What a litre of diesel actually costs you

Every figure below is hypothetical. The pattern is not.

You run a diesel supply business. Your tank took three deliveries this month, at three different landing costs, because the market moved between them. Yesterday you sold 8,000 litres to a factory in Ikeja.

Here is the question almost nobody in this business can answer in the moment: what did those 8,000 litres cost you?

Not what you paid for your last truck. What those litres cost.

Why "what I paid last" is the wrong number

The tank does not keep your batches separate. Diesel from Monday's delivery and diesel from the one three weeks ago are the same diesel once they are in there. So the cost of a litre stops being a fact you can look up and becomes a figure you have to decide how to calculate.

Most owners default to the price of the most recent truck, because it is the number they remember. In a flat market that is close enough to be harmless. In a market that moves — which is the only market anyone here has ever traded in — it is wrong in a specific and expensive direction: when prices are rising, using your latest cost makes your margin look worse than it is; when prices are falling, it makes your margin look better than it is.

The second one is what does the damage. It flatters you exactly when you should be tightening up.

The arithmetic, worked

Say the tank started empty and took three deliveries:

Delivery	Litres	Landing cost / litre	Value
3rd of the month	10,000	₦1,180	₦11,800,000
14th of the month	12,000	₦1,240	₦14,880,000
26th of the month	8,000	₦1,300	₦10,400,000
In tank	30,000	—	₦37,080,000

"Landing cost" means what the litre cost you by the time it was in your tank — product, freight, any losses on the way, everything. Not the depot price on the invoice.

Weighted average cost per litre is the total value divided by the total litres:

₦37,080,000 ÷ 30,000 litres = ₦1,236 per litre

So the 8,000 litres you sold yesterday cost you ₦9,888,000, not the ₦10,400,000 you would get from the latest truck price. That is a ₦512,000 difference on one delivery — and it is a difference in the direction of you thinking you made less than you did, then pricing the next job to fix a problem you did not have.

Run it the other way, with the market falling, and the same arithmetic hands you a margin that is not there. That is the version that ends in a bad quarter nobody saw coming.

Weighted average or batch-by-batch?

There are two defensible ways to do this, and the choice matters less than making one and sticking to it.

Weighted average is the one above. Every litre in the tank carries the same cost, recalculated each time a delivery lands. It is simple, it is what most supply businesses should use, and it survives the fact that your tank physically mixes everything anyway.

Batch (FIFO) tracks each delivery separately and assumes the oldest litres leave first. It gives you a truer cost per sale and it is worth the extra bookkeeping if you hold product a long time, if your batches differ a lot in cost, or if a customer contract is priced off a specific consignment.

What you must not do is switch between them depending on which one makes the month look better. Pick one, write it down, and let it tell you the truth.

The part that catches people: litres, kilogrammes and temperature

Diesel is sold by the litre and moved by weight. A truck is weighed in and out; your tank is dipped in litres. Those two numbers will not reconcile unless you are explicit about density.

Convert with a stated density, not a remembered one. AGO is commonly taken at roughly 0.84 kg per litre, but it varies by product and by supplier. Whatever figure you use, write it into your records so the same conversion is applied every time and any argument later is about the density, not about who is lying.
Diesel expands and contracts with temperature. A litre loaded cool in the morning and dipped hot in the afternoon is not the same volume. Over a single truck it is small; over a year of deliveries it is the difference between a shortage you can explain and one you cannot.
A dip is a measurement, not an opinion. If your dip and your book figure disagree, one of them is wrong and you need to know which today, not at year end.

The smallest ledger that actually works

You do not need software to start. You need four things recorded consistently, and most businesses in this trade are missing at least two:

Every inbound delivery, with its own landing cost per litre. Product cost plus freight plus known losses. Dated. This is the one people skip, and without it nothing downstream can be right.
Every outbound sale in litres, tied to a customer and an invoice. Not a total at the end of the week — each movement, so a gap has a date and a name attached to it.
A running stock figure the system calculates, and a physical dip you take independently. Two numbers that should agree. The value is entirely in the moments they do not.
A reconciliation you actually do, on a schedule. Weekly is fine. Never is what most people do.

If you have those four, your cost per litre falls out of the arithmetic instead of being a guess, and so does the answer to the only question that really matters: which customers are you making money on?

What this is really about

Fuel businesses rarely fail because of the price they buy at. They fail because the gap between what they think a litre costs and what it actually costs sits unexamined for long enough to price a year of contracts wrongly.

Every one of these numbers is knowable. Most of them are already written down somewhere in your operation — on a waybill, in a notebook, in somebody's phone. The work is not collecting them. The work is putting them in one place where the arithmetic happens by itself, every day, without anybody remembering to do it.

That is the whole job.