Five causes, one symptom
Every one of these presents identically in your register: a negative variance. They are not equally serious, not equally controllable, and not equally common — but the number on the page cannot tell them apart. Only the pattern over weeks can.
| Cause | Typical shape | How controllable |
|---|---|---|
| Evaporation | Small, steady, worse in summer | Partly — vapour recovery, tank condition |
| Temperature at delivery | Appears after a delivery, warm days | Yes — correct to 15 °C at decantation |
| Calibration drift | Proportional to litres, one nozzle | Yes — stamping and recalibration |
| Decantation loss | Step at a delivery, then normal | Yes — procedure at unloading |
| Leak | Negative and growing, volume-independent | Urgent — tightness test |
The five sources of stock loss and how much control you have over each.
The largest recoverable one is temperature
This is the loss most outlets never identify, because it does not feel like a loss — it feels like a normal delivery followed by an unremarkable week of small negative variances.
Fuel gauged warm in a tanker on an Indian summer forecourt occupies genuinely more space than the same fuel once it has cooled to underground temperature. You were invoiced for litres measured hot; your dip measures them cool. The difference is arithmetic, not theft, and it recurs on every warm delivery.
The fix is to correct both the invoiced quantity and your tank gain to 15 °C before comparing them. Anything that survives that correction is a real shortage, raised the same day while it is still a measurement rather than a memory.
Record the temperature, always
A density reading without its temperature cannot be corrected by anyone afterwards, including you. It is one extra number on the delivery record and it is the difference between a shortage you can demonstrate and one you can only assert.
Calibration drift hides in a nozzle, not in a tank
A dispenser that over-delivers gives away fuel one transaction at a time. The customer gets slightly more than they paid for, the totaliser records what was actually dispensed, and the tank empties faster than the sales say it should.
It is identifiable because it is proportional. The variance tracks litres dispensed on that specific nozzle, and it is absent on its neighbours drawing from the same tank. That signature is why per-nozzle records matter — a pump that only records tank-level movement can see the loss but never locate it.
It is also why a variance that appears on a low-volume day is much more interesting than one on a busy day. Drift cannot produce a large loss on a day that moved almost nothing.
Decantation losses are procedural
Fuel left in a compartment, a spill, a nozzle dispensing from the tank during unloading, or a dip taken before the tank settled — all of these produce a one-off step in the register at a delivery, followed by normal days.
They are the easiest category to eliminate, because every one of them is a habit rather than an equipment condition.
- Stop dispensing from the receiving tank during decantation.
- Watch each compartment empty completely.
- Let the tank settle before the post-delivery dip — a dip taken too early reads high and manufactures a loss over the next few days.
- Record which tank each compartment went into.
And the one that is not ordinary
A leak is the only cause on this list that grows. The variance is negative every day, larger every day, and — critically — largely independent of how much fuel you moved that day. A quiet Sunday showing a bigger gap than a busy Saturday is the signature.
That independence is what separates it from everything else here. Evaporation tracks volume and temperature. Drift tracks litres on a nozzle. A leak tracks time.
It is also the only one where the correct response is not analysis. A tank and line tightness test costs a fraction of the fuel a slow leak loses in a season, and vastly less than the environmental liability if it reaches groundwater.
What actually reduces the number
In rough order of how much loss each recovers for the effort involved.
- Reconcile daily, per tank — by hand, or with petrol pump management software. Not because reconciling saves fuel, but because nothing else on this list is identifiable without a daily series.
- Correct deliveries to 15 °C with the density calculator. The largest recoverable category at most outlets.
- Keep per-nozzle records so drift can be located rather than merely observed.
- Fix decantation procedure. Free, and it eliminates an entire category.
- Get dispensers stamped and recalibrated on schedule, not when someone complains.
- Test the tank the first time the variance grows independently of volume. Not the third time.
Frequently asked questions
What is a normal evaporation loss for a petrol pump?
There is no honest universal figure — it depends on throughput, climate, tank condition and whether vapour recovery is fitted. Petrol evaporates considerably more readily than diesel. What matters is that your outlet knows its own normal band and can see the week it leaves that band.
Can I claim stock loss from my supplier?
A delivery shortage, yes — raised the same day, in writing, with the pre-delivery dip, the post-settlement dip and both quantities corrected to 15 °C. Ongoing evaporation from your own tanks is your cost, not theirs.
My variance is positive. Is that good?
No, it is a recording error. Fuel does not appear from nowhere. A persistent gain almost always means a delivery was logged late, a dip was taken before the tank settled, or a nozzle reading was missed so dispensed litres are understated. It will correct itself into an apparent loss later.
How quickly does a slow leak become obvious?
Slower than most people expect, which is the danger. A leak losing a few litres a day sits inside the noise of ordinary variance for weeks, and only becomes unmistakable once it has grown. It is visible far earlier in the trend than in any single day's figure — which is the entire argument for reconciling daily.