In short
Weighbridge fraud shrinks when manipulation becomes visible. Send the reading from the indicator straight into the system with no hand-typing, write every correction to an audit log that cannot be erased, and let no vehicle leave without an exit permit linked to a ticket. The system raises flags rather than catching fraud on its own.
Weighbridge fraud is rarely one big loss. It is usually a small difference, repeated: two hundred kilos on a load, twenty times a day. Nobody notices it on the ticket, and it surfaces months later at stocktake as a shortfall no one can explain.
And most of the methods are not clever. They are simple enough that they only work because nobody checks. Here is what actually happens on site, and what stops each one.
1. Part of the vehicle off the platform
The oldest and simplest: the driver stops so that a rear wheel — or a whole axle on a trailer — stays off the deck. The display shows less than the truth, the ticket prints with the short figure, and everything looks normal.
What stops it: clear ground markings before and after the deck, and a barrier or sensor that blocks recording unless the vehicle is within bounds. In practice the biggest effect comes from a photo of the vehicle at the moment of weighing, attached to the ticket: once a driver knows their position is photographed and kept with the number, the attempt stops on its own.
2. The borrowed tare
Here the trick is in the empty weight, not the loaded one: the vehicle is recorded with an old tare, or with a lighter vehicle's tare, which inflates the net on inbound or shrinks it on outbound. No need to drive onto the scale at all.
What stops it: never accepting a hand-typed tare. Both weights — empty and gross — are captured from the scale in the same session, within a sensible time of each other. If you do allow a stored tare for regular vehicles, give it an expiry and flag it when it differs sharply from the last real reading.
3. Editing the ticket after it prints
The weight is recorded correctly, then the figure is changed in the notebook or the spreadsheet before it reaches accounting. This is the most dangerous one, because it leaves no trace: the final version is the only version.
What stops it: a record that cannot be overwritten. A correction has to be an addition, not a replacement — the old value and the new one, who changed it, when, and why. Editing stays possible, which it must be, but it becomes visible.
4. The ticket that was never recorded
The vehicle arrives, is weighed, and leaves — and no ticket is created at all, or one is cancelled after it has gone. The goods left and the record knows nothing about it. No audit of existing tickets will find this, because the problem is a ticket that does not exist.
What stops it: tying the gate to the weighing. Every vehicle is logged in at the gate, and none leaves without an exit permit linked to a ticket. Then one question exposes everything: which vehicle entered today and left without a permit? Sequential ticket numbers show any gap.
5. Collusion with a supplier or customer
The hardest, because every step looks correct: the weight is right and recorded, but the material written down is cheaper than the one actually moved, or the quantity is booked to another customer. There is no wrong number to catch.
What stops it: comparison, not inspection. Average load per vehicle, tickets per operator per shift, how materials distribute across customers — a departure from the pattern shows up in analytics before it shows up at stocktake. An outlier proves nothing, but it tells you where to look.
One principle underneath all of them
Every method above depends on the same thing: a moment where the number passes through a person before it is recorded. If there is one manual step between the scale reading and what reaches accounting, there is a gap.
Which is why the most effective control is not procedural but technical: the reading should travel from the indicator into the system directly, and everything after it should be written to a record that cannot be erased. That does not prevent human error — it makes both error and fraud visible, which is all you need.
Where to start if you run on paper today
Do not start with everything. Start with the two changes that cover most of it:
- Stop hand-writing the weight. Connect the indicator so no figure passes through anyone's hands. That alone closes the first two.
- Tie the exit to a permit. No vehicle leaves without a printed permit linked to its tickets. That closes the fourth — the one with the biggest impact, because nothing else finds it.
Documented corrections and analytics then tighten the rest. They are not the starting point.
One last thing, about the scale itself
All of the above assumes the scale gives a correct figure. If it is out of calibration you are recording an error precisely. Calibration is not an administrative formality — it is the ground every other control stands on, and it has its own article.