Connected Vehicle & Fleet Operations

Four capabilities disappear with the data. Service stops being predictable, because fault codes and mileage are unknown until the customer complains. Warranty stops being adjudicable, because 'serviced on schedule' cannot be evidenced. Fleet customers cannot be served, because they buy on cost per kilometre and uptime. Residual value stops being defensible, because condition and mileage are read off a dashboard at resale.

1. Service after the fact. Without mileage and fault history, maintenance can only be reactive — and by the time the customer calls, they are already unhappy.
2. Warranty decided by negotiation. Claims depend on service records that do not exist in a usable form.
3. Fleet buyers excluded. Ride-hailing, leasing and logistics evaluate on cost per kilometre and uptime; without those numbers there is no conversation.
4. Residual value guessed. Buy-back and trade-in pricing without continuous condition data has to be conservative, and conservative pricing loses the deal.

There is also a constraint that has to be settled before any of this: vehicle data includes location and usage behaviour. Consent is a design precondition, not a technical obstacle — what is collected, who may see it and how long it is kept are agreed at connection time, and the system treats consent as a precondition of ingestion rather than a form to fill in later.

The second practical reality is that connectivity differs sharply by market. Some vehicles ship connected, some need a retrofit terminal, some have expensive data tariffs. That is why ingestion is layered rather than solved one way.

<strong>Factory interface.</strong> Most complete, and dependent on the manufacturer. No retrofit cost, high data quality — but the scope is set by the OEM and differs by model. This line adapts per model and does not claim universal coverage.

<strong>Aftermarket terminal.</strong> Controllable, with ongoing cost. Suits fleets, leasing and used vehicles, and covers older models; terminal and data tariff are recurring costs, and installation quality affects data quality, so terminal model and installation date must be recorded.

<strong>Manual or workshop entry.</strong> The least fashionable layer and the most reliable one. Workshop visits give mileage and work orders, which is the strongest evidence available for a warranty decision — stronger than any remote reading.

Downtime is the item that changes conclusions. A ride-hailing vehicle off the road for three days loses the repair cost and three days of revenue — which is why fleets pay a premium for fewer stops, and why a TCO that omits it systematically undervalues reliability.

Two rules keep the analysis honest. Report TCO at three granularities — per vehicle, per model, per fleet — because each answers a different question, and a single average answers none. And when data is incomplete, state the coverage rather than presenting a partial figure as a complete one: remote coverage at 60% with the rest from work orders is a legitimate result, as long as it says so.

Three hard rules. Route replay is authorised on demand and logged — who looked at which vehicle for which period. A fleet customer sees only its own vehicles, and vehicles returned at end of lease leave the visible set automatically, or the customer keeps watching someone else's movements. And where behaviour data is used for driver assessment, drivers are informed in advance: that is both a compliance requirement and the only way the assessment works, because an uninformed driver will find a way around the device.

1. Collecting first and consenting later. Being technically able to read a value is not permission to collect it. Collection scope must match the consent scope, and retroactive consent is not available in most markets.
2. Alerts reported and never handled. A system that raises two hundred alerts a day and handles twenty trains everyone to ignore it — including the serious ones. Closure rate matters, alert count does not.
3. Mixing sources without quality marks. Factory data, terminal data and manual entry differ in reliability. Without a quality mark, conclusions look precise and are wrong.
4. TCO as a single average. Per vehicle, per model, per fleet — all three are needed, or the number answers nothing.
5. Ignoring downtime. For operating vehicles it often costs more than the repair. Leaving it out understates TCO systematically.
6. Routes visible by default. Location reveals routine and residence. Replay on demand, with a log, is the floor, not an option.

Failure is easier to notice than distortion, and distortion is more expensive: a system that obviously does not work gets fixed, while one that produces confident wrong numbers gets used for decisions.

1. For each vehicle in service, record whether position, condition and behaviour data are available.
2. Check whether workshop records carry mileage and work-order detail — the strongest fallback evidence.
3. Compute cost per kilometre for three operating vehicles and see which of the six components cannot be filled.
4. Confirm how consent and notice are handled today.
5. Walk through alert closure and the TCO view in the demo environment.

• After-sales repair & warranty
• Overseas 4S retail & showroom
• Country packs & overseas delivery