Inventory
Replace the Annual Stocktake With Cycle Counting
Shutting the warehouse for a weekend once a year gives you one accurate day and eleven months of drift. Counting a little every day gives you a number you can trust in March.
The annual physical count is a ritual with a clear purpose — satisfying the auditors — and a poor operational return. It is disruptive, it is done at speed by people who do not normally count, and its accuracy decays from the moment the warehouse reopens.
What cycle counting changes
Instead of counting everything once, count a small selection every day, weighted so that valuable and fast-moving items come round more often. The warehouse never closes, the counters get good at it, and discrepancies surface within days of the error rather than months.
• A items: every month or two.
• B items: two or three times a year.
• C items: annually.
• Anything with a discrepancy: again next cycle, until it is clean twice.
Count where the errors are
A fixed rotation is fine to start, but the discrepancies cluster and you should follow them. Trigger an extra count on any of these:
• A negative on-hand balance — always a process error, never a real quantity.
• A pick that came up short.
• A location the system says is empty but is not, or vice versa.
• Anything received into the wrong location last week.
A negative stock balance is not a counting problem. It is a receipt that never happened or a movement that happened twice.
Measure accuracy properly
The honest metric is location accuracy: the share of counted locations where system quantity matched physical quantity exactly. Net value variance is the metric people prefer because offsetting errors cancel and the number looks good — two mistakes in opposite directions report as zero.
Fix causes, not counts
Adjusting the balance closes the ticket without improving anything. Each discrepancy should get a cause: mis-pick, unrecorded scrap, receipt to the wrong bin, unit-of-measure confusion, theft. A month of causes usually points at two or three fixable process defects, and fixing those moves accuracy far more than counting harder.
On the auditors
A documented cycle-count programme with good accuracy statistics is frequently acceptable in place of a full annual count, but that is a conversation to have with your auditors in advance and to confirm against the standards that apply to you. Build the programme, keep the records, and get the agreement in writing before you cancel the stocktake — not after.
Location accuracy versus net value variance is the whole game. Our net variance was 0.3% and management were delighted. Location accuracy was 71%. Same warehouse, same month, completely different picture.
Negative on-hand as a trigger caught so much for us. Every single one traced back to a receipt posted to the wrong bin or a movement processed twice. None were counting errors.
Worth stressing the auditor conversation happens first. We built the programme, cancelled the stocktake, then discovered our auditors wanted a full count in the transition year anyway. Order of operations matters.