The four diesel leaks hiding in your mileage average

Every fleet office we walk into has a mileage number. It is usually written on a whiteboard, it is usually a fleet-wide average for the month, and it is usually within a decimal point of the same number as last month. This is taken as a sign that things are under control.

It is more often a sign that the number is incapable of moving. A fleet average is a mean across vehicles of different ages, drivers of different habits, corridors of different gradients and loads of different weights. Almost any real problem — one vehicle developing a fault, one driver selling diesel, one corridor being run the expensive way — is a small enough fraction of that mean to disappear inside it.

Here are the four leaks the average is best at hiding, and the specific comparison that exposes each.

Leak one: the vehicle that is quietly getting worse

Mechanical degradation is gradual. An injector problem, a dragging brake, a clogged filter, a slow fuel-system leak — none of these announce themselves. They take a vehicle from 4.1 km/l to 3.8 km/l over six or eight weeks.

Across a forty-vehicle fleet, that single vehicle moves the fleet average by less than a hundredth. It is invisible. But on a truck covering eight thousand kilometres a month, 4.1 to 3.8 is roughly ₹18,000 a month in diesel, and it will continue until something breaks loudly enough to reach the workshop.

The comparison that finds it: each vehicle against its own rolling twelve-week history, not against the fleet. You are not asking “is this vehicle good?” — you are asking “is this vehicle worse than it was?” That question has a clean answer, and it is the one that catches faults while they are still cheap.

Leak two: the corridor that costs more than you charge for it

If your fleet runs a mix of terrain, your average is the mean of two or more distributions that should never have been combined. A hill corridor at 3.2 km/l and a plain corridor at 4.6 km/l average out to something near 3.9 — a figure that describes neither, and that will be used to price both.

This is how corridors end up underpriced for years. The rate was set using a fleet-average cost per kilometre. On the plain routes that rate is generous. On the hill routes it has never covered the diesel, and the loss is subsidised by the rest of the network so thoroughly that nobody can see it.

The comparison that finds it: mileage segmented by corridor, then costed per kilometre against the rate actually charged on that corridor. Most fleets doing this for the first time find at least one route that has been running at a loss since it was signed.

Leak three: pilferage that looks like variance

This is the uncomfortable one, and it is worth being precise rather than either coy or accusatory about it.

Diesel theft — siphoning, short-filling in collusion with a bunk, fills booked against a trip that did not happen — is real in Indian fleet operations, and it does not look like theft in the data. It looks like variance. A litre or two a fill, on a fleet doing hundreds of fills a month, sits comfortably inside the noise of a monthly average.

What it does not survive is being matched. Pilferage requires that a fill exist without a corresponding movement, or that a tank’s arithmetic not close. Both are checkable, and neither is checkable from an average.

The comparisons that find it:

• Fill against movement. Every fill should correspond to a trip that consumed roughly that much diesel. Fills that do not tie to a movement are the first thing to look at.
• Tank reconciliation. Opening stock plus receipts minus issues should equal closing stock. It never closes exactly; the question is whether the gap is stable or growing.
• Driver against driver on the same corridor. Not fleet-wide — same road, same vehicle class, similar load. Persistent gaps between drivers on identical work are worth a conversation, and are often about driving style rather than dishonesty. Both are worth knowing.

A note on tone: most variance is not theft. Driving style, tyre pressure, load distribution and route choice explain the large majority of driver-to-driver difference. The value of matching fills to movements is that it separates the two, so you stop suspecting everyone and start addressing the one thing that is actually happening.

Leak four: idling, which nobody bills to anyone

A loaded truck idling burns somewhere between two and four litres an hour depending on engine and ancillaries. Two hours a day of idle across a forty-vehicle fleet is, at current diesel prices, comfortably over ₹15 lakh a year.

Almost none of it is on anyone’s cost sheet. It is inside the mileage figure — the kilometres denominator does not grow while the litres numerator does — which means idling presents as poor mileage rather than as idle time. Fleets then go looking for a mechanical explanation for a problem that is entirely operational.

The comparison that finds it: engine hours against moving hours, per vehicle and per site. Once idle is separated out and attached to the location that caused it, two things follow. Mileage figures become interpretable for the first time, because they now describe motion. And waiting time at a customer’s plant becomes a line item in a commercial conversation rather than an invisible subsidy.

Idle hours attached to a site turn a standing complaint into a negotiable number. That is usually worth more than the diesel.

What to do on Monday

You do not need to solve all four at once, and you should not try. In order of return on effort:

1. Match every fill to a trip. This alone catches most of leak three and a good part of leak one. It requires only that fills and trips exist in the same place, which is largely a filing problem rather than a technology problem.
2. Segment mileage by corridor. One afternoon of work. It will tell you whether any route is structurally underpriced, which is the highest-value single finding on this list.
3. Track each vehicle against its own history. A rolling twelve-week baseline per vehicle, with an alert when a vehicle drops more than five per cent below its own norm.
4. Separate engine hours from moving hours. This one usually needs telematics, which is why it is last — but it is also the one that most changes how you talk to your customers.

The whiteboard number can stay. Just stop making decisions with it.