Challenge
The client ran a network of 12 fulfillment centers serving mid-size e-commerce brands across Central Europe. Their warehouse management system was a patchwork of spreadsheets, manual queue assignments, and a 15-year-old ERP with no API. Order-to-dispatch time averaged 4.5 hours. During peak seasons (Black Friday, Christmas), it ballooned to 11+ hours with frequent mispicks.
Solution
We built a real-time order orchestration layer that sits between the client’s sales channels and warehouse floor:
- Event-driven order ingestion via Kafka, processing 800+ orders/minute at peak
- Dynamic wave planning algorithm that batches orders by zone, carrier, and SLA deadline
- Real-time inventory sync across all 12 warehouses with conflict resolution (last-write-wins with manual override queue)
- Route optimization engine (Python/OR-Tools) for last-mile carrier selection
- Operational dashboard with live throughput, bottleneck detection, and shift performance metrics
- Infrastructure on AWS (ECS Fargate) with auto-scaling tied to order queue depth
Results
- 62% reduction in average order-to-dispatch time (4.5h down to 1.7h)
- 94% on-time delivery rate vs. 71% before the system went live
- 34% lower shipping costs through carrier optimization and zone consolidation
- Zero downtime during 2025 Black Friday (2.1M parcels processed that month)