Distributor margin leakage

Margin leakage is an unhelpful term, because it suggests a leak: a single identifiable hole that somebody could find and plug. That is almost never what it looks like from inside a distribution business.

What it actually looks like is a set of individually reasonable decisions, each made by someone competent, each defensible on the day it was made, which together produce a number nobody intended.

The arithmetic nobody runs

Consider a distributor turning over eleven thousand order lines a month. A pricing gap of four hundred rupees — or five dollars, or four euros — on a single line is genuinely trivial. It is inside the noise of any reasonable commercial judgement, and if you flagged it to a sales manager they would be right to ignore you.

The same gap on every line, across a year, is roughly fifty-two thousand transactions. At that scale the number stops being trivial and starts being a budget line, and quite often it exceeds the operating profit of an entire branch.

The reason it never gets counted is that nobody experiences it at that scale. The rep experiences one order. The branch manager experiences one month. The finance director experiences a gross margin percentage that has moved by half a point, which is well within the range that gets attributed to mix.

Why it accumulates

The mechanism is almost always the same. A price is agreed in a contract. A rep adjusts it for a specific order, for a reason that made sense. A branch applies a local override because a competitor quoted lower that week. Then a supplier raises cost by seven per cent in March, and the price file is updated in September.

Every one of those steps is defensible in isolation. None of them is an error in any sense that an audit would recognise. The gap only exists in the relationship between them — between what should have been charged given everything the business knew, and what was actually invoiced.

That relationship is not stored anywhere. The ERP holds what was charged. The contract holds what was agreed. The cost record holds what it cost. Nothing in the system joins the three and asks whether they still make sense together.

Why the reports do not show it

Standard margin reporting aggregates. It tells you margin by product line, by branch, by customer segment, by period. Aggregation is the correct way to report performance and the worst possible way to find exceptions, because the arithmetic of an average is specifically designed to make outliers disappear.

A product line running at 22% margin might be forty customers at 24% and four customers at 11%. The report is accurate. It is also actively concealing the only thing on the page worth investigating.

This is not a failing of the ERP. An ERP is a system of record, not a system of exception. It is built to tell you accurately what happened, and it does. Asking it to tell you what should have happened instead is asking it to do a job it was never designed for.

What finding it actually requires

To locate this you have to work at transaction level and reconstruct, for every line, the price that should have applied given the contract terms, the price file, the cost record and the quantity-break tier in force on that date. Then compare that reconstruction against what was invoiced.

Most lines will match. Those get set aside, and they are the majority — which is the point, because the remaining few per cent are where the entire finding lives.

The remaining variances then have to be grouped by probable cause before they are ranked, or a single systemic pattern affecting eight hundred lines presents as eight hundred unrelated problems and nobody acts on any of them.

A caution worth stating

None of this means every variance is money someone can go and collect. A flagged line might reflect a negotiated exception that was never recorded centrally, promotional pricing, a strategic account with terms outside the standard file, a data quality problem in the export, or simply a decision somebody made deliberately and correctly.

An anomaly is a reason to look. It is not a conclusion. Any analysis that presents a variance total as recoverable revenue is overstating what the data can support, and the first time a finance team investigates three findings and discovers all three were legitimate, they stop trusting the entire exercise.

The useful output is narrower and more honest: here is where the numbers deserve attention, here is how often, here is what it would be worth if it turns out to be what it looks like, and here are the exact transactions so you can decide for yourself.