What a Warehouse Count Checks
A site total that agrees with the system proves less than it seems: two bins wrong in opposite directions add up to a correct total and a warehouse nobody can pick from.
A warehouse audit therefore works location by location. For each bin it covers, it asks whether the quantity found matches the quantity the system holds for that bin, and it reads every difference against the locations around it before deciding what the difference means. Four checks make up the core of the work.
- Bin-level quantity. The count for each location compared with the system figure for that location, not with the item total for the site.
A correct total can hide wrong bins - Put-away errors. Stock present at a location the system does not associate with it, usually a pallet placed in the nearest free space.
Reads as a shortage plus an excess - Picking errors. A location short by exactly what a neighbouring location holds in excess, because the picker took from the wrong bin.
Pairs with the bin next door - Negative balances. Any item or location the system shows below zero, which is physically impossible and so always a recording error.
Always traceable to an entry
What a bin-level checklist does not settle on its own is ownership and timing: whether the goods on the floor belong to the business, and what was between two places at the cut-off. Those are tested from documents rather than bins, and the guide to what a warehouse stock audit has to settle covers both.
Selecting Which Bins to Count
Most warehouse counts cover a selection of locations rather than every bin, and three design choices decide how much that selection can show.
Choice one
Which locationsWeight the selection by value and by movement. High-value lines are where a difference costs most; fast-moving locations are touched most often and collect errors fastest. A slow line at the back of the building has had few chances to go wrong since the last count.
Choice two
Blind or system-assistedA blind count gives the counter no expected quantity, so nothing pulls the result towards the system figure. It is slower and throws up more differences to investigate. Showing the system figure is quicker, and risks the counter confirming what they were shown.
Choice three
When to recountSet a threshold, by quantity or by value, above which a difference is recounted before the count sheet closes. A transcription slip caught on the floor costs a minute; the same slip recorded as a variance can take a week to unwind.
A workable pattern is to count blind on the first pass and bring in the system figure only at recount, which keeps the first reading honest and the recount quick. Adding a few bins picked at random to the weighted selection means no location can count on being skipped.
Put-Away Errors and Their Signature
Put-away errors are among the largest sources of apparent loss in a warehouse, and they leave a pattern that identifies them quickly once someone looks for it.
The classic form is the adjacent bin. The assigned location was occupied or blocked, so the pallet went into the next one, and the system still shows it where it was meant to go. The count then finds the assigned bin short and its neighbour over by the same quantity of the same item. Read together, the two lines explain each other in seconds; read separately, each one is an unexplained difference.
Mixed items in one location are harder to unpick. Where a bin holds two items and the system believes it holds one, the count either records the wrong item or includes goods that belong elsewhere, and the error now sits in two records instead of one.
Nothing has left the building. In a count reported as totals, the shortage is listed while the excess is absorbed or reported somewhere else, and what remains looks exactly like theft.
The working rule is simple to state: before any shortage is recorded as a loss, look for the same item in the surrounding locations and in any bin holding more than the system expects. Only a shortage that survives that search is a candidate for loss.
Negative Stock: A Records Problem
A location cannot hold less than nothing, so a negative system balance is never a stock condition. It is a recording error, and it can always be traced.
The usual cause is an issue booked before its receipt. Goods arrive, are put away and are picked before anyone posts the receipt, so the system records an issue from a location it believes is empty. When the receipt is finally posted the balance corrects itself, which is why negatives often appear and vanish with nobody asking why. Receipts posted to the wrong item or location, and units of measure that differ between purchase and issue, produce the same symptom.
- List the transactions for that item and location in time order: receipts, issues, transfers and adjustments.
- Find the entry that took the balance below zero, and the entry that should have come before it.
- Name the process gap that let the order reverse, such as picking before the receipt is booked.
- Correct the record with a stated reason, and change the process rather than only the number.
Adjusting a negative balance to zero removes the symptom and leaves the cause in place. The same sequence produces the same negative the following week, and the adjustments pile up as unexplained write-offs that eventually surface in the shrinkage figure as loss.
Evidence Collected During the Count
Working papers from a warehouse count are organised by location, because that is how every difference will have to be investigated afterwards.
- Count sheets by bin, recording what the system said, what was found, who counted it and when.
Item-wise sheets cannot trace a put-away error - An exception list kept while the count runs, so a shortage in one bin and an excess in another are matched while both can still be recounted.
Close exceptions on the day - Photographs of disputed lines: stock in an unrecognised location, damaged pallets, goods bearing another party's markings, any line the count was challenged on.
The one record that cannot be rebuilt later
An exception closed on the floor, with both locations recounted and the explanation written down, is worth far more than one investigated a week later from a spreadsheet. Goods carrying another party's markings are recorded and set aside for the ownership check rather than counted into the business's own stock.
Running the Checklist Yourself First
A self-check before the audit turns the audit into confirmation rather than discovery. Take a sample of bins weighted towards value and fast movement, count them blind, and compare with the system location by location. What comes out is a list of today's differences, and most of them will be put-away and picking errors that can be corrected while the stock is in front of you.
What to do with what the self-check finds
Correct genuine errors, disclose what is missing, and fix the process that produced both.
Correct
- Stock found in the wrong bin: move it or correct the location record
- A reason noted against every correction
Disclose
- A quantity that is simply missing
- Quantified and investigated, never adjusted away
Fix the process
- Receipts booked before stock is picked
- Put-away confirmed to the assigned bin
Adjusting the system to match a missing quantity removes the evidence of the difference and answers nothing. Where the count supports the accounts, where a lender has a charge over the stock, or where the facility holds goods for other businesses, the result has to come from someone independent of the operation; an independent warehouse stock audit covers bin-level verification and the ownership question together.
