Client Concentration Risk: When One Account Is 40% of Your Revenue

There is a question I ask agency owners that makes them slightly uncomfortable, which is usually how I know it is worth asking: what share of your revenue is your single biggest client, and what would happen to the agency the month after they left? Most can answer the first part to the nearest five percent. Almost nobody can answer the second part without going quiet for a moment.

Client concentration risk is the gap between those two answers. It is easy to track the first number and tempting to never confront the second, because the biggest client is usually also the most flattering one: the logo you put on the deck, the relationship the founder is personally proud of, the account that paid for the office move. The trouble is that the same account is the one that can take a third of your turnover out of the building with one procurement decision, and the one that — when you look honestly — is often earning you a worse real margin than the smaller clients you treat as second-tier.

This is for the owner or finance lead carrying a top-heavy book. I want to do three things: separate the two risks concentration actually creates, show you how to measure it on real margin rather than invoice value, and lay out how to de-risk on purpose. There is a worked example in the middle where the biggest account by revenue turns out to be middle-tier by margin, because that is the case I see most.

Two risks wearing the same coat

Concentration is usually discussed as one risk. It is two, and they behave differently.

The first is margin risk, and it is live right now, every month, whether or not you ever sell. A client large enough to matter knows they are large enough to matter. That knowledge does the work quietly: they win the annual rate discount because losing them is unthinkable, they get the senior people on their work because you cannot afford a wobble, they get the scope creep absorbed because the relationship is too valuable to fight over a change order. Each of those is individually defensible. Together they mean the big account is frequently your worst real margin, not your best — even though it is comfortably your biggest invoice.

The second is valuation risk, and it is dormant until the day it isn't: the day you raise, sell, or take on a partner. An acquirer looks at a book where one client is 40% of revenue and does not see £X of turnover. They see a single point of failure they are being asked to pay for. Customer concentration is one of the first things diligence quantifies and one of the heaviest discounts they apply. A clean, diversified book of work at the same total revenue is worth materially more than a concentrated one, because the buyer is pricing the probability that the keystone client walks within eighteen months of the deal — which, post-acquisition, with new owners and a changed relationship, is not a small probability.

Both risks share a cause: a single relationship has been allowed to grow faster than the rest of the book without anyone deciding that it should. And both are usually under-stated, because both are measured on the wrong number.

Measure concentration on margin, not on the invoice

The standard concentration metric is simple: each client's revenue as a share of total revenue over a rolling twelve months. Rank them, look at the top one, look at the top three. If your top client is over ~25% you should be paying attention; over ~40% you are running a real exposure.

That is a fine starting point and a poor finishing one, because revenue concentration tells you what you would lose at the top line if the client left. It tells you nothing about what you would lose at the bottom line, and those are very different numbers when the big account is discounted and overstaffed with seniors.

So compute it twice.

• Revenue concentration — each client's invoiced revenue ÷ total revenue, trailing twelve months. The number everyone has.
• Margin concentration — each client's real margin contribution ÷ total real margin, trailing twelve months. The number almost nobody has, because it requires costing each client's delivery at real employee cost, allocating overhead, and counting unbilled scope before you divide.

That second number is the one that matters, and it usually moves in an uncomfortable direction. Because the big client tends to carry below-average margin, its share of profit is generally lower than its share of revenue — which sounds reassuring until you turn it around. It means a disproportionate amount of your delivery capacity, your senior bench and your overhead is being consumed to produce a client relationship that contributes less profit per pound of revenue than the rest of the book. You are concentrated on the thing that pays you least well per unit of effort.

If you have never separated real margin from invoice value, what project margin actually is is the place to start, because every number in this article depends on costing delivery properly rather than subtracting timesheets from invoices.

A worked example: the biggest account that wasn't

Take an agency, anonymised and scaled to round figures. Roughly 70 people, four significant clients plus a tail of smaller work. Trailing twelve months.

Here is the book as the bookkeeping reports it — and then as it actually is once delivery is costed at real employee cost, overhead is allocated pro-rata, and unbilled scope is counted.

Client A is unmistakably the biggest client. At 40% of revenue, A is exactly the kind of account that gets called "our anchor" and earns the founder's personal attention. On the invoice, A is the agency.

On real margin, A is not the agency. A runs an 11% real margin — well below the ~18% blended — because over six years A has negotiated the deepest discount in the book, takes first call on the two best architects, and has a standing habit of small unpapered extras that nobody wants to invoice for fear of the relationship. So although A is 40% of revenue, A is only 24% of profit. Client B, at 25% of revenue and a healthy 22% margin, contributes more actual profit than A does. The account everyone treats as second-tier is the one quietly carrying the agency.

Two things follow that an owner should sit with.

First, the margin risk is already being paid, every month, in the form of the discount and the senior allocation A commands. The agency is funding the privilege of being concentrated.

Second, the valuation risk is worse than the revenue table implies in one direction and slightly better in another. Worse, because a buyer sees 40% revenue concentration and a keystone client on the thinnest margin in the book — a client with both the leverage and, given the discount, possibly the motive to renegotiate or leave. Slightly better, because the agency's profit is less concentrated than its revenue, which is the more honest picture of resilience — if you can show the buyer that picture, with the costing behind it, rather than leaving them to assume the worst from the revenue line alone.

What a healthy mix actually looks like

There is no single correct distribution, but there are shapes that survive a bad year and shapes that don't. The ones I trust have three properties.

No single client over roughly a quarter of revenue, and no client over a quarter of margin contribution. Both ceilings, not just the revenue one. A client can be a modest share of revenue and a large share of profit; that is its own concentration, and losing it hurts the bottom line more than the top line suggests.

The top three clients under ~50% combined. Concentration is not only about the single biggest account. Three clients who are each "only 20%" are a 60% cliff if they share an industry, a buying season, or a single introducer who could take all three at once. Look for hidden correlation, not just individual size.

A live tail that can grow. A healthy book has smaller engagements that are genuinely profitable and could absorb more work if a big account shrank. A tail of loss-making small clients is not diversification; it is just more risk at a worse margin. The point of the tail is optionality, and optionality only counts if the tail makes money.

Measured this way, the example agency is more exposed than its owner believes on revenue (40% in one client) and slightly less exposed than the revenue line implies on profit (24%). Neither reading is comfortable. Both are true. Concentration is a range you manage, not a number you eliminate — every agency has a biggest client, and the goal is to keep that client from quietly becoming the agency.

De-risking on purpose

You bring a top-heavy book down in one of two ways: the client does it for you, on their schedule, with no warning; or you do it deliberately, over twelve to twenty-four months, while you still have leverage. The deliberate version has three moves.

1. Recover pricing power on the anchor. The reason the big client is your thinnest margin is usually that the discount was set years ago and has never been revisited, while scope quietly grew. You are unlikely to claw the discount back in one negotiation, but you can stop the unpapered extras, re-scope at renewal with the real cost in front of you, and decline the reflex of always staffing your best two people there. Recovering even a few points of margin on a 40%-of-revenue account moves the whole agency. If the discount is the heart of it, the mechanics in realised rate and discounting are the lever.

2. Grow the rest faster than the anchor. The cleanest way to reduce concentration is rarely to shrink the big client; it is to grow everything else. That means a pipeline that is actually fed — not the founder fitting sales around delivery for the anchor — and a deliberate bias towards landing the kind of mid-sized, healthy-margin work that Client B represents in the example. Diversification is a sales-capacity decision before it is a finance decision.

3. Productise the repeatable part. The work you do for the big client is often the most repeatable work you have, precisely because you have done it for years. Some of it can become a defined, fixed-scope offer that several smaller clients buy — which both diversifies the book and de-risks delivery, because a productised offer does not depend on the two architects the anchor monopolises. This is also where margin on the smaller end of the book tends to improve, since repeatable work estimates better.

None of these is fast, and that is the point. Concentration that took six years to build does not unwind in a quarter. The agencies that handle it well are the ones that started the year they first measured it honestly — not the year the client gave notice. The closest I have written to the wrong-end-of-this is the three clients we cancelled, where the accounts we had carried longest were the ones losing the most; concentration and silent unprofitability are usually the same story told from two angles.

If you want to see your own book on both lenses — revenue concentration and real-margin concentration, side by side, on your actual Jira data rather than a reconstructed spreadsheet — that is one of the first views Saldo puts in front of you, and the reference customer page shows what it looks like at scale across a 120-person book.

The short version: the biggest invoice is not the biggest risk, and it is rarely the biggest profit. Measure concentration on real margin, watch both ceilings, and bring a top-heavy book down while the decision is still yours to make.