Why your physical stock doesn't match the system: seven causes and their fixes
In every business that holds stock, the day arrives when you're staring at two numbers that don't agree: the system says 120 units, the shelf holds 107. The first question asked is usually "who took them?" — and it's usually the wrong question.
Stock discrepancies are rarely a single incident. They're normally the result of a small leak that repeats every week and only surfaces at count time. Looking for a culprit before looking for a cause wastes your time and costs you your team's trust.
Why the number matters at all
Inventory isn't goods on a shelf; it's cash you've frozen. An inaccurate number means inaccurate purchasing: you buy what you already have, you run out of items you believed were in stock, and you promise a customer a date you can't meet. The damage doesn't show up in the warehouse — it shows up in sales and cash flow.
The seven most common causes
1. Receiving before recording
Goods arrive, someone signs the delivery note, and the boxes go on the shelf. Entering them into the system is postponed until "this afternoon," then forgotten — or entered later from memory. It's the most common cause and the least dramatic, which is exactly why it gets ignored.
2. Inconsistent units of measure
An item is bought by the carton and sold by the piece. If the conversion factor isn't fixed in the system, an employee will enter one carton where they meant twelve pieces. This isn't missing stock — it's an arithmetic error that repeats with every transaction.
3. Returns that never return
A customer returns an item, gets refunded, and the unit is set aside pending inspection. Then it stays there. The system deducted it from sales but never added it back to stock — or added it back damaged and unsellable.
4. Damage and expiry with no entry
Something breaks or expires and goes straight in the bin. Nobody stole it, but it left the shelf and stayed in the system. Warehouses without a clear write-off procedure accumulate a steady gap every month.
5. Samples and internal use
One unit goes to a customer as a sample, another is used in the showroom, a third is taken by an employee for an operational need. All legitimate uses — and none of them recorded, because nobody thinks of them as a "sale."
6. Look-alike items
Two colours of the same product, or two close sizes, with unclear coding. One is issued in place of the other, so one item shows a surplus and another shows an identical shortfall. If you see a positive and a negative gap of the same size, look here first.
7. Open permissions
When any user can edit a quantity directly and the system keeps no record of who changed it, when, and why, you don't have inventory — you have an estimate. This is the only point where theft becomes a reasonable line of inquiry — the missing audit trail is what makes it possible.
Read the shape of the gap before you investigate it
The discrepancy itself tells you where the fault sits:
| Shape of the gap | Where to look first |
|---|---|
| Shortfall in one fast-moving item | Receiving and manual issuing |
| Equal surplus and shortfall in two items | Coding and look-alike items |
| Small shortfalls spread across many items | Units of measure and write-offs |
| A gap that appears all at once on one day | A manual edit in the system |
| The same percentage missing every month | Unrecorded internal use |
An annual count isn't enough on its own
A business that counts once a year discovers the gap twelve months after it happened, by which point every trace that could explain it is gone. The practical alternative is cycle counting: a small group of items every week, chosen by value and movement speed, without halting operations.
A count that happens once a year measures the loss. A count that happens every week prevents it.
Start with your highest-value, fastest-moving items — they produce most of the variance. An item that sells twice a year doesn't deserve the same attention.
A checklist before your next count
- Does every item have one unique code and a defined location in the warehouse?
- Is the conversion factor between purchase unit and sales unit fixed in the system?
- Is there a written procedure for recording damage, returns and samples?
- Are manual quantity edits restricted to a specific permission, and do they leave a trail with user, time and reason?
- Is receiving recorded the moment goods arrive, rather than at the end of the day?
- Do you know what variance rate is acceptable, and when a gap becomes an investigation?
Three "no" answers mean your next count will hand you a new number, not an explanation.
Where to start if things are messy
Don't fix everything at once. Pick the ten highest-value items in your warehouse, tighten coding, receiving and write-offs on those for one month, then measure the gap. If it improves, you know what works, and rolling it out to the rest is a matter of time.