Inventory Accuracy: Cycle Counting with Barcode Scanners
Implementing barcode-based cycle counting — ABC analysis, count scheduling, scanner workflows, variance investigation, and accuracy targets.
Inventory Accuracy: Cycle Counting with Barcode Scanners
Cycle counting is the practice of counting a small subset of inventory on a regular schedule rather than conducting a full physical inventory. Barcode scanning makes cycle counting fast, accurate, and minimally disruptive to operations.
Why Cycle Count
Traditional annual physical inventory: - Requires shutting down operations for 1-3 days - Counts everything at once (expensive, disruptive) - Accuracy degrades throughout the year
Barcode-based cycle counting: - Counts a subset daily during normal operations - Every SKU counted multiple times per year - Maintains high accuracy continuously
ABC Analysis
Prioritize counting frequency by inventory value:
| Class | % of SKUs | % of Value | Count Frequency |
|---|---|---|---|
| A | 10-20% | 70-80% | Weekly or monthly |
| B | 20-30% | 15-20% | Monthly or quarterly |
| C | 50-60% | 5-10% | Quarterly or annually |
A-class items (highest value) are counted most frequently because errors have the greatest financial impact.
Cycle Count Workflow
- WMS selects locations: System generates today's count list based on ABC class, last count date, and discrepancy history
- Counter scans their badge: Identifies who is performing the count
- Scans location barcode: Confirms which bin/shelf is being counted
- Scans each product barcode: Identifies the SKU
- Enters quantity: Physical count for that SKU at that location
- WMS compares: System quantity vs physical count
- Variance handling: If within tolerance, auto-adjust. If outside, flag for investigation
Tolerance Thresholds
| Metric | Typical Tolerance | Action |
|---|---|---|
| Unit variance <1 | Auto-adjust | No investigation |
| Unit variance 2-5 | Recount required | Verify with second counter |
| Unit variance >5 | Investigation required | Research cause before adjustment |
| Dollar variance >$100 | Management review | Document root cause |
Blind vs Non-Blind Counting
Blind count: Counter does not see the system quantity. Forces an actual physical count rather than confirming the expected number. More accurate but slower.
Non-blind count: Counter sees the system quantity and verifies or corrects it. Faster but prone to confirmation bias.
Best practice: Use blind counting for A-class items and non-blind for C-class items.
Scan-Based Counting Methods
Location-based: Count everything at a location. Best for dense storage areas.
SKU-based: Count all locations containing a specific SKU. Best for high-value items stored in multiple locations.
Triggered: Count when an event occurs (receiving error, pick shortage, negative inventory). Addresses discrepancies immediately.
Accuracy Metrics
| Metric | Formula | Target |
|---|---|---|
| Unit accuracy | Locations with correct count / total counted | >98% |
| Dollar accuracy | (1 - abs(variance) / total value) x 100 | >99.5% |
| Perfect order rate | Orders with zero picking errors / total orders | >99.7% |
Root Cause Analysis
When cycle counts reveal discrepancies, investigate:
- Receiving errors (wrong count or wrong product)
- Picking errors (wrong item or wrong quantity)
- Putaway errors (product in wrong location)
- Damaged/discarded items not recorded
- Theft
- System errors (duplicate transactions, timing issues)