Fuel Fraud: How to Detect and Prevent Unauthorized Fuel Purchases in Your Fleet

Fuel fraud covers any fuel-related spend that violates policy or intent. The obvious cases: stolen or cloned cards, "buddy fill-ups" for vehicles that aren't yours, purchasing premium when regular is specified, adding gift cards or merchandise to a fuel transaction, falsified odometer entries, and swipes that happen far from where the vehicle actually was. It also includes subtler behaviors that blur the line between mistake and misuse — topping off several times a day to collect loyalty points, fueling outside business hours without a reason, or buying diesel or diesel exhaust fluid (DEF) for a gasoline vehicle.

A modern fuel-fraud program has three layers that work together:

• 1. Controls — rules on the card itself that prevent bad transactions from happening.
• 2. Analytics — signals that detect whatever slipped past the controls.
• 3. Response — a fast, fair process for investigating cases and recovering money.

The objective isn't to create friction for honest drivers. It's to remove ambiguity, keep the books clean, and turn thousands of small, frequent purchases into a dataset your cost and sustainability reporting can actually trust.

Why it matters in fleet management

Margins

Fuel is often the largest variable expense on a fleet's P&L. Fraud is a direct leak from it, and because individual transactions are small, the leak can run for months before anyone notices the total.

Data integrity

Fraudulent fills distort the KPIs you steer by — cost per mile, fuel efficiency, route economics. Decisions built on polluted data are wrong even when the analysis is right.

Back-office load

Every disputed transaction consumes reconciliation time, and chargebacks drag on for weeks. Tight controls mean fewer disputes to fight in the first place.

Driver trust

Unclear rules and inconsistent enforcement erode trust faster than fraud itself. A transparent program with fair exception handling protects the honest majority — and clean fuel data then feeds better coaching on idling, route design, and vehicle specification, benefits that reach well beyond security.

Example in practice

A regional distributor notices one fuel card averaging three fills per day, with occasional purchases of 95–110 liters — despite the assigned van having a 75-liter tank. Telematics shows the van was 18 km away from the station during several of the swipes.

The system auto-flags both anomalies — a capacity mismatch and a location mismatch — suspends the card, and opens a case. Station CCTV confirms a cloned card. The card issuer processes chargebacks, and the fleet tightens its controls: driver PINs are rotated, a per-fill liter cap is enforced, a 5 km corridor geofence (a virtual boundary around approved routes) is added, and weekend blocks are switched on for that depot.

In the next quarter, off-hours fuel spend drops 82%, reconciliation time is cut in half, and the team publishes cost-per-route metrics to leadership with confidence in every number behind them.

Prevention controls

Bind cards to vehicles and drivers. Issue vehicle-specific cards paired with unique driver PINs, and avoid generic spare cards — an unassigned card is an unaccountable one.

Set sane limits. Per-fill liter caps, daily and weekly spend caps, a maximum number of transactions per day, and merchant-category restrictions that allow fuel purchases only. Block premium grades unless a specific vehicle requires them.

Time and place rules. Business-hours windows, country or region locks, and geofences around depots and route corridors, with small grace buffers so a station just off the highway doesn't trigger a false alarm.

Required prompts at the pump. Odometer reading, vehicle ID, and optionally a job or route code. Prompts discourage card sharing and feed the data quality your analytics depend on.

Network discipline. Prefer merchants that report rich transaction data — product code and exact volume — and negotiated pricing. Block cash-like add-ons and gift cards outright.

Lifecycle hygiene. Automatically deactivate cards when a driver leaves, when a card is reported lost or stolen, or after prolonged inactivity. Rotate PINs periodically.

Telematics pairing. Validate a swipe only if the vehicle's GPS places it at the station within a defined time window. This single check eliminates most cloned-card and buddy-fill scenarios.

Detection analytics

Capacity mismatch — liters purchased exceed the tank size plus a tolerance (for example, more than 105% of capacity) or exceed a realistic fill rate.

Velocity checks — too many transactions in a short window, back-to-back fills at different stations, or frequent top-offs when the tank is still above 75% full.

Location mismatch — the card was swiped while the vehicle was more than a set distance from the station, or off its planned corridor.

Product anomalies — premium or diesel on vehicles specified for regular gasoline, unusual DEF purchases, or non-fuel items added at the counter.

After-hours spikes — nights, weekends, or fills outside scheduled shifts without an approved exception.

Pattern mining — the same driver ID appearing across multiple vehicles, repeat merchants charging above-index prices, or driver cohorts whose fuel efficiency trends look wrong after their reported fills.

Telematics corroboration — idle and fuel-rate signals that don't match reported fills. If fuel efficiency improves unrealistically after a "fuel-only" transaction, the fuel likely went somewhere other than the tank.

Response playbook

Triage fast. Auto-suspend the implicated card, contact the driver to validate the transaction, and — if it can't be confirmed — escalate to the card issuer and merchant with evidence: timestamps, GPS history, and photos.

Document everything. Keep a single case file per incident: receipts, card logs, telematics snapshots, CCTV requests, driver statements, and all issuer correspondence. A complete file is what turns a suspicion into a recovered chargeback.

Recover and remediate. Pursue chargebacks (reversing the transaction through the issuer) or subrogation (recovering costs from the responsible party) where applicable. Then close the gap that let it happen: rotate PINs, adjust limits, or change preferred stations in the affected region.

Coach vs. discipline. Distinguish error from intent. An honest mistake calls for retraining on policy and cab setup — a phone mount for odometer capture solves more data problems than a warning letter. Willful misuse calls for progressive discipline under a policy everyone has seen in advance.

Close the loop. Publish quarterly "fuel wins" — fraud caught, dollars recovered, rules tightened. Visible results are what keep drivers, managers, and finance invested in the program.

KPIs to track

Fraud rate — the percentage of fuel spend flagged and confirmed as fraudulent, tracked by depot, route type, and merchant.

Mean time to detection and response — hours from swipe to flag, and from flag to resolution. Faster detection means smaller losses and better evidence.

Recovery rate — dollars recovered divided by confirmed fraudulent dollars.

False-positive rate — the share of flags that turn out legitimate. Keep it low: every false accusation costs trust, and every ignored alert trains the team to ignore the next one.

Off-policy rate — transactions that break time, place, or product rules without being fraudulent. Each one is an opportunity to fix a process or clarify a policy before it becomes a loss.

Data quality score — the percentage of transactions with a valid driver, vehicle, odometer, and location. This is the foundation every other number stands on.

Implementation tips

Start with the 80/20. Connect fuel cards to telematics, enable the capacity, location, and after-hours checks, and route only the exceptions to humans. Three automated checks catch the large majority of fraud.

Standardize identifiers and units. Consistent vehicle, driver, and card IDs — and consistent miles/kilometers and gallons/liters — prevent mismatches that look like fraud but are really data entry.

Use clear exception codes. PTO fueling, storm response, and remote jobs are legitimate outliers; code them so honest drivers aren't penalized for doing their job.

Review the merchant mix quarterly. Migrate volume toward networks with better data quality and pricing.

Pair analytics with scorecards lightly. Reward trend improvement — clean months, accurate odometer entries — rather than raw rankings. The goal is better habits, not a leaderboard.

Run a retro after every confirmed case. Adjust limits, geofences, and training based on what actually happened, and share two concrete changes each time. It shows the program learns.