You do not need to close, and you certainly do not need a big annual inventory. A variance found ten months later cannot be explained, only recorded: nobody remembers March’s delivery. The method that works in a small shop is the rolling count, one shelf a week during a quiet hour. Here is how, and what to do with the variance.
In short
- Count one shelf a week, not the whole shop once a year. In two months you have covered everything.
- A recent variance is explainable, an old one is not. That is the whole difference.
- Count with a scanner if your products have a code: twice as fast and no transcription errors.
- Never count while it is selling. A quiet hour, or before opening.
- The variance is not the result, its cause is. Note it next to the figure.
Annual inventory or rolling count
Both exist and they serve different purposes. One satisfies the accountant, the other corrects the business.
| Full annual inventory | Rolling count | |
|---|---|---|
| When | Once a year, often on a Sunday. | One shelf a week, at a quiet hour. |
| The shop | Closed, or counted in disorder. | Open, selling continues elsewhere. |
| The variance found | Accumulated over twelve months, unexplainable. | Recent, attributable to a delivery or a week. |
| The effort | A whole day, the whole team. | An hour, one person. |
| What you do with it | Correct the stock and start again next year. | Correct the cause, and the next shelf is better. |
The rolling count does not replace a year-end accounting inventory if your accountant requires one. It does something else: it gives you usable stock all year, which the annual inventory does, at best, for one week in twelve months.
The three things that corrupt a count
They explain nearly all the “variances” that are not variances.
- Counting while it sells: An item goes through the till while you are counting the shelf, and the variance was created by the count itself. Count before opening, or with the shelf closed to sales for the hour.
- Goods with no code: Loose, imports without a label, repackaged: impossible to scan, so counted by hand, so transcribed, so wrong. That is the first reason to label at goods-in rather than later.
- Two people counting differently: One counts the boxes in the store room, the other does not. One counts damaged items on the shelf, the other sets them aside. Decide the rule first, and write it on the sheet.
A rolling count, in one hour
Six steps. The fifth is the one everybody skips and the only one that produces value.
- Pick a shelf and keep a fixed order: always left to right, top to bottom.
- Pick the moment: before opening, or the quietest hour of the week.
- Count everything belonging to that shelf, including the store room and the back of the shelf. That is where positive variances hide.
- Enter as you go, with a scanner if the products have codes, otherwise on a sheet per sub-section.
- Note the cause beside every variance over a unit or two: known breakage, unsold thrown out, doubt about a delivery.
- Correct the stock, then move to the next shelf the following week.
Without step 5 you have up-to-date stock and no information: the same variance will return next time round. With it, after two months you can see which shelves drift and why, and that is where counting starts to pay.
What to do with the variance
Always correct stock to the real figure: keeping a false theoretical stock “so as not to lose the trace” traces nothing at all. But correct it while recording a reason, not by silently overwriting the number. Stock corrected with no reason will produce exactly the same variance next quarter, and you will never have the list of products that genuinely drift.
What complicates stock takes in Morocco
Goods with no barcode, first: loose, imported, craft and repackaged items make up a significant share of many shelves, and each one doubles the counting time. Labelling at goods-in, covered in our piece on https://blog.bellocommerce.com/en/ean-gtin-barcodes/, is what makes a rolling count sustainable rather than heroic.
Then seasonality: Ramadan and the two Eids move unusual volumes in and out within a few weeks. Count the affected shelves before the season rather than after: accurate stock at the start makes the end-of-season variance readable, whereas stock already wrong at the start makes the whole month uninterpretable.
The list before you start
Five points, two minutes.
- The rule is written down: what counts, what does not.
- The shelf is not selling during the counting hour.
- The scanner is charged and stock is up to date with today’s deliveries.
- A “cause” column exists on the sheet or on screen.
- Next week’s shelf is already chosen.
Mistakes to avoid
- Counting once a year. You get a figure, not an explanation.
- Counting during selling hours. You manufacture the variance you are looking for.
- Forgetting the store room. That is where half the positive variances are.
- Correcting without noting the cause. The same variance returns next time.
- Counting a whole shelf of unlabelled loose goods. Label first, count second.
Frequently asked questions
How often should I do a stock take?
For a small shop, one shelf a week rather than a full inventory once a year. In two months you have covered the whole shop, and above all every variance found is recent, so attributable to a delivery or a specific week. An annual variance can no longer be explained.
Do I have to close the shop to do a stock take?
No, if you count shelf by shelf. It is enough that the shelf being counted is not selling during that hour: before opening, or at the quietest time. Counting while the shelf sells mechanically creates a variance that did not exist.
How do I count products with no barcode?
By giving them one. Generate the code, print the label and apply it at goods-in: the item becomes countable with a scanner like any other. While it has no code it is counted by hand and transcribed, which is the first source of counting error.
What do I do with the variance I find?
Correct stock to the real figure, always, but record a reason: breakage, unsold, doubt about a delivery. Stock corrected with no reason will produce the same variance next time, and you will never have the list of products that genuinely drift.
Does a rolling count replace the accounting inventory?
No, if your accountant requires a year-end inventory. The two serve different purposes: the annual inventory values stock at a date, the rolling count gives you stock you can actually use all year for ordering and for calculating margins.
What to take away
Stop waiting for December. Take one shelf, a quiet hour, a written rule and a column for the cause, and do it again next week with the shelf beside it. In two months you will have covered the shop, corrected stock as you went, and above all you will know which shelves drift and why, which no annual inventory will ever tell you.
Stock that stays accurate between counts
BelloPOS keeps stock current on every sale and every goods-in, accepts counting with a scanner, and keeps the reason for every correction. The Lite licence is free for life and entirely offline.
Read next
Other practical guides on the same subject: