When to fire a client, and how to know before it costs you a year

Most advice on this question is about behavior. The client is rude, they miss calls, they pay late, they treat your team badly. Those are real, and they are also the easy cases. Nobody needs a framework to leave a client who shouts at a designer.

The expensive cases are the pleasant ones. A client everybody likes, who pays on time, whose account has quietly run at a loss for eleven months.

Run the number before you run the feeling

For each account, take last month's fee and subtract the fully-burdened cost of every hour that touched it. Include account management, QA, and the time you personally spent. Divide by the fee.

That is gross margin per account. Sort the list by it.

You now have a decision list that does not depend on how anyone feels about the client, which matters because the accounts that feel worst and the accounts that perform worst are rarely the same ones.

The floor

Forty percent gross margin per account, checked quarterly.

Below that there is no capacity left to fund hiring or investment, so the account is being subsidized by the rest of the business. A single account under the floor is a problem to fix. Three of them is why the agency cannot afford the hire it needs.

Three options, in this order

For every account below the floor:

Re-scope. Same fee, less delivery, honest about what fits. This is the easiest conversation and it works more often than people expect, because most clients would rather keep the price and lose the monthly report than lose the relationship.

Re-price. Ninety days' notice, reason tied to scope rather than to your costs. Expect a portion to leave, and treat that as the mechanism working.

Release. Only after the first two have been offered and declined. At that point the client has told you they want delivery you cannot provide profitably, which is a straightforward answer rather than a conflict.

Doing these in order matters. Agencies that skip to the third option lose revenue they could have kept, and agencies that never reach it keep accounts that quietly cost them a hire a year.

The other reason to leave a client

Concentration. Any client above 20 percent of revenue is a risk regardless of its margin, and above 30 percent the owner effectively has a job with extra steps and a boss who can fire them by email.

A concentrated account is also the one where you have least pricing power, because both sides know what happens if the relationship ends. That is worth correcting while it is still your decision.

What to say

Give the reason once, keep it about scope, and do not build a case.

"The scope has grown past what this fee supports. I can either reduce the scope to match, or move the fee to X. If neither works, I would rather hand this over cleanly than keep delivering something that is not working for us."

Offer a proper handover. Files, access, a call with whoever takes it on. The agencies that do this get referrals from clients they released, which sounds unlikely until it happens.

Expect to lose some, and count that as the point

In our experience the accounts that leave over a return to a 40 percent floor were the ones funding their own discount out of your profit. The revenue drop is visible immediately and the margin recovery takes a quarter to show, which is why this decision is easier with the list in front of you than without it.

Fraggell Productions went from £23k to £110k a month over 24 months, with the account mix corrected early rather than tolerated. The growth came from capacity that was no longer being spent on work that did not pay.

If you want the margin run across your own accounts first, the diagnostic does exactly that.

Read next

• Why your agency is busy and still not profitable
• Agency financial benchmarks: the numbers a healthy agency actually hits