Warehouse cycle counting is the practice of counting a small subset of inventory every day, on a rotating schedule, instead of shutting the building down once a year for a wall-to-wall physical count. Done properly it keeps inventory accuracy above 98% year-round, catches errors within days rather than months, and eliminates the annual shutdown entirely.
Why cycle counting beats the annual physical
An annual physical inventory has three structural problems: it stops shipping for a day or more, it finds errors that may be eleven months old, and it tells you nothing about why the variance happened. By then the receiving clerk who mis-keyed the pallet has moved on and the audit trail is cold.
Cycle counting inverts that. Small daily counts mean variances surface while the paperwork is still fresh, root causes are traceable, and the correction is a conversation instead of an investigation. Operations never stop.
How to choose what to count: the ABC method
The standard approach ranks SKUs by annual usage value and counts them at frequencies matched to their importance.
A items
Roughly the top 20% of SKUs driving about 80% of value. Counted most often — monthly or even weekly. These are where a 2% error costs real money.
B items
The middle tier. Counted quarterly. Moderate value, moderate movement.
C items
The long tail — many SKUs, little value. Counted once or twice a year. Spending daily labor here is waste.
ABC frequency is only the starting point. Mature operations layer on trigger-based counts too: count a location when it hits zero, when a pick short is reported, when a return posts, or when a OS&D exception is raised. Zero-balance counts are especially cheap and effective, because verifying an empty location takes seconds.
Running counts that actually work
- Count blind. Do not show the counter the expected quantity. If the system says 48 and the counter sees “about 48,” they will write 48. Blind counts are the difference between an audit and a rubber stamp.
- Count by location, not by SKU. Location-based counting is faster and matches how the WMS actually stores data.
- Freeze the location. Suspend picks and putaways for that bin during the count, or you are counting a moving target.
- Recount before adjusting. A second independent count on any variance filters out simple miscounts before they hit the books.
- Scan, don’t type. Barcode or RFID capture removes the largest single source of counting error.
- Log the root cause, not just the number. “Variance -3” teaches nothing. “Variance -3, mis-slotted to adjacent bin” fixes a process.
The metric that matters
Do not measure accuracy as net units — overs and shorts cancel out and flatter the number. Measure location accuracy: the percentage of counted locations where system quantity matched physical quantity exactly. World-class distribution operations run 99%+ on this measure. Below 95%, you have a process problem — usually in receiving, slotting, or pick confirmation — and no amount of counting will fix it.
Track variance root causes in a Pareto chart. In most warehouses, three or four causes account for the large majority of variances, and fixing those beats counting more.
Cycle counting in a 3PL environment
When a third party holds your inventory, cycle counting is also your visibility. Ask any prospective warehousing provider three questions: what is your location accuracy percentage, how often are A items counted, and can I see the variance report? A provider who cannot answer with a number is not counting systematically.
Cycle counting matters even more in a bonded environment, where CBP recordkeeping requires accurate accounting of every entry and withdrawal. See how bonded warehousing works and what a warehouse receipt covers.
Go Freight operates a 104,000 sq ft bonded warehouse at 6901 NW 26 Ave in Miami, with scanning-based cycle counts and customer-visible variance reporting. We have been an asset-based 3PL since 2004.
Frequently asked questions
How often should I cycle count?
Match frequency to value. A items — the roughly 20% of SKUs driving most of your inventory value — should be counted monthly or weekly. B items quarterly, C items once or twice a year. Add trigger-based counts whenever a location hits zero or a pick short is reported.
Does cycle counting replace an annual physical inventory?
It can, and that is usually the goal. Most auditors will accept a documented cycle counting program with strong location accuracy in place of a wall-to-wall count, provided the program covers all SKUs within the year and variance procedures are documented. Confirm with your auditor before eliminating the physical.
What is a good inventory accuracy target?
Measure location accuracy — the percentage of counted locations that match exactly, without netting overs against shorts. 98% is a solid operating benchmark and 99%+ is achievable in a well-run, scan-driven warehouse. Anything below 95% points to a process defect rather than a counting problem.
Need a Miami warehouse that counts its inventory?
104,000 sq ft bonded, scan-driven, with variance reporting you can see. Request a warehousing quote or call (786) 445-0150.