Turning a Mixed Vehicle Fleet Into One Set of Numbers

The challenge

The fleet spanned several manufacturers, and the operational data was split accordingly — telemetry in each maker's own portal, service history and costs in the shop management system. Nobody could see true cost of ownership for a vehicle, so maintenance stayed reactive and decisions about replacing a vehicle were made on instinct.

Where we came in

The trap in fleet analytics is building against whichever telematics provider you have today, which guarantees a rebuild the next time the fleet changes. We made the architecture data-source agnostic from the start and opened with a discovery phase on the hardest integration, so the shape of the problem was known before the pipeline was committed to.

What we did

• Opened with a discovery phase on the service-records integration to establish feasibility before committing to the pipeline
• Stood up production cloud infrastructure, then modelled the warehouse schema as a distinct phase
• Built an ETL pipeline plus a real-time webhook listener, so the platform takes both scheduled and event-driven data
• Kept the ingestion layer source-agnostic, currently integrating two telematics providers
• Delivered reporting in a tool the business already uses rather than a bespoke dashboard
• Scoped an insights layer on top of the warehouse once the data foundation was proven

Where it landed

• Telemetry, maintenance and cost data land in one analytics environment rather than three portals
• Cost of ownership can be reported per vehicle instead of estimated across the fleet
• The ingestion layer is provider-agnostic, so changing telematics vendor does not mean rebuilding the platform