Margin by Customer, Margin by Product: The Report That Changes Decisions

The P&L answers one question: did the business make money? It is structurally incapable of answering the better question: where? Which customers, which products, which jobs? And "where" is the report that changes behaviour — because almost every SME that builds it for the first time discovers the same uncomfortable shape: a profitable average concealing a chunk of the business that loses money with real commitment. The blended number was hiding cross-subsidies for years. Here's how to build the view, and what to do with what it shows.

Build it rough; rough is enough

Perfect cost allocation is an enterprise hobby. The SME version is an afternoon with your accountant or bookkeeper: take your top customers (or product lines, or job types — whichever cut you actually make decisions about), assign the direct costs you can trace (materials, subcontractors, direct labour hours at loaded rates), and produce gross margin per customer in dollars and percentage. Resist allocating every overhead — fine-grained allocation arguments are where these projects go to die. Direct costs alone reveal the pattern, and the pattern is the point.

Two refinements pay for themselves. Count the labour honestly — for service businesses, time is the cost, so even a few weeks of rough time-tracking against your biggest accounts beats assumptions (the account everyone "knows" is easy is frequently the one consuming a senior person's Thursdays). Add a service-intensity sniff test — the columns no system tracks: who generates the urgent reworks, the extended payment terms, the scope creep. A customer at decent gross margin who consumes disproportionate management attention and pays at 70 days is not, in any sense that matters, decent margin.

Reading the result: the four quadrants

Sort customers by margin percentage and size, and four populations appear:

Large and profitable — the core. The action isn't gratitude; it's protection and replication: deepen these relationships deliberately, and turn their profile into your sales targeting (more like these is the whole growth strategy in one sentence).

Small and profitable — the quiet good. Often servable with less effort than they get; sometimes growable into the first quadrant. At minimum, the marketing question: where do more of these come from?

Small and unprofitable — the long tail. Usually fixable with policy rather than drama: minimum order values, standard pricing instead of legacy rates, self-service options, or a respectful migration to a lighter service tier. Most of this quadrant became unprofitable through accumulated exceptions nobody priced.

Large and unprofitable — the report's reason for existing. The big account everyone celebrates that the numbers reveal as a subsidised passenger. The sequence here: verify the costing (be sure before the conversation), then re-price or re-scope — a direct, respectful negotiation armed with specifics ("at current volumes and service levels, this needs to move by X"). Some convert into your best accounts; they respected the relationship enough to expect honesty. Some leave — releasing capacity your profitable customers were queueing for, which is the part owners only believe after they've watched it happen.

Products and jobs: the same lens

The product/job-type cut runs identically and catches different leaks: the service line priced in 2022 and never revisited, the product whose input costs crept past its price, the job type that always overruns its quote (an estimating problem wearing a margin costume — feed actual-versus-quoted back into the quoting model and the leak closes at the source). One genuinely common finding: the offering the business is proudest of, the complicated flagship, earning half the margin of the boring bread-and-butter work. What you're best at and what pays best are different questions, and only this report asks the second one.

What the first run usually finds

Having watched many businesses do this exercise, the recurring discoveries are worth pre-announcing, because forewarned owners act faster: a fifth to a third of customers contributing little or negative margin; one celebrated major account revealed as break-even once real labour lands on it; legacy pricing surviving on accounts everyone assumed had been updated; and at least one boring, unloved service line out-earning the flagship by a wide margin per hour. None of these findings demand drama. Together they typically point to two or three points of recoverable gross margin — which, for context, usually exceeds the profit impact of any cost-cutting program the same business has ever attempted, and arrives without firing anyone or buying anything.

Make it a rhythm, not a revelation

The first run is an event; the value is the habit. Quarterly is plenty: the same cut, the trend per major account, one page in the management pack, and one decision per quarter arising from it — a price moved, a tier changed, a target profile sharpened. Within a year the portfolio visibly reshapes itself, not through any dramatic firing of customers, but through a hundred small decisions finally being made with the lights on. The blended P&L average will drift upward and look like luck. It will be the opposite of luck.

---

Profitability by customer is exactly the kind of blind spot our free Business Health Check is designed to catch. Five minutes, and you'll know what to look at next.