Cycle Counting: Inventory Accuracy Without Stopping Work

The cost of a wrong number

Inventory accuracy is not a warehouse hygiene metric. It is a growth metric wearing overalls. When your system says 12 units are on the shelf and only 9 are really there, three orders will be accepted that cannot be shipped. On your own site that is a refund and an apology. On Amazon, Flipkart, or a quick-commerce panel it is a cancellation, a hit to your cancellation rate, and a dent in the account health that decides your Buy Box and your visibility.

Phantom inventory, stock the system believes exists but the shelf does not hold, is the specific enemy. It is invisible until an order fails. Cycle counting is how you find it before a customer does.

Why annual stock takes fail

The default in many Indian warehouses is the annual physical count. Everyone stops work, counts every SKU over a weekend, reconciles, and moves on. It satisfies the auditor and fixes almost nothing operationally. Accuracy is correct for exactly one day, then drifts for the next 364. Worse, the pressure to finish fast makes the count itself error prone.

Cycle counting replaces one large disruptive event with a small daily habit. You count a defined slice of the warehouse every working day. Nobody stops shipping. Accuracy stays current because you are always looking at some part of the shelf.

Classify before you count

You cannot count everything with equal frequency, and you should not try. Use ABC classification, ranked by order-line velocity and margin contribution, not by rupee value of stock sitting idle.

• A items: the roughly 20 percent of SKUs behind about 80 percent of your dispatches. Count weekly.
• B items: steady mid-tier movers. Count monthly.
• C items: the long tail and slow movers. Count quarterly.

A brand with 600 SKUs might have 120 A items. Counting 24 of them each working day means every fast mover is verified weekly, and the whole catalogue is covered on a rolling basis without a single full shutdown. Reclassify quarterly, because velocity shifts. A new launch that becomes a bestseller should move into the A tier, and a fading SKU should drop to C so you stop spending scarce count time on stock that barely moves.

Count by location, blind

Two rules make counts trustworthy. First, count by bin location, not by SKU wandering the aisles. Go to a shelf, count what is physically there, and match it to what the system says should be at that exact location. This catches misplacements, which are a huge source of phantom stock. Second, count blind. The counter should not see the system quantity beforehand, or the number on the screen quietly becomes the number they write down.

Record the physical count, then compare to system. The gap is your variance. A parcel scanned to the wrong bin, a return put back untracked, a picking error, damage never written off, all surface here.

Reconcile, then investigate

Correcting the system number is the easy half. The valuable half is asking why the gap existed. Every variance is a clue about a broken process upstream.

If returns keep creating positive variances, your reverse logistics desk is not scanning stock back in properly. If a fast mover keeps showing negative variances, you may have a picking process that grabs from the wrong bin, or theft. If damages appear as phantom stock, your write-off discipline is weak. Log the cause code for every variance, not just the adjusted quantity. Over a month the cause codes tell you which of five or six process gaps is costing you the most, and you fix the process instead of re-counting the symptom forever.

The daily rhythm

Keep the ritual small and unskippable. A supervisor generates the day’s count list each morning, a picker counts those bins during a lull, someone independent enters the physical numbers, and variances above a set tolerance get a cause code and a same-day correction. Fifteen to thirty minutes of disciplined effort a day beats a two-day annual scramble by a wide margin.

Track two numbers weekly. Inventory record accuracy, the percentage of counted locations that matched exactly. And variance value, the rupee size of the errors you found. Healthy warehouses hold accuracy above 98 percent and watch variance value trend down as process fixes land.

Accurate inventory lets you promise stock you can actually ship. That protects cancellation rate, which protects marketplace visibility, which protects revenue. It also sharpens replenishment, because your reorder points are only as good as the on-hand number they trust. And it removes the quarter-end panic where finance and operations argue about a stock figure nobody believes.

Cycle counting is not glamorous work. It is a small daily discipline that quietly removes a whole class of failures, the oversell, the surprise stockout, the year-end reconciliation fight. For a brand scaling past a few hundred SKUs, it is one of the highest-return operating habits you can build, and it costs almost nothing but consistency.