Restaurant stocktaking: a fast count that explains your food cost

If opening stock is €4,820, purchases €2,140 and theoretical usage €2,380, expected closing stock is €4,580. A physical count of €4,310 creates a €270 variance, or 5.9%. The kitchen thinks it used 20 kg of chicken; the books say 24. The difference is four kilos — either the kitchen is over-prepping, the portion sizes have crept, or someone is taking home portions. Without a system to track all three, you will never know which, and you will keep paying for the difference in food cost.

Stocktaking

• Par level: the amount of each item you should have on hand to cover demand without over-ordering. Set from four weeks of actual usage, not from a recipe or a supplier's suggestion.
• On-hand: what you actually have right now. Count it, don't estimate it. A count that is 'roughly 20 kg' is a guess, not a number.
• Usage: what the kitchen pulled from the pantry between counts. This is the number that catches over-prep, plate waste, and unrecorded usage.

Where losses hide

• Spoilage: items that expire before they are used. Track by category — dairy, produce, meat, prep. Each has a different shelf life and a different countermeasure.
• Over-prep: the kitchen prepares more than the service will use. The surplus is either thrown away, stored (and wasted), or 'forgotten' in the back of the fridge.
• Plate waste: food that goes out on the plate and comes back. Track it by dish. A dish that generates 10% plate waste is a menu problem, not a kitchen problem.
• Unrecorded usage: ingredients that leave the kitchen without a record. Tasting portions, staff meals, and 'just one more' all add up.

A worked example

Making it stick

1. 1Set par levels. Track actual usage for four weeks. Set par levels based on the 75th percentile of usage — enough to cover most weeks without over-ordering.
2. 2Count weekly. Count on-hand inventory every week, not every month. A weekly count catches a leak in the week it starts; a monthly count catches it after it has already cost you.
3. 3Flag variances above 5%. Any item with a variance above 5% gets flagged for investigation the same week. A variance that waits two months is a mystery, not a number.
4. 4Fix the cause, not the symptom. If the variance is from over-prep, fix the prep schedule. If it is from portion creep, fix the portion size. If it is from unrecorded usage, fix the record-keeping.