When your dip doesn't match your books

Every figure below is hypothetical. The pattern is not.

You dip the tank on Saturday morning. Your book figure says there should be 12,400 litres in there. The dip says 12,050.

Three hundred and fifty litres are missing. At a landing cost of ₦1,236 a litre — see what a litre actually costs you for where that number comes from — that gap is worth ₦432,600.

What happens next in most businesses is one of two things. Either somebody adjusts the book figure to match the dip and moves on, or somebody starts accusing the pump attendant. Both are wrong, and the second one is worse, because it burns a good staff member over a gap you have not yet explained.

First, the arithmetic that produces the gap

Stock reconciliation is one line:

opening stock + deliveries in - sales out = book stock

For the month above:

The variance on its own means nothing. What matters is the variance as a share of what moved through the tank:

350 / 30,000 = 1.17%

That is the number to watch, month after month. An absolute figure of 350 litres is alarming on a small tank and unremarkable on a large one. A percentage is comparable across months, and it is the only version that tells you whether things are getting better or worse.

Four things cause the gap, and they are not equally serious

Measurement error. A dipstick read at an angle, a tank that is not perfectly level, a calibration chart that was made for a different tank. This is the most common cause and the most ignored. Before anything else, ask when the tank was last calibrated and whether two people reading the same dip get the same answer.

Temperature and density. Diesel expands when it is warm. Product received cool at 6am and dipped hot at 2pm genuinely occupies a different volume, and if you are converting between kilogrammes and litres with a remembered density rather than a stated one, you are introducing error at every single delivery. On 30,000 litres, a small density assumption error is easily a hundred litres.

Genuine operational loss. Evaporation, line fill, the residue that never comes out of a hose, small spills. Real, mostly unavoidable, and it should be stable. Stable loss is a cost of doing business. Loss that moves around is a signal.

Theft. It exists, and pretending otherwise is not kindness. But it is the last explanation to reach for, not the first, because you cannot distinguish it from the other three until you have eliminated them. A business that has never reconciled properly has no basis for the accusation, and no defence for the person accused.

The entire value of doing this monthly is that it separates these four. One unexplained gap tells you nothing. Twelve months of gaps, each one measured the same way, tells you almost everything.

Set a tolerance before you need one

Decide in advance what variance is acceptable, write it down, and let the number rather than your mood decide when something is wrong.

A tolerance of around half a percent of throughput is a reasonable starting point for a well-run tank, but the right figure for your operation is whatever your own reconciliations settle at once you have fixed the measurement problems. Establish your baseline first, then hold yourself to it.

The 1.17% above is not a crisis. It is a number above tolerance and therefore requiring an explanation — which is a completely different thing, and a much more useful thing to say out loud to your team.

Never silently edit the stock figure

When you find a variance, the correction goes in as its own entry: the quantity, the date, the reason, and the person who authorised it. Not an edit to the original number.

This matters for three reasons, and only one of them is about honesty.

• You lose the history. If you overwrite the book figure, next month's reconciliation starts from a number that was quietly invented, and every subsequent month inherits the error.
• You lose the pattern. Adjustments with reasons attached are the data. Six adjustments all reading "short on delivery" from the same supplier is a conversation with that supplier. You cannot see that if the adjustments were edits.
• You lose the ability to defend anybody. Including yourself, if a customer or a partner ever asks.

The routine

Weekly is better than monthly, and monthly is enormously better than never.

1. Dip at the same time of day, by the same method, ideally by two people. Consistency matters more than precision here — a consistent method with a known bias is more useful than an inconsistent one that is occasionally exact.
2. Close off sales and deliveries before you dip. A truck discharging while you measure guarantees a gap you will spend an hour chasing.
3. Compute the variance in litres and as a percentage of throughput. Record both, every time, in the same place.
4. Explain anything outside tolerance before you close the week. Not later. The information you need to explain a gap — who was on, which truck, what the weather was — decays within days.
5. Post an adjustment with a reason. Then start the next period from the reconciled figure.

What this is really about

Nobody reconciles stock because they enjoy it. They do it because it is the only mechanism that converts "I think somebody is stealing from me" into "our variance has run at 0.4% for eight months and this month it was 1.9%, and here is what changed."

The first sentence poisons a workplace and resolves nothing. The second one is a question with an answer.

And it is worth being blunt about the cost of not knowing: at ₦1,236 a litre, a variance of one percent on 30,000 litres a month is ₦4.4 million a year leaving the business through a hole nobody has measured. That is not a rounding error. For most suppliers, that is the year's profit.