Cycle Counting by ABC Class: Ditch the Annual Stocktake

Every January, someone in the business books out the warehouse for three days, pulls in agency staff, and counts every single item on every single shelf. By February, the stock file is accurate. By April, it isn’t.

That pattern repeats itself on most manufacturing sites. The annual stocktake is treated like a reset button, and it works, briefly. The problem is the gap between resets. Twelve months is a long time for errors to accumulate, and in a busy plant they accumulate fast.

Why the annual count fails

The fundamental issue is frequency. If you count everything once a year, you find out once a year how wrong your records are. By then the damage is done: production plans built on phantom stock, purchase orders raised for parts you already had, expediting costs to cover shortages that shouldn’t have happened.

There’s also the disruption. Closing a warehouse or stores area for two or three days costs money in lost productivity. You’re pulling supervisors and operatives off value-adding work to walk the aisles with clipboards.

And the accuracy you get from a full stocktake degrades quickly. Counts done under time pressure, by people who don’t know the parts, in a warehouse that hasn’t been properly organised, are rarely as clean as the spreadsheet suggests.

What ABC classification does

ABC analysis divides your inventory into three groups based on annual usage value. The exact thresholds vary, but a common split is:

• A items: the top 20% of parts by value, which typically account for 70–80% of total annual spend
• B items: the middle tier, perhaps 30% of parts, accounting for 15–20% of spend
• C items: the remaining 50% of parts, often representing less than 10% of spend

The logic is straightforward. A mistake on an A item costs you more than a mistake on a C item. So A items deserve more attention.

Building a cycle count schedule

Once you have your ABC classification, you set counting frequencies that reflect the risk each tier carries.

A common starting point is:

• A items: counted every month, or every four weeks
• B items: counted quarterly
• C items: counted twice a year, or annually

That means your highest-value, highest-risk stock gets reviewed twelve times a year instead of once. And you do it in small batches, counting perhaps fifty or a hundred lines a day rather than thousands in a single panic.

The daily count becomes a routine task, five to fifteen minutes for a stores operative or a team leader. No warehouse closure, no agency staff, no disruption to the production schedule.

Making it work in practice

A few things separate a cycle count programme that works from one that quietly dies after six weeks.

First, assign ownership. Someone needs to be responsible for completing the counts and investigating discrepancies. A count that finds a variance and does nothing about it is worse than useless, because it creates a false sense of control.

Second, investigate root causes rather than just correcting the number. If the count finds twenty fewer widgets than the system shows, the correct response is not to adjust the record and move on. The correct response is to find out where those widgets went. Scrap not booked? Issued to the wrong job? Sitting in a quarantine area that nobody remembered? The investigation is where the value is.

Third, track your accuracy over time. Count accuracy is a number worth measuring. If you start at 78% and you’re at 94% six months later, you have evidence the programme is working. If accuracy is flat, something in the process needs looking at.

Fourth, keep the count locations tidy. You cannot count accurately in a disorganised stores area. If your shelves are unlabelled, if mixed parts share the same location, if unprocessed returns are sitting in random spots, you will get bad count results regardless of how diligent your team is. Sorting that out before you start will save a lot of frustration.

The end of the big annual panic

A well-run cycle count programme does not eliminate the need for stock accuracy checks. What it does is spread them across the year, concentrate effort on the parts that matter most, and give you an ongoing picture of how your inventory is performing rather than a snapshot that’s stale within weeks.

For manufacturers who want to reduce working capital, cut expediting costs, and stop being surprised by shortages, this is one of the more straightforward improvements available. It doesn’t require new software. It requires a classification, a schedule, and consistent daily discipline.

If you want to see where stock accuracy sits within the broader picture of operational losses, losses.cadencestandard.com is a practical place to start sizing the problem. And if you’re looking for the templates and scheduling guides to build the programme itself, they’re in the toolkit.