Fuel shrinkage: which variance is normal and which one is theft

Every fuel site runs some variance. Temperature changes volume. Meters drift between calibrations. Deliveries come in slightly over or under. Evaporation is real. Nobody reconciles to zero, and an operator who claims to is not measuring carefully.

The problem is that this normal, expected noise is exactly the place a theft can live undetected — sometimes for months. Below is how to tell the two apart.

Start with the honest baseline

You cannot spot an anomaly without knowing what normal looks like at your site. Before anything else:

• Reconcile book inventory to stick or ATG readings on a fixed cadence — daily if you can, weekly at the outside. Monthly is too slow to be useful.
• Track variance as a percentage of throughput, per grade, not as raw gallons. A 400-gallon variance means something different on 30,000 gallons than on 6,000.
• Keep at least 90 days of history, so you have a distribution rather than a single data point.
• Record it per tank and per grade, not site-wide. Site-wide totals average away the signal.

Most operators who get caught out did not lack the data. They had it and never plotted it.

The patterns that are worth a second look

Ordinary variance is noisy, symmetric and grade-agnostic. It wanders both directions and does not respect a schedule. Theft usually is not any of those things.

Variance that only goes one way. Real measurement error over- and under-states in roughly equal measure. Ninety days of losses with almost no offsetting gains is not measurement error.

A step change with no operational cause. Variance that was steady, then moved to a new level and stayed there. Ask what changed on that date — a new delivery schedule, a service call, a staffing change, or nothing at all. “Nothing at all” is the answer worth chasing.

One grade, especially diesel. Diesel is the target in most of the cases we cover, because it resells easily and moves in volume. Variance concentrated on a single grade while others stay clean is a strong signal.

A time-of-day or day-of-week shape. If you can resolve inventory overnight versus daytime, a loss that accrues consistently in the small hours means something different from one spread evenly.

Dispenser totals that disagree with tank drops. This is the important one for tampering specifically. If a dispenser’s metered volume is being under-counted, the dispenser will not report the discrepancy — it reports a normal sale. The tank knows what actually left. Reconciling dispenser totals against tank movement is the only place a metering discrepancy has to show up. If your process compares POS totals to dispenser totals, you are comparing two numbers that share the same source and will always agree.

Repeat “false” cabinet alarms on the same position. Log the nuisance ones. Clusters are data.

The scale problem

Here is why this is worth building a habit around. In the Bay Area case we covered, one crew hit at least 15 stations for more than 10,000 gallons total — an average of roughly 700 gallons per site. (CBS News Bay Area) At a high-volume location, across a month of deliveries, 700 gallons is comfortably inside the noise. Several of those operators may never have booked it as theft at all.

At the other end, the Covington, Tennessee co-op lost 18,000 gallons — $63,000 — from unattended pumps. (Action News 5) That one is not hiding in anybody’s variance. But it also did not happen in one night, and a site watching daily numbers has a chance to catch it partway through rather than at the end.

The gap between those two outcomes is almost entirely cadence.

If the pattern looks real

Do not start by confronting anyone or by opening equipment yourself. Tighten the measurement first: go to daily reconciliation, isolate by grade and dispenser, and pull camera footage for the windows where the loss accrues. Then read the first 24 hours after suspected pump tampering, which covers who to call and — importantly — what not to touch.

---

Want to see what a single incident is worth at your site? Run the numbers on the fuel theft cost calculator.

"We've never had a fuel theft" and "we've never identified one" are very different statements, and most operators can only honestly make the second. In one Bay Area case a single crew hit 15+ stations for 10,000+ gallons total — average around 700 gallons per site. At a busy location, spread across a month of deliveries, 700 gallons can sit entirely inside normal variance. It looks like temperature. It looks like meter drift. It looks like a light load. We wrote up the variance patterns that distinguish ordinary shrinkage from someone taking fuel, and what to do before it becomes a six-figure number. What does your reconciliation actually look like week to week?

https://fueltheftprevention.com/fuel-theft-news/fuel-shrinkage-red-flags?utm_source=linkedin&utm_medium=social&utm_campaign=fuel-theft-news

#fuelshrinkage #fueltheft #cstore #inventorycontrol

Every station runs some fuel variance. But some patterns aren't normal shrinkage — they're theft. Here's how to tell the difference before it turns into a big number.

https://fueltheftprevention.com/fuel-theft-news/fuel-shrinkage-red-flags?utm_source=facebook&utm_medium=social&utm_campaign=fuel-theft-news

"We've never had a fuel theft" usually means "we've never identified one." A 700-gallon hit can hide inside a month of normal variance. Here's what to look for.

https://fueltheftprevention.com/fuel-theft-news/fuel-shrinkage-red-flags?utm_source=x&utm_medium=social&utm_campaign=fuel-theft-news