Chasing a shortage alone
The matching excess one bay away is never looked for.
Stock audit glossary · Warehouse and quick commerce
noun · also called picking accuracy; its opposite is the mispick rate
Pick accuracy is the share of order lines picked with the right item in the right quantity. It matters to stock records as much as to customers: one wrong pick makes two lines wrong at once, because the system deducts the item the order asked for while the shelf loses the item actually taken.
Meaning
Pick accuracy is usually discussed as a customer service figure, but its effect on the stock record is larger. When a picker takes the wrong variant, the system deducts the variant on the order while the shelf loses a different one, so two lines are now misstated in opposite directions. Nothing in the transaction looks unusual, so nothing is flagged.
At a count the pattern is distinctive: one line short by some quantity and a neighbouring line over by the same amount. Nothing has left the building, and the remedy is a shelf or labelling change rather than an investigation. Telling that pattern apart from one-sided shortages, where goods really are missing, is one of the most useful things a line-level count report does, and it cannot be seen in a summary total.
Where it is used
The more alike the items, the more it matters.
In practice
A stated quantity of a stated item from a stated location.
Location and item both scanned, so the system checks the right thing was taken.
Nothing checks the physical action, so a wrong pick is recorded as right.
An order that is broadly right can still contain several wrong lines.
Equal and opposite differences point to picking, not loss.
Worked example
Sample figures.
| Cause | Lines | Share of errors |
|---|---|---|
| Right item, wrong quantity | 148 | 41% |
| Look-alike pack, wrong variant | 122 | 34% |
| Right item, wrong location scanned | 61 | 17% |
| Item not at its location | 29 | 8% |
99.14% sounds strong and hides four different faults. Quantity errors are a scanning discipline issue. Look-alike variants are rarely fixed by training; they are fixed at the shelf, by separating the two packs. Only 8% of the errors were stock genuinely missing from its location.
Reported as one number, the result would have prompted general retraining. Split by cause, it points at shelf layout, which is cheaper and deals with about a third of the errors.
Watch out
Reading the symptom as theft is the costly error.
The matching excess one bay away is never looked for.
Money goes on a loss the data says did not happen.
Each error corrupts two stock lines until someone counts.
Look-alike packs keep being confused.
Wrong lines inside broadly right orders disappear.
A wrong pick enters the records as a right one.
Patron Accounting LLP. This page makes no regulatory statement.
Sources: general stock-counting and trade practice; no statutory statement is made on this page. General information, not professional advice.
Need help with pick accuracy? See our dark store stock audit service. A dark store stock audit reads count differences line by line to separate mispicks from loss.