When to Turn Down Work (and How to End a Client Relationship Properly)

Every agency has had the client who took 40% of the team's attention for 8% of the revenue. The one whose feedback arrived at 11pm, whose scope moved weekly, whose invoices needed chasing three times.

The revenue from that client was real. So was the cost, and the cost was larger — it just didn't appear on any line item.

What a bad-fit client actually costs

Opportunity cost. Every hour on a difficult client is an hour not spent on a good one, or on the marketing that would find more good ones.

Team morale. Nobody quits over one hard client. People do quit over the accumulation of them, and replacing a good designer costs more than the account was worth.

Your standards. Chaotic clients produce rushed work. Rushed work goes in your portfolio, or doesn't, and either way you've spent months producing nothing you can show.

Cash flow. Late payers are their own category of expensive. You're financing their business at your expense.

The maths is usually brutal once you actually run it. A client at 8% of revenue consuming 40% of attention is losing you money even at a healthy margin.

Spotting it before you sign

Most bad-fit relationships were visible in the sales process. The signals get ignored because revenue is revenue and the pipeline looked thin that month.

They lead with price. Not "what does this cost" — that's normal. But an opening move of "what's your best rate" tells you what they'll optimise for through the whole engagement.

They badmouth their last agency at length. Sometimes the last agency was genuinely bad. But listen to how they describe it. If everyone who's ever worked with them was incompetent, consider the common factor.

They won't say who decides. Either they don't know, or they don't want you talking to the person who does.

The timeline is urgent and the reason is vague. Real deadlines have causes. Manufactured urgency is a negotiating posture that continues after you sign.

They want to renegotiate scope before starting. This is the single most reliable predictor. The proposal is the most cooperative they will ever be.

They're evasive about budget. Not "we don't have a number yet" — that's fine. But refusing to give a range while asking you to spec in detail means they're using your time to price-shop.

Any one of these is a yellow flag. Two or more and the price should go up substantially or the answer should be no.

How to say no well

You will work in a market with a reputation, so decline gracefully. A well-handled no generates referrals surprisingly often.

Be prompt. Don't ghost. A quick no is a favour.

Be honest but not brutal. "We're not the right fit for this one" is enough. You don't owe a diagnosis, and offering one invites an argument.

Refer when you genuinely can. A client who's wrong for you may be right for someone else. That referral is a deposit in a relationship you'll want later.

Price yourself out honestly, if you'd take it at the right number. Quote what the work is actually worth given the difficulty. If they say yes, the money compensates. If they say no, you got the outcome you wanted. What you shouldn't do is quote an insulting number as a covert rejection — people talk.

Ending an existing relationship

Harder, and occasionally necessary.

Try to fix it first. Have the direct conversation before the termination one. Name the specific behaviour, propose the specific change. "Feedback arriving after each round of revisions is pushing the timeline. Can we agree to consolidate it into one response within three days?"

Some clients genuinely don't know they're difficult. A clear conversation fixes a real share of these, and you keep the revenue.

If it doesn't improve, exit cleanly:

• Check the contract first. Notice period, deliverables owed, payment terms on termination.
• Give real notice. 30 days is common. Abandoning someone mid-project is the thing that produces the public complaint.
• Finish what's committed where you reasonably can. Half-delivered work is what people write reviews about.
• Hand off properly. Files, access, documentation, a written summary of where things stand. Make it easy for whoever comes next.
• Keep the reason simple. "We're not the right long-term fit for your needs." Don't itemise their failings — it changes nothing and escalates everything.
• Get it in writing, including what's been paid and what's owed.

Settle the money before you go quiet. Chasing a final invoice from a client you've just terminated is difficult. Where you can, structure so you're never significantly ahead on delivery.

The client you keep for the wrong reason

Watch for the account you'd never take today but keep because you always have. Often an early client, often at old rates, often the most demanding relative to what they pay.

The loyalty is real and worth something. But if they're at half your current rate and twice your average effort, that's not loyalty — it's an unexamined subsidy.

The honest move is a conversation about rates, not a slow resentful drift. Give notice, explain the change, and let them decide. Some will pay. The ones who won't were going to leave eventually anyway.

What to do with the capacity

Firing a client only works if you replace the revenue. Otherwise it's a panic move in three months.

Practical sequencing: fire from a position of strength, not desperation. Fill the pipeline first, then let the bad-fit account go. Ideally you're turning them down at renewal rather than terminating mid-engagement.

And use the freed capacity deliberately. The temptation is to fill it with whatever arrives next, which is how you end up with the same problem and a different name on it.

The uncomfortable summary

Most agencies know exactly which client they should let go. They keep them because the revenue is certain and the replacement isn't.

That's a reasonable fear and usually a bad trade. Run the actual numbers — hours spent versus revenue, against your average — before you decide it's not.

---

We work with agencies as a fixed-scope build partner, which tends to make the difficult conversations easier because the boundaries are written down. See how it works.