Starting point
A leading rail freight operator ran an IoT business unit with high complexity and an uneconomical cost base. The systems had grown organically, not by design: tracking accuracy too low for track-level positioning in marshalling yards and ports, IT operations expensive, sales lacking a clear market focus. The unit was under pressure to demonstrate its contribution to the group's bottom line.
What we did
We restructured the IoT unit along three dimensions: operationally (processes and resources), technically (migration to an auto-scaling Kubernetes cluster), and commercially (market-oriented sales strategy, end-customer app for mobile and desktop). In parallel, we developed a high-precision locomotive tracking solution in cooperation with DLR (German Aerospace Center) — EUR 250,000 in grant funding, tracking accuracy improved from 7.5 to 1.75 meters.
Results
80 %
Operating cost reduction
750.000 €
Follow-on funding secured
1,75 m
Tracking accuracy (from 7.5 m)
+25 %
Market penetration
What we learned
IoT units rarely fail because of the technology — they fail because of the business logic. Anyone looking to cut costs and grow market presence simultaneously must move operational, technical, and commercial levers in parallel — not sequentially. Sequential is the more expensive option.
This is the summary. How we approached it methodologically — which architectural decisions we made, what we discarded and which patterns can be transferred to other contexts — we discuss in a personal conversation.
Not because we want to sell you something. But because this depth is what our clients engage us for — and it does not belong on the open internet.
More case studies
Insurance
AI platform for claims assessment in insurance
AI platform for claims assessment taken from concept to integration approval in 12 weeks.
Read case studyFood & Ingredients
AI Portfolio Management at a Food Corporation
49 use cases identified and EUR 2-5m annual savings potential validated.
Read case studyInsurance
Disability Claims Review: AI PoC with 98% Extraction Accuracy
98% extraction accuracy across 187 document classes — validated technically, economically, and for regulatory compliance.
Read case study