Fuel is usually the largest controllable cost in a Kenyan transport business — and the easiest place to bleed money without noticing. Owners feel it long before they can prove it: the fuel bill grows faster than the work, the numbers never quite reconcile, and every conversation about it turns into an argument nobody can win.
One thing should be said honestly up front: no software makes a fleet theft-proof, and anyone promising that is selling you a story. What you can do is make fuel loss visible, fast to catch, and expensive to attempt — and that changes behaviour more reliably than any lock on a tank.
Most fleets police fuel the same way: collect the receipts, add them up at month end, and frown at the total. That approach cannot catch anything, for a simple reason — the unit of loss is a single fill, and the monthly total destroys exactly that information:
Knowing the patterns is not about suspicion — it is about knowing what a control system has to be able to see. The common ones in East African operations:
Notice what every pattern has in common: each one leaves a trace in the relationship between litres, money, odometer, and time. None of them survives contact with a per-fill record.
Make every fill answer for itself.
Kora logs each fill from the driver's phone — litres, price, station, odometer, receipt photo — and automatically flags excess fills, mileage mismatches, rapid refills, and suspected siphoning for review.
The control is not a gadget — it is a record with the right shape. Every fill becomes a small, checkable transaction:
The review queue is the part most fleets skip, and it is the part that matters. A flag is a question, not a verdict — sometimes the answer is a faulty odometer, a genuine price change, or a legitimate double stop. The system's job is to surface the question early with the evidence attached; a person decides what it means.
Measured fuel closes the gap paper cannot: what happened to the tank between fills. But the hardware picture is more nuanced than vendors admit:
A fleet without fuel records runs on suspicion, and suspicion lands on everyone — the honest majority included. A per-fill record works in both directions: it exposes the fill that does not add up, and it quietly clears every fill that does. Drivers who fuel honestly stop being accused by a vague monthly number, because their record answers for them. Framed that way — policy for every vehicle, not surveillance of any person — the discipline is one most crews accept quickly.
Start small and let the numbers make the case. The comparison that convinces an owner is not a vendor demo — it is the cost per kilometre of the pilot vehicles, before and after the record existed.
The reliable signals are in the numbers, not in suspicion: consumption that does not match distance covered, refills that come sooner than the kilometres justify, fills that exceed the space left in the tank, and receipts priced above the pump rate. The only fair way to act on any of these is a per-fill record — station, litres, price, odometer, receipt — reviewed case by case, so a conversation starts from evidence instead of accusation.
No system makes theft impossible — the goal is to make fuel loss visible, fast to catch, and unattractive to attempt. That means logging every fill at the pump with an odometer reading, reconciling litres against tank capacity and distance driven, flagging excess fills, mileage mismatches and rapid refills automatically, and reviewing every flag while the trip is still fresh. Fleets that move from monthly totals to per-fill records typically find the losses were hiding in specific, provable fills.
It depends on the vehicle and the sensor type. CAN-connected telematics read odometer and engine data well on most common makes, but factory fuel-level data over CAN is unreliable or absent on many trucks operated in Kenya. External tank-level sensors fitted to the tank are the more dependable route to measured fuel on most vehicles — they can detect sudden drops while parked, which is the signature of siphoning. Confirm compatibility per vehicle model before buying.
There is no single normal figure — consumption depends on the vehicle class, load, route profile, and condition. The useful benchmark is the vehicle’s own history: a truck that suddenly consumes materially more per kilometre than its own baseline, on the same route and load, is the signal worth investigating. Fleet software builds that baseline automatically from per-fill records.
GPS alone does not — it shows where a vehicle is, not what happened to its fuel. Position history helps explain route padding and unexplained stops, but catching short fills, inflated receipts, and siphoning requires fill records reconciled against odometer readings and tank capacity, plus measured tank levels where sensors are fitted. Tracking is one input; the fuel record is the control.
Visibility instead of suspicion
Kora gives Kenyan transport teams per-fill logging with receipt photos, automatic anomaly flags, a review queue with evidence attached, and tank-sensor support where fitted — so fuel stops being an argument.
