The owner I am thinking about as I write this had wanted to fire her biggest client for two years. The client was 35 per cent of the firm’s revenue. They were demanding response times the team could not sustain, squeezing the margin every renewal, and exhausting the two senior people the work depended on. She had wanted to have the conversation for two years and kept finding reasons to defer it. The renewal was coming up. The team had just landed a new project for them. The cash flow forecast still leaned on them. Then it was the autumn, then it was the new year, then it was two more years.
She is not unusual. Almost every owner-managed firm I work with eventually has a 30 per cent client, and almost every owner who has fired one says the same sentence after the fact. “I wish I had done it sooner.” The cost of deferring the conversation is rarely the revenue, it is the team energy bleeding out, the margin compression, and the work the firm is not doing because the big client is taking up the room.
This post is about why the call is so hard, the three tests that simplify it, and what AI changes about the calculation. None of it makes the conversation easier. It can make the conversation clearer.
What does firing the 30 per cent client actually mean?
Firing the 30 per cent client means giving structured notice to the largest revenue concentration in your book, with a planned three-month handover, when the client is extracting from the firm rather than improving it. At 30 per cent the relationship is load-bearing, the cost of getting the decision wrong is real, and the call cannot be made on instinct alone.
The 30 per cent number is approximate. It marks the rough point at which the conversation has to be deliberate rather than incremental. A client at 12 per cent of revenue who is mildly annoying gets pruned at renewal without ceremony. A client at 30 per cent who is exhausting the firm needs an explicit decision, a transition plan, and a cash flow forecast that holds during the handover.
Why is this the call owners defer most?
Owners defer this call because four things make it uniquely uncomfortable, the revenue feels load-bearing, the team has the relationship and may resist losing it, the alternative pipeline is unclear, and the conversation itself is hard. Each one alone is manageable. Stacked together they become a reason to find one more renewal cycle to wait. The deferral is rational, not lazy, the decision has obvious cost today and uncertain benefit later.
The revenue argument is the loudest. A 35 per cent client paying steady invoices is a cash flow line on the forecast that the firm has built its operating cost base around. Losing that line means cutting costs or replacing revenue inside a short window, and neither feels like a positive action. The team argument runs underneath it. The two or three senior people who carry the relationship have built years of context with this client, and they will lose part of their working identity when the engagement ends. The pipeline argument is the quietest but the most corrosive. The firm has been so busy carrying the big client that the partner has not built out the next significant relationship, which makes the replacement question feel even more daunting than it really is.
Where do the three tests cut through?
The three tests separate the financial picture from the team picture from the strategic one. Financial extraction asks whether the client is materially less profitable than the rest of the book on full cost to serve. Team extraction asks whether they are taking disproportionate team energy for their revenue share. Optionality extraction asks whether they are preventing better work elsewhere. Two failures is a strong signal. Three is the conversation.
Financial extraction is the easiest test to run honestly, and the one firms commonly run dishonestly. The headline revenue number says the client is profitable. The cost-to-serve number, including senior partner time, the team’s compressed delivery schedule, and the discounted day rate the client negotiated three years ago, tells a different story. Run the calculation with full loaded cost. If the client is materially below the firm’s average gross margin per client, the financial test has failed.
Team extraction is the test owners commonly feel in their gut before they put numbers to it. The work feels heavy in a way the rest of the book does not. The senior team is tired in a way that does not match their billable hours. There are evening emails, weekend escalations, response times the rest of the book does not require. The proxy metric is utilisation against satisfaction. A client that runs the team at high utilisation and low satisfaction has failed the team test, even if the revenue still looks good on paper.
Optionality extraction is the strategic test. Is the firm turning down better work because the big client is consuming senior capacity. Is the partner saying no to interesting engagements because no senior person is free. Is the team building a narrower CV than it should be. A firm that fails the optionality test is paying a cost that does not appear on the P&L.
When to make the call and when to hold
Make the call when two or three tests have failed for two consecutive quarters, when you have deferred the conversation for over twelve months, and when the firm has the financial buffer to absorb a three-month transition. Hold the call only when one test has failed and the underlying issue is fixable inside two quarters, or when the firm is in a working capital squeeze that genuinely cannot survive the loss.
The structured exit is the part owners describe as the bit they wish they had planned more carefully. Give the client written notice three months out. Be honest about the reason at a high level, the work has changed, the firm is repositioning, the relationship has not been a good fit for the firm’s direction for some time. Offer to warm-introduce them to a competitor or a friendly firm that fits them better. Agree a handover plan for current matters. Stage the revenue reduction so the team has time to redirect capacity and the firm has time to absorb the cash flow shift.
The AI piece changes the cash flow calculation in one specific way. The marginal cost of taking on a new significant client at acceptable quality has come down, because the productivity gain on drafting, research, and routine analysis is real. The bar has not gone to zero, but it has come down enough that the historical defence we cannot replace this revenue is weaker than it was two years ago. The team time released by firing the wrong client is itself usable capacity for the new business work.
What changes after the client is gone
What changes is consistent across the owners I have watched do this. The team energy comes back inside a quarter. The margin recovers within two quarters because cost-to-serve was higher than the revenue suggested. The book fills back to the same revenue or higher within nine to fifteen months, often with two or three new clients at better margin and lower team strain. The change is boring, the work just gets better.
The risk to flag is not the revenue gap, it is the temptation to fill it with the wrong replacement. A firm that has just fired a 30 per cent client is tempted to take on the next significant prospect with the same dynamics, because the cash flow forecast is asking for it. The discipline is to use the three tests on the prospect before signing, not after the engagement is six months old.
None of this is a recommendation to fire concentrated clients on principle. Concentration is a separate question with its own thresholds. The 30 per cent client question is the harder one, whether the client is improving the firm or extracting from it. The three tests make the decision simpler. AI lowers the cost of replacement, not the difficulty. The conversation gets clearer, not easier.
If you have been deferring this call, Book a conversation.
