Cycle Counting Inventory: Count Without Stopping Ops

What is cycle counting inventory

Cycle counting is a physical inventory method in which a small part of the stock is counted every day, on a rotating schedule, instead of shutting down the plant or warehouse once a year to count everything. Each item is counted at a set frequency, usually higher for items of higher value or usage.

The point is not just matching a number: every variance is investigated, the root cause is fixed, and the system balance ends up matching what is on the shelf. That gives MRP and production planning and control reliable on-hand balances all year, not only after the annual physical count.

Cycle counting visual map

The map shows count frequency by ABC class, the counts-per-day calculation and the inventory record accuracy formula.

Setting the frequency: ABC classes

The most common way to build the cycle is to classify items with an ABC analysis of inventory and count more often the items that carry more value. A typical starting rule:

These numbers are a starting point, not a standard. Adjust for risk: items critical to the line, items with a history of variances or shrinkage can be counted more often even if their value is low.

Counts per day = Σ (items in class × counts per year) ÷ working days per year

Worked example: 500 items

Fictional example: a stockroom with 500 SKUs and 250 working days a year, split into 50 A items (10%), 100 B items (20%) and 350 C items (70%).

That is about 5 to 6 items a day, a few minutes per item for one person. Compare it with a yearly wall-to-wall count: 500 counts at once, over a shutdown weekend. Cycle counting makes more counts in total (1,350 against 500), but they are small, spread out and need no production stop.

Step-by-step implementation

1. Clean the item master. Duplicate part numbers, wrong units of measure and nonexistent locations create false variances.
2. Classify the items by ABC and set the frequency for each class.
3. Build the calendar. Spread the counts over the working days; do not pile everything on Monday or on month end. An ERP can generate the daily list; without one, a spreadsheet works.
4. Set the tolerance. For example: class A, zero variance; class B, up to 2%; class C, up to 5%. A variance outside tolerance triggers an investigation.
5. Freeze movement only for the item being counted. Count right before or after a transaction, or block the location for a few minutes. The rest of the operation runs as usual.
6. Count blind. The counter does not see the system balance, so the result is not bent toward the expected number. On a variance, have a second person recount.
7. Find the root cause (wrong transaction, item in the wrong location, loss, receiving error) and record the corrective action. Only then adjust the balance.
8. Measure and react. Track accuracy and raise the count frequency of items that keep failing.

Measuring: inventory record accuracy

Accuracy (%) = items counted within tolerance ÷ items counted × 100

Example: in one month 120 counts were made and 108 were within tolerance. Accuracy = 108 ÷ 120 × 100 = 90%. Also track accuracy by value (value of variances ÷ value counted) and by class, because an overall 90% can hide 70% in class A.

The target depends on the operation, but plants that run MRP or automatic kanban replenishment need high accuracy to avoid shortages and excess. The metric belongs on the warehouse board next to the other performance indicators.

Cycle counting or wall-to-wall count

• Cycle counting: large stock, continuous operation, an ERP with location control and a need for reliable balances all year.
• Wall-to-wall: small stock, an operation that stops anyway, or a requirement from accounting or auditors. Check with your finance team whether a documented cycle counting program is accepted for the year-end close in your company.
• Combination: many companies run cycle counts all year and still do a lean annual full count, with clean balances, which shortens the downtime.

The better the inventory control and the physical organization (see 5S), the faster each count goes.

Common mistakes

• Counting before cleaning the item master and the locations.
• Adjusting the balance without finding the cause: the error comes back at the next count.
• Showing the system balance to the counter.
• Using the same frequency for every item and wasting time on low-value C items.
• Letting the calendar slip: a late cycle count turns into a stopped inventory.
• Tracking only the number of counts done, not accuracy or root causes.

Frequently asked questions

What is cycle counting?

It is counting parts of the inventory on a rotating schedule, with a set frequency per item, without stopping operations for a full physical count.

What is the difference between cycle counting and a physical inventory?

A physical inventory counts everything at once, usually with operations stopped. Cycle counting counts a few items every day through the year.

How often should each item be counted?

A common rule is monthly for class A, quarterly for B and yearly for C, adjusted for risk and variance history.

How many items should be counted per day?

Add up the yearly counts for all classes and divide by working days. In the example, 1,350 ÷ 250 = 5.4 items a day.

How do you calculate inventory accuracy?

Items counted within tolerance divided by items counted, times 100. With 108 of 120, accuracy is 90%.

Can cycle counting replace the annual physical inventory?

In many companies it can, but it depends on accounting and audit requirements. Confirm before dropping the full count.

About the author

Vagner Soares

Lean Manufacturing & Behavioral Management Specialist

Over 20 years in the automotive and metalworking industries (GM and Dana), Lean Manufacturing practitioner since 2006. SENAI instructor and mentor in Brazil’s Brasil Mais Produtivo program, delivering consulting, training and audits for 50+ companies, combining quality, productivity and people development.

LinkedInInstagramFull profile →