Most businesses discover their inventory is inaccurate at the worst possible moment: when production needs a material the system says is in stock, and it isn't. The traditional answer — a once-a-year stock take — corrects records for a day, after which errors start accumulating again. Cycle counting offers a better approach: small, frequent counts that keep inventory accurate all year and reveal the causes of errors.
Why annual stock takes fall short
An annual stock take usually means shutting down receiving and issuing, mobilising many people for a day or two, and counting everything in a rush. Counts are often made by people unfamiliar with the items, under time pressure, which introduces new errors. Differences are written off, but the causes — wrong issues, unrecorded scrap, unit errors — are rarely investigated, because there are too many to analyse at once. And by the time the next problem surfaces, the stock take is months old.
What is cycle counting?
Cycle counting is counting a small subset of inventory every day or week, so that every item is counted at a planned frequency over the year. Counts are done by trained staff, in normal working conditions, and differences are investigated while the cause is still traceable. Over time, accuracy improves because errors are found and fixed at their source.
Deciding what to count and how often
Not all items need the same frequency. ABC classification is the usual basis: A items (high consumption value) counted monthly or more often, B items quarterly, C items once or twice a year. Critical spares and items with frequent errors can be counted more often regardless of class. See ABC analysis.
Other triggers for counts include: a location reaching zero (a quick check that it is truly empty), a negative stock balance in the system, an item involved in a recent error, or high-value items before a period close.
How to run a count
- Generate the count list — items or bin locations due today.
- Freeze transactions briefly for those locations, or record the count time precisely.
- Count blind — the counter records what they find without seeing the system quantity.
- Compare — the system compares the count with the expected quantity.
- Recount differences beyond tolerance, ideally by a different person.
- Investigate confirmed variances before adjusting.
- Adjust with a reason code and approval for significant values.
Investigating variances
Investigation is where cycle counting earns its value. Common causes include: material issued but not recorded, issues recorded against the wrong item, unit of measure errors, receipts posted to the wrong location, unrecorded scrap or damage, and theft. Each cause points to a process fix — better issue procedures, scanning, clearer units, controlled access. Recording the cause of each variance and reviewing patterns monthly drives lasting improvement.
Tolerances
Set tolerances by item class: very tight for high-value items, looser for low-value bulk items (for example, small percentage tolerances for screws counted by weight). Tolerances prevent wasted effort recounting trivial differences, while still catching real problems.
Measuring inventory accuracy
Track the percentage of counted items whose count fell within tolerance, by class and by location. Also track the value of adjustments. Both should improve over months as causes are addressed. Share results with stores and production teams — accuracy improves when people see it measured.
Location accuracy matters too
In warehouses with bin locations, being in the right location matters as much as quantity. An item in the wrong bin is effectively missing. Counting by location — rather than by item — catches misplaced stock and keeps bin records accurate. See bin location setup.
Batches, serials and expiry
For batch-controlled or serialised items, counts should confirm batch or serial numbers as well as quantities. This keeps traceability and expiry management reliable. See batch management.
Mobile counting apps
A counting app on a phone or scanner shows the counter which locations to visit, lets them scan item and location labels and enter quantities, and sends results instantly for comparison. Paper count sheets, manual entry and reconciliation spreadsheets disappear, and counting a few locations each day becomes a ten-minute task rather than a project.
Common mistakes
- Counting without investigating variances.
- Letting counters see expected quantities.
- Counting while transactions continue in the same location.
- Adjusting stock without reason codes.
- Stopping cycle counting when the team gets busy.
Getting started
Begin with A items in one store: count a few locations daily, investigate every variance, and fix causes. Once the routine is stable, extend to B and C items and other stores. Within a few months, most businesses see a clear improvement in accuracy — and far fewer production surprises.
Corposol builds inventory management software with mobile cycle counting. Talk to us.
Frequently asked questions
Does cycle counting replace the annual stock take?
Often it can, if counts are well-run and auditors accept the process. Many businesses reduce the annual count to a smaller verification once cycle counting is established.
Should counters see the system quantity?
No. Blind counts — where counters don't see the expected quantity — give more honest results.
What accuracy should we aim for?
Measure the percentage of counted items within tolerance. Many businesses aim for high-90s percentages on A items over time.
Do we need to stop operations to cycle count?
No. Counts are small and can be scheduled during quieter periods, with transactions for counted locations briefly paused.
Can counts be done on a phone?
Yes. A mobile app shows the items or bins to count, accepts counts by scanning, and records variances for review.
Who should do the counting?
Trained stores staff who know the items, ideally with a second person for recounts of significant variances. Consistency matters more than the number of counters.
