How Diesel Theft Happens in Indian Fleets (2026)

The simplest method and the most common. Fuel is drawn from the tank while the vehicle is stationary, typically overnight at a dhaba, a parking area or a yard.

Volumes are usually modest per event, twenty to fifty litres, because the objective is to stay below the noise floor of normal consumption variation. Repeated across a fleet across a year, modest becomes substantial.

A rapid fall with the engine off and the vehicle stationary is a reason to investigate. Check slope, sensor faults, leaks, calibration, and any authorised transfer before attributing the change to siphoning.

Here the fuel never reaches the tank. The receipt says one hundred litres, ninety go in, and the difference is settled between the driver and the attendant.

This is harder to detect from fuel bills alone because the paperwork is internally consistent: the invoice matches the payment and the entry in the register.

Comparing the receipt with the measured refill can reveal a discrepancy. A ninety-litre sensor rise against a hundred-litre receipt needs investigation, including tank calibration, measurement error, vehicle position, and dispensing evidence; it is not proof of fraud by itself.

A stationary engine burns fuel. Long idling for cabin cooling, at loading points, in queues, or simply from habit, produces consumption with no distance to show for it.

Operators frequently discover that idling accounts for more loss than siphoning does. It is entirely legitimate, entirely invisible on a fuel bill, and considerably easier to reduce than theft.

It is also the fastest win: reporting idle hours by vehicle and driver, and acting on the outliers, usually cuts it substantially without confrontation.

Extra distance means extra fuel, and the fuel is legitimately consumed. The trip was not.

Detours for personal errands, unofficial passenger or goods carriage, and vehicles used outside working hours all appear as ordinary consumption unless distance is compared against planned route.

This is why fuel data alone is insufficient. Consumption without route context cannot distinguish a heavy load on a hilly route from a driver making an eighty-kilometre detour.

A fuel bill records what was purchased. It says nothing about what entered the tank or what the vehicle did with it.

Reconciling bills against kilometres gives an average, and averages absorb exactly the variation you are looking for. A vehicle running at 3.8 km per litre against a 4.2 benchmark looks like a maintenance issue, a load difference or a route difference. It can equally be forty litres a month walking away.

Level data plus location is what separates those explanations, because it shows when and where fuel left the tank rather than only how much was bought.

In rough order of effect for effort.

• Alert on rapid drops while stationary. This catches siphoning as it happens rather than at month end.
• Match receipt volume against sensor-recorded fill volume. This catches short-filling, which bills alone never will.
• Report idle hours by vehicle and driver and act on outliers. Largest and easiest saving in most fleets.
• Compare actual against planned route distance to surface deviation.
• Tell drivers the system exists. Deterrence outperforms detection, and a fleet that quietly monitors catches individuals while a fleet that openly monitors reduces losses.

The last point matters commercially. The aim is lower fuel cost, not a disciplinary process.