Key Takeaways
- Why apparel retail is the best starting point for RFID
- Deployment path: from pilot to full rollout
01Why apparel retail is the best starting point for RFID
Apparel retail has three natural characteristics that make RFID ROI materialize fastest: high SKU density (5,000-30,000 SKUs per store), moderate per-item value, and rigid cycle-count frequency. Barcode-counting a 10,000-SKU store takes 4 people × 8 hours = 32 labor-hours, with accuracy of only 75-85% (missed scans, wrong scans, duplicate scans). RFID handhelds bulk-read, completing a full-store count in 30 minutes at 98%+ accuracy. 50-store pilot data: cycle-count time dropped from 8 hours to 30 minutes (-94%), count frequency increased from quarterly to weekly, and inventory accuracy rose from 82% to 99.2%.
- Count time: 8h → 30min (-94%)
- Count frequency: quarterly → weekly
- Inventory accuracy: 82% → 99.2%
02Deployment path: from pilot to full rollout
The 50-store pilot proceeded in three phases. Phase 1 (2 weeks, 3 flagship stores): deployed RFID handhelds + fixed readers at store entrances/exits, completed tag encoding mapping and EPC-to-SKU binding. The key here is cleansing master data — apparel's size×color matrix causes SKU explosion, so each SKU's EPC encoding rule must be correct. Phase 2 (4 weeks, 15 stores): connected to Odoo ERP inventory module, auto-writing count data to stock.quant and triggering replenishment suggestions. Phase 3 (8 weeks, 32 stores): deployed smart shelves and fitting-room readers for real-time inventory visibility and theft alerts. After full deployment: stockout losses down 42%, shrink down 62%.
ROI calculation based on a 10,000-SKU store: RFID handhelds 2 × ¥4,500 = ¥9,000; fixed readers 2 (entrance) × ¥12,000 = ¥24,000; tags 30,000 × ¥0.35 = ¥10,500; Odoo middleware annual ¥18,000. Year-1 investment ¥61,500. Annual savings: count labor 32 hrs/quarter × 4 × ¥80/hr = ¥10,240; stockout recovery ¥45,000; shrink reduction ¥38,000. Total annual savings ¥93,240, ROI 152%, payback 8 months.


