers. Count at the same time every period, stop movement while counting, use two people, check the count before it feeds the food cost, and test your accuracy with an independent recount now and then.