A Queensland service station operator was ready to call the police. Every second fuel delivery was landing 200 to 400 litres short, the pattern was too regular to ignore, and the obvious suspect was the tanker driver.
The data said otherwise. Theft is rarely that tidy: same shortfall, same size deliveries, like clockwork. When the reconciliation compared dips before and after each delivery against the invoiced volumes, the losses only appeared on large drops, and they scaled with the size of the drop. That’s not a thief. That’s maths.
The real culprit was the tank’s calibration chart, the table that converts a dip in millimetres into litres. It was outdated, off by nearly two per cent through the middle of the tank. Every big delivery pushed the fuel level through the miscalibrated band and manufactured a phantom loss on paper. The chart was corrected, the “theft” stopped instantly, and an innocent driver kept his reputation.
This is the uncomfortable truth about tank monitoring: your measuring equipment lies to you a little every day, and the lies look exactly like losses. Three places it happens:
The calibration chart. Tanks shift and settle over decades. A chart made when the tank was installed may not describe the tank you have now. The tell: variances that appear at particular fuel levels or on large deliveries, rather than steadily.
The dip itself. A dipstick read at a slight angle, a different person reading the meniscus differently, a tank dipped at 6am one day and 9pm the next. Each small inconsistency creates noise, and enough noise can hide a genuine slow leak inside it. The tell: variances that spike and reverse, that follow rosters rather than physics.
The dispenser meters. A meter over-pouring by half a per cent creates a real loss (you’re giving fuel away) that looks statistical. The tell: losses that track your sales volumes precisely.
The job of proper reconciliation is stripping this measurement noise out so you can see what’s left. Correct for the noise and one of two things remains: nothing, in which case your equipment needed attention and your tank is fine, or a steady, direction-holding trend, which is when you take a leak seriously.
Before anyone accuses a driver, digs up a forecourt or pays for a tank test, interrogate the measurement first. It’s the cheapest suspect to eliminate, and in our experience it’s guilty more often than not.
