Start by assuming it is not theft
This is not naivety, it is arithmetic. At a typical outlet the four largest sources of unexplained loss are evaporation, delivery temperature, dispenser calibration drift and bookkeeping — and all four are bigger, on average, than theft.
The practical consequence is that acting on suspicion before ruling those out is expensive twice over. It costs you the actual loss, which continues, and it costs you the trust of staff who were not responsible. An outlet where everyone knows they are suspected is an outlet where nobody tells you when something genuinely goes wrong.
So the sequence below is not squeamishness. It is the order that gets you to the answer fastest.
The one thing that makes all of this possible
Every check here compares two records that were made separately — meters against dip, litres against cash, invoice against tank gain. If your outlet only keeps one of each pair, none of them is available to you, and no amount of watching people will substitute.
The six ways it actually happens
Four involve fuel, two involve money. Each leaves a different signature in a daily reconciliation.
| What | Where it shows | Signature |
|---|---|---|
| Evaporation | Stock variance | Small, steady, negative, worse in summer |
| Warm delivery | Stock variance | A step down after a delivery on a hot day |
| Calibration drift | Stock variance | Proportional to litres on one nozzle only |
| Fuel removed | Stock variance | Not proportional to volume; often overnight |
| Cash short | Cash reconciliation | Shortfall on the cash line, one shift |
| Unrecorded sale | Both | Stock down, cash flat — the only one that moves both |
How each loss shows up in the daily numbers.
The one signature nothing else produces
An unrecorded sale is the only loss on that list that moves the stock number and the cash number in opposite directions. Fuel left the tank, so book stock says you should have more than you do — and no money arrived, so collections are lower than the litres imply.
Everything else moves one side or the other. Evaporation, temperature and drift move stock and leave cash untouched, because the sales that were recorded really happened. A short drawer moves cash and leaves stock untouched, because the fuel really was sold.
That is why a pump that reconciles only cash, or only stock, can be losing fuel continuously and see nothing unusual on either. The signal exists only in the relationship between the two.
Ruling out the innocent explanations
In this order, because each is cheaper than the one after it.
- Temperature first. Correct every delivery to 15 °C — the density and temperature calculator does it — before comparing invoiced quantity against your tank gain. A summer shortage that vanishes under the correction was never a shortage.
- Then per-nozzle proportionality. If the loss tracks litres dispensed on one specific nozzle and is absent on its neighbours drawing from the same tank, that is calibration drift. Get it stamped.
- Then the delivery log. A gain that appears and then corrects itself over the following days is almost always a dip taken before the tank settled, or a delivery logged late.
- Then the seasonal curve. Evaporation tracks temperature and volume. A loss that is worse in May than in January and scales with throughput is physics.
- Then credit. A sale on account looks exactly like missing cash if it was never written up. This is the single most common cause of a persistently short drawer at outlets with commercial customers.
What is left after that
A loss that survives all five is characterised by what it is not. It is not proportional to volume — it happens on quiet days as much as busy ones. It is not seasonal. It does not track any single nozzle. It does not appear on delivery days in particular.
Two things fit that description: a leak, and fuel being removed. A leak grows steadily and does not care about your opening hours. Removal tends to cluster — particular shifts, particular days, or overnight — because it requires someone to be there.
Which is why the timing pattern matters more than the size. Plot the daily variance against shift, and against day of week. A loss that is flat across all of them is a leak and needs a tightness test today. A loss concentrated in a repeating slot is a different conversation.
Test the tank before you accuse anyone
A tightness test costs a fraction of what a slow leak loses in a season, and vastly less than the environmental liability if it reaches groundwater. It also definitively removes the most serious innocent explanation, which is worth doing before a conversation you cannot take back.
The cash side
Cash shortfalls are easier to attribute and much easier to prevent, and almost all of the prevention is structural rather than supervisory.
- Reconcile per shift, not per day. A shortfall found at close belongs to people who are still on the premises; the same figure at month-end belongs to nobody.
- Count the drawer before looking at what it should be. Somebody who knows the target first is not counting, they are confirming.
- Record drawer payouts as they happen. An expense nobody logged is indistinguishable from money taken, and it is far more common.
- Keep the five payment types apart. A UPI transaction that showed success and later failed is not a person — but pooled into one collections figure it looks exactly like one.
- Log credit sales at the pump, not at the end of the day from memory.
What not to do
Three responses that feel decisive and make the problem harder to solve.
- Do not start with cameras. They tell you what happened at a moment you already have to suspect. The numbers tell you which moment to look at, and without them footage is just hours of forecourt.
- Do not rotate staff to see if the loss follows. It is slow, it is visible to everyone, and a loss that is really a drifting nozzle will appear to follow whoever happens to be on that pump.
- Do not act on a monthly figure. A month collapses thirty days of signal into one number that any of six causes could have produced. Every diagnostic on this page needs the daily series.
The honest summary
The reason theft is hard to find at a fuel outlet is not that it is well hidden. It is that a pump losing fuel and a pump losing nothing look identical in the records most outlets keep — a stack of meter slips, a cash total and a monthly summary.
Reconcile daily, per tank and per shift, and the six causes separate themselves within a few weeks without anyone having to watch anyone. That is the entire method, and it works because the signal was always there. Doing it by hand is a register and a habit; doing it without the habit is what pump software is for.
Frequently asked questions
How much stock loss is suspicious?
The size of the loss tells you very little; its shape over time tells you almost everything. A steady small negative that worsens in summer is evaporation at any magnitude. A loss that is flat across volume and season is worth investigating at any magnitude. Reconciling monthly destroys the shape and leaves you with only the size — which is why it is the wrong tool for this question.
Can software prevent fuel theft?
No software prevents anything on a forecourt. What it does is make the six causes distinguishable within days instead of never, by comparing records that were made separately — meters against dip, litres against cash. Most outlets discover their loss was calibration or temperature, and the ones that do not now know where and when to look.
Should I install CCTV?
Cameras are useful once the numbers have told you which shift and which pump to look at. Installed first, they mostly produce hours of footage nobody watches, and they signal suspicion to the whole staff without producing an answer.
My cash is short every single day by a small amount. Is someone taking it?
Possibly, but check three things first, in this order: credit sales being recorded late or not at all, UPI payments that showed success and later failed, and drawer payouts that were made but never logged. Between them those three account for the large majority of persistent small shortfalls at outlets with commercial customers.
What if the loss stops when I start reconciling daily?
That happens, and it is worth being clear about why it is not proof of anything. Daily reconciliation changes several things at once — attention, recording discipline, and the delay between an event and someone noticing it. A loss that was really unlogged expenses stops for the same reason a loss that was really theft would.