Real-time supply chain visibility across 12 carriers
Engagement
22 weeks
Team
1 tech lead, 5 engineers, 1 data engineer, 1 designer
Industry
Logistics
Stage
Enterprise
The numbers that mattered to the client.
What we walked into.
The client — a top-20 US 3PL — had visibility into 60% of shipments. The other 40% went dark between pickup and delivery. Customer service was reactive, not proactive. ETA accuracy was 62%, meaning customers were surprised by delays more often than not.
Each of the 12 carrier integrations was a snowflake — different API, different data model, different update cadence. The team couldn't add a new carrier in less than 8 weeks.
What we built.
We built a unified visibility platform: a normalised data model across all 12 carriers, a streaming pipeline (Kafka) for real-time status updates, an ETA prediction model (gradient-boosted trees on historical lane data), and a customer-facing dashboard.
The carrier integration framework uses a plugin architecture — each carrier is a Go package implementing a Carrier interface. Adding a new carrier dropped from 8 weeks to 2 weeks.
ETA prediction was the highest-leverage piece. The model runs every 5 minutes on every active shipment and pushes updated ETAs to the customer dashboard. Accuracy went from 62% (carrier-reported) to 94% (model-predicted).
Tech stack
What changed for the client.
- ETA accuracy 94% (up from 62%)
- Customer-reported exceptions down 71% (proactive alerting catches issues first)
- Carrier onboarding time 8 weeks → 2 weeks
- Status update latency under 2 minutes (was 15+ minutes for some carriers)
“We went from 62% ETA accuracy to 94% in 22 weeks. Customer service became proactive. The framework means new carriers take 2 weeks, not 8.”
VP Operations
Top-20 US 3PL
Want results like these?
Book a free 30-minute consultation. We'll review your situation and propose a small first milestone — same playbook as the case studies above.