Quick answer
A stock variance is a difference to explain, not proof of theft. First check count dates, units, receipts and transfers. Then compare physical usage with recipe usage, recorded waste, staff food and remakes. Recount the few items with material gaps before deciding what action is justified.
Make both sides cover the same period
A sales report ending at midnight cannot be compared cleanly with a stock count taken halfway through the next lunch service. Choose a clear cutoff and record any movements around it. If you must count while trading, use a movement record for the items being counted.
Keep the calculation in one unit. Opening stock plus receipts and transfers in, minus closing stock and transfers out, gives physical usage over the period. Damaged or quarantined goods need a consistent treatment. If they are still physically present but unavailable for sale, show that status rather than quietly forgetting them.
Build a bridge from sales to actual usage
For an illustrative coffee example, opening beans are 8 kg, receipts are 12 kg and closing beans are 6 kg. Physical usage is 14 kg. Suppose 600 recorded drinks use an 18 g recipe: that explains 10.8 kg. Recorded calibration and training use is 1.2 kg, staff drinks use 0.6 kg and spills account for 0.4 kg.
Those known uses total 13 kg, leaving 1 kg unexplained. It would be wrong to call all 3.2 kg above customer-drink recipes waste, or to call the remaining kilogram theft. The gap could include an incorrect recipe, an unrecorded transfer or a faulty count. This is an arithmetic example, not an acceptable-loss benchmark.
| Movement | Quantity |
|---|---|
| Physical usage | 14 kg |
| Customer drinks at recorded recipe | 10.8 kg |
| Calibration and training | 1.2 kg |
| Staff drinks | 0.6 kg |
| Recorded spills | 0.4 kg |
| Unexplained difference | 1 kg |
Check the cheap explanations first
Ask a second person to recount the item without first telling them the expected answer. Check whether an unopened bag has a different weight, whether a case was entered as one unit and whether stock is stored in another location. Look for a receipt entered after the count even though the goods arrived before it.
Next check the sales recipe and exceptional uses. A doubled espresso dose, staff drinks recorded under a generic button or a product used in several recipes can all distort theoretical usage. BCcampus's inventory guidance provides background on maintaining physical and perpetual records; the investigation sequence here is a practical suggestion.
- Recount and verify the unit.
- Match delivery timing to the count cutoff.
- Check all storage and transfer locations.
- Review recipe quantities and menu-button mapping.
- Include legitimate non-sale uses before labelling the remainder unexplained.
Further reading: BCcampus: Basic Inventory Procedures
Agree a short investigation, then review it
The linked bar-owner discussion shows how quickly a discrepancy can become a trust dispute. Keep your language factual: the records differ by this amount, for this period, and these possibilities have been checked. Involve the people doing the work because they may know about an undocumented practice.
For a persistent gap, choose a limited observation period with clearer records and a named reviewer. Explain what is being measured and why. Avoid collective accusations or informal deductions from wages. Any employment action needs a separate fair process and appropriate advice; an inventory calculation is not that process.
Close the investigation with the cause if known, the correction and a date to check whether it worked. If it remains unexplained, say so. A reliable system can tolerate an honest unresolved entry more safely than a confident story unsupported by evidence.
Unpickle can hold stock counts and recurring checking tasks. Keep your till sales and reconciliation calculations alongside those records when investigating a difference. Make stock checks easier to follow up.