The Single-Invoice Question: How Much Revenue Should One Customer Be Allowed to Carry?

There is a particular kind of comfort in a large, dependable customer. Invoices are paid, the bank manager is content, and forecasting becomes a good deal simpler. That comfort has a price, and it is usually disclosed only at the moment the relationship changes hands, changes buyer, or changes mind.

Customer concentration is rarely a secret. In most SMEs the number is known to the finance lead, suspected by the managing director, and quietly ignored by everyone else. The purpose of an audit is not to induce panic but to replace a vague worry with a figure that can be managed against.

Why concentration bites cash before it bites sales

A concentrated customer book affects four things at once, and only one of them is revenue.

First, cash. Losing a customer worth a quarter of turnover does not simply reduce next year's sales; it leaves overheads stranded and stock already committed. Second, working capital exposure — if that customer also owes you money when trouble starts, you carry both a bad debt and a revenue hole simultaneously. Third, negotiating leverage: a buyer who knows they represent a fifth of your business will ask for terms you would not otherwise grant, and they will be right to. Fourth, enterprise value. Any acquirer or lender will apply a discount to earnings that depend on very few relationships, and that discount is applied before any conversation about multiples.

None of these problems is solved by growth for its own sake. A business can double turnover and remain just as exposed, because the new revenue often arrives from the same handful of names.

The 30-minute audit

The exercise needs the sales ledger for the last twelve completed months and nothing more sophisticated than a spreadsheet. Do it in this order.

1. Export every invoice line for the period, by customer.
2. Sum revenue and, where you have it, gross margin per customer. Where margin data is missing, use revenue now and flag the gap — margin is the number that matters.
3. Rank customers from largest to smallest and calculate cumulative percentage of the total.
4. Note the number of customers required to reach 80% of revenue, and again for gross profit.
5. For the top ten names, record contract status, notice period, payment terms, renewal date and the name of the relationship owner inside your business.

The two figures worth writing on a whiteboard are the count of customers providing 80% of gross profit, and the percentage held by the largest single account. In a healthy mid-market SME the first is often between five and eight; when it falls to two or three, the business is effectively a subsidiary of somebody else's procurement department.

Setting thresholds you can defend

Thresholds are a matter of judgment and sector. A manufacturer supplying one automotive tier-one will inevitably run higher concentration than a services firm with two hundred small clients, and there is no virtue in pretending otherwise.

As a starting position, many owner-managed firms adopt something close to the following: no single customer above 20% of gross profit; the top three below 50%; the top five below 65%. Where a single account legitimately sits above the limit — a public sector framework, a distribution agreement — the policy should record the exception, the mitigation and the date for review, rather than rewrite the rule.

The point of a written threshold is not bureaucracy. It gives a sales director a defensible reason to decline an order that would breach the ceiling, and it gives a board a metric to report each quarter. It also aligns the sales incentive: if commission is paid only on new revenue, concentration will rise and nobody will notice until it is expensive to fix.

Stress-testing the loss

Run the arithmetic before the market does. If your largest customer gave notice tomorrow, model three things: the cash position at 30, 60 and 90 days; the resulting earnings before interest, tax, depreciation and amortisation; and the headcount or cost base you would need to remove to hold margin.

Then ask the harder questions. How much stock or work in progress is committed to that account? What would a bad debt from them do on top of the revenue loss? Are there covenant tests with a lender that would fail? Does the timing make it worse — a loss announced in December is a very different event from one announced in April with a full pipeline behind it.

Where the exposure is material, credit insurance and tighter payment terms are sensible mitigations, but they protect cash, not confidence. There is no substitute for reducing the dependency itself.

Diversifying without diluting

The instinct when concentration is diagnosed is to chase anything that moves. This is how focused businesses become unfocused ones.

The better approach is to define the ideal customer profile that already works — sector, size, buying trigger, operational fit — and then widen the count of customers within it rather than the range of things you sell. Adjacent services sold to the same type of buyer are lower-risk than a new market sold to a new type of buyer.

Three practical levers tend to earn their keep. First, formalise referral: your best customers know others like them, and few businesses ever ask systematically. Second, productise the offer so that smaller accounts can be served profitably without loading the sales team. Third, set a pipeline coverage rule — three times the revenue at risk — so that diversification has a target rather than a sentiment.

Protect the core offer while doing this. A diluted proposition sold to a broader market usually produces lower margin and a slower sales cycle, which is the opposite of what a concentrating business needs.

Making it a standing item

Concentration is not a one-off project. Put the two headline figures on the monthly management pack, review customer-level profitability each quarter, and require any contract above the threshold to carry a documented exit assessment.

None of this is glamorous, and it will not appear in a pitch deck. What it will do is ensure that the day a major customer reconsider their supplier list, the decision is an operational problem with a plan attached rather than an existential one.