You have a stock count variance: how to find the cause and close it
Who it is forGrocery chain finance and accounting teams, loss prevention, warehouse managers, branch managers, internal audit
The count is done and the list is in front of you: hundreds of lines where the record and the count disagree. At this point the most common reflex is to post the variance and close it. That does not solve the problem; it only makes it invisible, and six months later the same variance reappears. A count variance is not an accounting item but a diagnostic tool: it is the only concrete data telling you which process is broken. This guide covers how to read the variance, find its cause and close it.
Negative and positive variances say different things
Looking at the variance as a net total is the biggest mistake. Negative and positive variances offset each other and the net figure looks small; in fact the sum of the two sides is the real size of the problem.
A negative variance — goods in the record but not on the shelf — comes from three causes: the goods have physically been lost, shrinkage was not recorded, or something that did not sell was recorded as sold. The first two are genuine losses, the third is a recording error.
A positive variance — goods on the shelf but not in the record — is usually a recording gap: a delivery was never entered, a return was not recorded, or a sale was rung through under another product's barcode. A positive variance is not good news; it is evidence that the recording system is unreliable.
| Variance type | Likely cause | How to verify |
|---|---|---|
| Negative | Unrecorded shrinkage | Compare shrinkage entry frequency across departments |
| Negative | Theft | Concentration in high-value, small-format items |
| Negative | Over-recorded sales | Till void and return records |
| Negative | Outbound transfer deducted twice | Transfer movement history |
| Positive | Goods receipt never entered | Compare delivery notes against system records |
| Positive | Sale under another barcode | Look for an opposite negative variance in the same group |
| Positive | Return not recorded | Customer return records |
| Negative + positive pair | Barcode mix-up | Put similar products' variances side by side |
If you see two equal variances on similar products, one negative and one positive, it is almost certainly a barcode mix-up. That is not a real loss and it is corrected differently.
The order of variance analysis
Examining every variance is both unnecessary and impossible. The analysis starts where the return is highest and stops at a defined point.
The ranking is done on value. The top 20 per cent of variances by value typically accounts for 80 per cent of the total; keeping the analysis to that group saves time and produces results.
- Total variance rate
(|Negative| + |Positive|) ÷ Inventory value counted × 100- Net variance rate
(Negative + Positive) ÷ Inventory value counted × 100- Stock accuracy rate
Correctly counted lines ÷ Total lines × 100
- Eliminate counting errorLarge variances are physically recounted. A meaningful share of variances close at this step and are never posted.
- Look for opposite variancesAre there equal and opposite variances on similar products in the same category? If so it is a barcode mix-up and both products are corrected together.
- Check recent movementsGoods receipts, transfers and returns in the two weeks before the count are examined. Movements that exist in the system but never happened physically surface here.
- Compare shrinkage recordsCompare against the shrinkage rate recorded by similar departments in the same period. If one department's shrinkage entry is notably low, that is where the negative variance comes from.
- Flag the remainder as unknownVariances with no identifiable cause are closed under an "unknown" code. That group's share is tracked; growth indicates process failure.
Acceptable variance and the accuracy target
A zero-variance target is not realistic; you need a measurable target band. Two separate measures are used and they say different things.
The value-based measure gives the financial impact: total variance as a share of inventory value. Approximate industry-practice bands are given in the table below; there is no sourced universal threshold.
The line-based measure — stock accuracy rate — gives operational quality: how many records were correct. It is harder and more informative: 97 per cent accuracy means three lines in every hundred are wrong, and ordering decisions on those three are being made incorrectly.
| Measure | Good | Acceptable | Problem |
|---|---|---|---|
| Total variance / inventory value | < 0.5% | 0.5% - 1.5% | > 2% |
| Stock accuracy rate (lines) | > 98% | 95% - 98% | < 95% |
| Share of unknown variance | < 20% | 20% - 40% | > 40% |
| Positive share of total variance | < 25% | 25% - 40% | > 40% |
A high positive share matters particularly: it means the goods receipt process is not working. That is both more common and easier to fix than theft.
Accounting treatment and the effect on cost
Variances with an identified cause are posted to their own account; those without go to a stock variance account. Without that separation the whole variance collects in one line and you cannot see which process is losing what.
On negative variances goods leave stock and are expensed. If the cause is shrinkage it goes to the shrinkage account, if theft to a loss account, if a counting error it is treated as an adjustment. On positive variances goods enter stock; at what cost they are valued is a separate decision, and the nearest purchase cost is used.
The effect on cost of goods sold is direct. A negative variance reduces closing inventory, which increases cost of goods sold and reduces gross profit. That is why a large count variance shows up as an unexplained decline in the margin report.
- Identified shrinkage → shrinkage expense account, tracked in the shrinkage report
- Theft and loss → separate loss account, an input to security measures
- Counting error → adjustment entry, not expensed
- Cause unknown → stock variance account, with its share tracked separately
Preventing recurrence
The job is not finished once the variance is closed. To stop the same variance appearing at the next count you have to go back to the process — and that is straightforward as long as the cause was coded.
The three highest-return fixes are: making barcode scanning mandatory at goods receipt, setting up daily shrinkage entry at department level, and increasing cycle count frequency for high-risk items.
What all three have in common is catching the error as it happens. A count is a diagnosis made months later; the real solution is preventing the error or catching it the same day.
- Make barcode scanning mandatory at receiptEvery goods receipt keyed manually from a delivery note is a potential count variance. Mandatory scanning removes most positive variances.
- Move shrinkage entry to the shelfShrinkage is entered on a handheld at the shelf, not from the office. Every step that makes recording harder increases unrecorded shrinkage and therefore negative variance.
- Build a high-risk item listSmall-format, high-value items and those with a history of variance are grouped on a separate list and counted weekly.
- Attach the cause to the process, not the personTying count variance to a performance penalty leads to variances being hidden. The measure should be the rate at which causes are identified, not the variance itself.
Frequently asked questions
- What does a stock count variance mean?
- It is the difference between recorded stock and the physical count. A negative variance means goods are in the record but not on the shelf; a positive variance means the reverse. It is not an accounting item but a diagnosis of a broken process.
- How is a stock count variance recorded in the accounts?
- Variances with an identified cause go to their own account: shrinkage to a shrinkage expense account, loss to a separate account, counting errors treated as adjustments. Variances with no identified cause collect in a stock variance account. Without that separation you cannot see which process is losing what.
- What is an acceptable stock count variance?
- There is no sourced universal threshold; approximate industry-practice bands are given in the table in this guide, and the exact threshold is set from the business's own count history. Assess on the sum of negative and positive variances rather than the net figure, by product and location.
- Why does a positive stock variance occur?
- Goods on the shelf but not in the record usually come from a recording gap: a delivery was never entered, a return was not recorded, or a sale went through under another product's barcode. A positive variance is not good news; it shows the recording system is unreliable.
- How does a count variance affect gross profit?
- Directly. A negative variance reduces closing inventory, which increases cost of goods sold and reduces gross profit. A large count variance appears as an unexplained decline in the margin report, and the cause cannot be found unless the variance is broken out.