How Do You Run a Physical Inventory Count With Your POS?

Every retailer has had the same moment. A customer asks for something, the POS says it is in stock, and it is not on the shelf, not in the back, and not on the layaway rack. Inventory drift is normal. What separates a store that stays close to accurate from a store that stops trusting its own numbers is a counting routine — and POS software that makes counting quick enough that you actually do it.

This is a practical guide to physical inventory counts: why numbers drift, whether to do one big count or ongoing cycle counts, how to prepare, what to do with the variances you find, and the questions to ask a POS vendor before you commit.

Why do my on-hand numbers drift in the first place?

Shrink from theft gets all the attention, but most of the gap between your POS and your shelf comes from ordinary daily friction:

• Receiving errors. A case of twelve gets received as twelve cases, or as one unit. This is the single most common source of large, sudden variances.
• Lookalike items rung up wrong. Two colours, two sizes, two flavours with nearly identical names. The sale is correct in the till and wrong in the inventory.
• Returns and exchanges handled at the counter. The item goes back on the shelf, the return never gets processed, or it gets processed as a different item.
• Transfers between locations. Stock moves in a car boot and nobody records it.
• In-house use, damage, and giveaways. Testers, samples, parts pulled for a repair, breakage swept up and forgotten.
• Supplier short-shipments that were never credited or corrected.

None of these announce themselves. You find out when you go to reorder, or when a customer is standing in front of you.

Should I do one big count or ongoing cycle counts?

Both have a place, and they answer different questions.

A full physical inventory counts every item in the store, usually with the doors closed. Accountants and lenders often want one, and it gives you a clean baseline for the value of your stock. The downside is obvious: it is a long day, it usually costs you trading hours, and by the time you finish the first aisle the last aisle has already moved.

Cycle counting breaks the store into sections and counts a few of them at a time — one shelf before opening, one category on a slow afternoon. Nobody has to close. Errors get caught weeks earlier, while the receiving paperwork is still fresh enough to explain them.

For most independent retailers the practical answer is cycle counts as the everyday habit, plus a full count when your accountant asks for one. Count your fastest movers and your highest-value items more often than the slow, cheap stuff. A shelf of accessories that sells constantly deserves more attention than a bin of spare parts that nobody has touched in months.

How do I get ready for a count?

Most bad counts are bad before anyone picks up a scanner. Prepare first.

Clean up the item list. Merge duplicate items. Retire items you no longer carry. If two records exist for the same product, your count will never reconcile.

Make sure everything scans. Any item without a barcode becomes a manual search, and manual searches are where mistakes happen. Print labels for the gaps before count day.

Close out anything in flight. Receive open purchase orders. Finish pending returns. Decide how you are treating layaways, items on hold, and units sitting on a repair bench — either count them and note them, or exclude them consistently. Consistency matters more than which choice you make.

Tidy the shelves. Like with like, facings straight, nothing hiding behind anything else. Ten minutes of tidying saves an hour of recounting.

Pick a quiet window. Before open, after close, or a section you can freeze while the rest of the store keeps trading. Counting a shelf that customers are actively shopping produces numbers you cannot defend.

Brief whoever is counting. Say out loud: count what you see, not what you expect. Do not fix anything. Do not restock during the count.

How should the count itself run?

Scan, do not type. Typed quantities and typed item numbers are the fastest way to turn a count into a new set of errors.

Count by physical geography, not by the order items appear in a report. Work a shelf left to right, top to bottom, then move to the next shelf. That way you can see where you have been.

Use blind counts if your POS supports them — the counter does not see the expected quantity, so there is no anchor pulling them toward the number the system already has. When staff can see the expected figure, close counts tend to become exact counts.

For high-value categories, put two people on it: one counting, one recording. Then recount only the items that came back with a variance, not the whole section.

Finally, record who counted what and when. When a variance is genuinely strange, the first useful question is who was holding the scanner.

What do I do with the variances?

A variance list is not a to-do list of adjustments. It is a list of clues. If you post every adjustment without asking why, you get a tidy inventory and the same problem again next quarter.

Work from the largest value impact down, and check the usual suspects in order:

1. Receiving history. Was this item received in cases, units, or both? A variance that is an exact multiple of the case pack is almost always a receiving error.
2. Sales of similar items. If one variant is short and its twin is long by the same amount, someone rang the wrong one.
3. Transfers. Did stock go to another location without paperwork?
4. Physical hiding places. Window display, back room, repair bench, staff room, the box under the counter.

Only then post the adjustment, with a reason attached. Over a few counts, those reasons turn into a pattern — one category that shrinks, one supplier that short-ships, one shift where returns go unprocessed. That pattern is worth more than the adjustment itself.

How does counting work when I have more than one location?

Each location needs its own on-hand figures and its own counts. A count at your second store should never touch the first store's numbers, and stock in transit between them has to be visible somewhere, or it will show up as a shortage at one end and a surprise at the other.

Before a multi-store count, close out transfers so nothing is in limbo. Count locations on the same day if you can, or at least before stock moves between them again. And check that your variance reporting can be read per location, because a company-wide total hides the one store that has a real problem. If you are weighing up systems for more than one site, our multi-location comparison goes deeper on per-location pricing and reporting.

What should I ask a POS vendor about inventory counts?

When you are demoing systems, counting is easy to skip and expensive to get wrong. Ask directly:

• Can I count one section without the system assuming every uncounted item is now zero?
• Can I save a count in progress and come back to it tomorrow?
• Can more than one person count at the same time?
• Do you support blind counts?
• Can I review a variance report before anything is committed to inventory?
• Is there an audit trail showing who counted and who approved the adjustment?
• Will the scanner I already own work for counting? (See our guide to reusing existing hardware.)
• Does adding a second device or an extra staff account for count night change my bill?

That last one catches people out. Plenty of systems charge per register or per employee, so the sensible move — a couple of extra devices in staff hands for an hour — costs money. PHP Point Of Sale does not charge per register or per employee, so staffing up for a count is an operational decision rather than a billing one. If cost per till is shaping your shortlist, our breakdown for independent retailers covers what to compare.

FAQ

How often should I count?

Count your top sellers and your most valuable items on a short cycle, and the long tail rarely. A workable starting point is one section per week, chosen so that fast movers come around often and slow movers get touched at least once a year. Frequency matters less than never skipping.

Do I have to close the store to count?

Not for cycle counts. Count a section before opening, or freeze one aisle while the rest of the shop trades. A full store count is easier with the doors shut, but if you are counting sections regularly you will need the full count less.

What is a blind count, and does it matter?

A blind count hides the expected quantity from whoever is counting. It matters because seeing the expected number nudges people toward it, especially when they are tired. Blind counts take the same time and produce numbers you can trust.

Should I count everything, or only what sells?

Everything, eventually — dead stock still sits on your balance sheet and still takes up shelf space. But not everything at the same frequency. Weight your effort toward what moves and what is worth money.

Who should do the counting, staff or the owner?

Staff can count. The owner or manager should review the variances. Separating the counting from the approving is a basic control, and it is also the point where you learn something about your process instead of just fixing a number.