Ask an operator what their margin is and most can answer immediately. Ask what it is by customer, and the answer slows down. Ask which customers earn nothing at all after everything is loaded in, and the conversation usually stops.
That gap is not a bookkeeping failure. Company-level margin is the number the accounting system is built to produce, so it is the number that gets looked at. The trouble is that it is an average, and an average of a distribution this uneven is close to useless for deciding anything.
The shape underneath the average
In most businesses that make or ship something, gross profit is not spread evenly across the customer book. A small share of customers produces most of it. A meaningful share of the book produces close to none, and some of it produces less than none once you load in the expedites, the short runs, the credit terms and the returns.
Neither of those facts is visible at the company level. Both of them change what you should do next.
If most of your gross profit comes from a small set of accounts, then your marketing budget has a target it did not have this morning, and it is a much narrower target than the one you were aiming at. If a third of your book is earning nothing, then the growth you were about to buy will arrive on top of work that is already not paying for itself, and doubling it doubles the problem rather than the profit.
Why it stays hidden
Three reasons, and none of them is negligence.
The general ledger is organized by account rather than by customer or by item, so the cut you need is a transformation somebody has to sit down and do. It is nobody’s weekly job.
Cost allocation is genuinely hard, and the fear of getting it slightly wrong stops people running it at all. That is the wrong trade. A rough cut that shows you the shape beats a precise cut you never build.
And the answer is often uncomfortable. The account that everyone is proud of, that the sales team named a conference room after, is sometimes the one at the bottom. Nobody goes looking for that on a Tuesday.
The first pass, which you do not need us for
You need one year of transaction-level data. Sales lines with customer, item, quantity, price and whatever cost you can attach.
- Cut gross profit by customer. Sort descending, then take a cumulative running total. Find the point where you cross eighty percent. Count how many customers are above it.
- Cut gross profit by item or product line. Same operation, same sort.
- Find the zero band. Everything at or below roughly zero gross profit after loaded cost. Add up the revenue in it, not the profit, because revenue is what it is consuming in capacity.
That is it. Three cuts, and the third one is the one that changes meetings.
You are not looking for a decimal place. You are looking for the shape, and the shape shows up even in a rough pass. If the answer surprises you, that surprise is the finding.
What we would do differently
Two things, and this is where a survey earns its fee rather than a spreadsheet.
We load cost properly, which means the setups, the changeovers, the expedite freight, the returns and the cost of carrying the receivable, not just materials and direct labor. The zero band moves a long way once those are in.
And we read it against the rest of the business rather than on its own. A product line losing money is a finance finding. A product line losing money because it runs short orders through a line that was not built for them is an operations finding wearing a finance costume, and the fix is a different fix in a different department. Which one it is decides where the money should go, and you cannot tell them apart from the margin table alone.
That is the whole reason we read four areas rather than one before naming anything. The number tells you where to look. It does not tell you what you are looking at.