The fuel invoice does not know which load burned it

The fuel card invoice arrives on the fifth of the month. It is long: every fill-up by every truck, listed by card, by station and by date. The owner looks at the total, compares it with last month’s total, and signs it off. It is correct, in the sense that every litre on it was really bought.

The same afternoon the dispatcher is asked how the big Thursday job went, the one with the difficult customer and the tight rate. The answer is the one every dispatcher gives: we charged this, the distance was that, so at our usual cost per kilometre it came out fine. Nobody in the room knows what that truck actually burned on that trip, and the invoice that could tell them is sitting on the desk with the total already signed.

The cause is not carelessness. Fuel is bought per card and per month, while money is earned per load. The two records were never joined, so every margin figure in the company is built on an average that no individual load ever matched.

Where the average goes wrong

The average is the whole fleet, not this truck. A fleet rate per kilometre blends the newest tractor with the oldest, the driver who coasts with the one who brakes late, and the lane that is flat with the one that climbs. It is accurate for the fleet and wrong for every truck in it.

Loaded and empty kilometres get the same number. A full trailer over hills burns one thing, an empty one on the motorway another. A load that was heavy out and empty back is costed as if it had been average in both directions.

The fill-up is not where the fuel was used. A tank filled on Monday carries into Tuesday’s job and part of Wednesday’s. Matching a fill-up to the load that happened the same day gets it wrong in both directions and nobody notices because the error cancels in the monthly total.

Waiting time burns fuel that no load pays for. Engines run while a truck stands at a dock in winter. That fuel appears on the invoice, disappears into the average, and is never attached to the customer whose dock it was.

Good customers subsidise bad ones. When every load is costed at the same rate, the lanes that genuinely cost less look no better than the ones that cost more. The price given on the phone next time is the same for both, and the loss on the second is paid by the margin on the first.

The number nobody checks is the one that moves. Fuel is the cost that changes week to week, by route, by driver and by season. Tolls and wages are known in advance. The one input that would change a pricing decision is the one treated as a constant.

What a per-load cost record needs

Every fuel card transaction tied to a vehicle and a driver. The card issuer’s file already has the card, the station, the time and the litres. The company knows which card lives in which cab and who was driving. That join is the first and easiest step, and most companies have never made it.

The vehicle’s legs between fill-ups. A tank of fuel covers a set of legs: positioning, loaded run, empty return, the next positioning. Fuel between two fill-ups is spread across those legs by distance, loaded or empty, rather than dumped on whichever order happened that day.

Each leg attached to an order. The loaded leg belongs to the customer’s order. The empty return belongs to it too, as the earlier article on empty running argued, unless the next load picked it up. Once a leg has an order, the fuel on the leg has a customer.

The result written back to the lane. Cost per load is only useful if it survives the month. It needs to sit next to the rate charged and the time taken, on the lane, where the next quote will look for it.

A person who checks the odd ones. A fill-up with no matching vehicle, a leg with no order, a card used two hundred kilometres from where the truck was: these are the lines that need a human eye, and they are a handful a month, not hundreds.

Decisions that change once the record exists

The first change is in the quote. A dispatcher who can see that the last three runs on this lane burned more than the fleet average, because of the terrain or the regular empty return, prices the next one differently or declines it. The one who can see it burned less can go lower and still win.

The second is in which work the company keeps. A customer whose loads are consistently costed above the average is not a bad customer, but they are not the customer the monthly total suggested. The conversation about rates with them starts from a number instead of a feeling.

The third is in the fleet. A truck that burns more on the same lanes with the same drivers is telling you something about itself. A driver who burns more in the same truck on the same lanes is telling you something else. Neither shows up in a card invoice.

A test to run this week

Take last month’s fuel invoice and pick one truck. List its fill-ups in order, with the litres. Then, from your order records, list every leg that truck ran in the same period, loaded or empty, with the distance. Spread each tank over the legs it covered, by kilometre.

You will end up with a rough fuel cost for each order that truck carried. Compare it with what you charged, and with what the fleet average would have said. If the two cost figures sit close together on every order, your average is doing its job. If they diverge on even a few, those orders have been priced on a number that was never true for them, and the next quote on those lanes should not use it.

A fuel invoice is a supplier invoice and belongs on the loads it paid for; that is the single-object view of a load that EdrIQ is built around, with the cost sitting on the lane where the next quote can find it. It does not connect to the card issuer today. The matching described here starts with the data your company already receives, and the sibling article on planning with driving hours covers the other input a plan quietly ignores.