Fuel is typically 30–40% of a transport fleet’s operating cost. It is also the line where a gap between the invoice and what actually entered the tank is easiest to create and hardest to prove without data.
In practice fuel monitoring answers three questions: how much fuel genuinely went into the tank, how much left it outside normal operation, and which vehicle or driver sits outside the fleet’s own normal range.
Three sources of fuel data
Capacitive sensor in the tank. The most accurate method and the standard in trucks and construction machinery. It measures level independently of the vehicle’s own electronics, at a resolution that catches losses of a few litres. Requires installation in the tank and calibration.
CAN or OBD reading. The tracker reads the factory fuel level and engine parameters — sufficient in newer cars and light commercial vehicles. The advantage is that nothing is fitted inside the tank; the limitation is the accuracy of the factory gauge, which on large tanks can be tens of litres out.
Fuel card transactions. Refuelling data from the card operator’s system. On its own it only tells you what was paid. Value appears when it is set against the measured increase in tank level.
Most fleets end up with a combination: CAN where it is good enough, sensors on heavy vehicles and on any vehicle where a discrepancy has already shown up.
What the system catches
Siphoning. A sharp drop in level with the engine off, away from a filling station. The alert carries time, location and volume.
Refuelling smaller than the transaction. The card shows 300 litres, the sensor records an increase of 240. The difference may be a calibration error — or fuel going into a different tank.
Refuelling off-route. A transaction in a place the vehicle never visited. The simplest abuse pattern to detect.
Excessive idling. An engine running while parked burns fuel without covering distance. In refrigerated and construction fleets this can be a low double-digit share of total consumption — and it is the share that responds fastest to a change in habits.
Outliers against the fleet norm. Consumption per 100 km across identical vehicles surfaces the units that need servicing and the drivers worth a conversation about driving style.
From data to savings
Measurement on its own saves nothing. The effect appears when the data enters a management routine:
- A weekly deviation report landing with one person who has the authority to call the driver.
- Monthly reconciliation of transactions against invoices as a fixed part of period close, not a quarterly exercise.
- An eco-driving ranking tied to a bonus — the most effective mechanism we have seen in deployments, provided the metrics are visible to drivers.
We cover the wider method in the article on cutting fleet costs with telematics.
Calibration is not optional
An uncalibrated sensor produces false alarms, and a month of false alarms teaches the team to ignore notifications. That is why calibration is part of the rollout rather than an add-on.
The procedure: controlled incremental filling, recording the sensor curve in the platform, then two to four weeks of observation with threshold correction. Only after that do alerts go to the operations team — before then they come to us.
Feeding the numbers into billing
Fuel data is worth something when it reaches the place where costs are counted. Reports export to Excel, and in larger fleets we push them automatically into an ERP or data warehouse through the Wialon platform API. Architecture options are on the Wialon integrations page.
Start with one vehicle
The standard starting scenario: the vehicle you have the strongest suspicions about, plus a reference unit with a similar duty cycle. A month of parallel data is usually enough to size the problem across the fleet and calculate the return. We prepare a quote covering devices, installation and calibration within 24 hours.