Direct answer: Bad inventory accuracy is almost never a counting problem. It is a transaction discipline problem. Inventory goes wrong at receiving, putaway, picking, and returns, and counting only measures the damage after the fact. The fix is to find where transactions break, close those gaps with written standard work and ownership, then use cycle counting to verify rather than to repair.
Where accuracy actually breaks
| Breakpoint | What goes wrong | Why it persists |
|---|---|---|
| Receiving | Quantity accepted without verification, wrong item or UOM | Pressure to clear the dock |
| Putaway | Product stored in an unrecorded location | No location discipline or confirmation step |
| Picking | Short pick substituted without a transaction | Operator solves the order, not the record |
| Returns and RMA | Product back on the floor before it is received in | No quarantine control |
| Customer or vendor owned stock | Not segregated, counted as sellable | Ownership rules never written down |
| Adjustments | Variances written off without root cause | Counting treated as the fix |
The last row is the one that keeps operations stuck. If every count ends in an adjustment and no one asks why, the count has become a maintenance ritual rather than a control.
Diagnose before you count
Run this before changing the cycle count program.
- Pull ninety days of adjustments and rank by item, location, and transaction type
- Identify which function the variance traces to, not just which SKU
- Check whether adjustments cluster by shift, by operator, or by process step
- Verify item master accuracy: dimensions, UOM, conversion factors
- Confirm whether customer owned and vendor owned inventory is physically and systemically segregated
Variance that clusters is a process defect. Variance that is evenly scattered is usually master data.
The control set that holds accuracy
- Written standard work for receive, putaway, pick, pack, ship, and return, with one named owner per process
- A confirmation step at every inventory movement. No movement without a transaction
- Location discipline with no unrecorded storage, including staging and overflow
- Quarantine control for returns, damage, and RMA so nothing reenters sellable stock uninspected
- Segregation and labeling rules for customer owned and vendor owned inventory
- Cycle count compliance tracked as a measure, with a named escalation path when counts are missed
- Root cause required on every adjustment above a defined threshold
What to measure
| Measure | Target behavior |
|---|---|
| Inventory record accuracy by location | Trend up and hold |
| Cycle count compliance | Counts completed on schedule, not caught up in batches |
| Adjustment value and frequency | Trend down with documented root cause |
| Receiving accuracy | Verified at the dock, not discovered later |
| Variance by transaction type | Shows which process still leaks |
Accuracy percentage alone is a vanity measure. Pair it with adjustment root cause or the number tells you nothing actionable.
What bad accuracy costs
Inaccurate inventory drives expedited freight, short shipments, lost sales, excess safety stock, and labor spent searching. It corrupts demand planning and makes every downstream decision less reliable. In acquired or consolidated operations it also distorts the working capital picture a sponsor underwrote.
How Pursuing Excellence approaches this
Warren Stout has led inventory accuracy and governance recovery across distribution operations, including customer owned and vendor owned inventory control, cycle count compliance, and RMA and quarantine process design. The approach fixes the transaction discipline first and uses counting to verify the fix, which is the sequence that makes the gain permanent.
Read the case study on post consolidation service recovery and inventory governance. Scope questions are answered on the FAQ.