How do you know when to fire a client?

Read the dread before you act on it. When a client's name in your inbox produces a physical reaction, something specific has gone wrong, and it's worth naming before you decide anything. Often it's scope creep against a vague agreement. Sometimes it's a deal that was structured badly from the start. Ending the relationship is one option. It isn't the first one.

What is the dread actually telling you?

I'll go first, because I've been on the wrong side of this.

Years ago I had one of my first fractional CMO clients. Fourteen months, and I spent roughly six of them thinking about ending it. Every time their name appeared in my inbox I'd feel my shoulders go up before I'd even opened the thing. The demands weren't realistic. There was no approval process, so everything moved at the speed of whoever happened to be free, which was nobody. It ground on.

Now the part I'd rather not include. A good share of that was my own doing.

I had dropped the price on purpose, because the work wasn't there yet. Their project wasn't ready. So we spent months doing preemptive marketing, getting everything into position for a thing that kept not happening. I set up an arrangement where I was cheap, the scope was vague, and the outcome depended on something entirely outside my control. Then I was surprised when it felt bad.

That was me growing up as an entrepreneur. I don't think there was a shortcut available.

Eventually I ended it. What I said was that I didn't think it was going the way it should, which is honest and about as much as anyone needs. And then the thing that still slightly annoys me: I replaced the revenue in about a week.

Six months of deliberation. One week of consequence.

So when I say the dread is worth reading, I mean read it, not obey it. Dread is a signal that something in the arrangement needs to be addressed. Sometimes the answer is to end it. Often the answer is a conversation you've been avoiding for five months.

What exactly is scope creep?

It starts with an undefined contract. That's the root of nearly every case of it I've seen.

If the agreement doesn't say what's included, what isn't, and what an addition costs, then every request is a negotiation conducted entirely inside your own head, usually at 10pm, usually resolved by saying yes. Do that forty times and the engagement you signed no longer exists.

Two things worth separating, because they get confused constantly.

Going the extra mile is good. I believe in it. Doing something small and unbilled because it obviously serves the client is most of what builds a relationship worth having, and a business that meters every fifteen minutes is unpleasant to hire.

Adding something significant to a retainer is a different animal. On a one-off project, the extra mile has an end date built in. On a retainer, whatever you absorbed this month becomes the baseline for next month, permanently, at no additional fee. That's where it compounds, and where founders quietly resent a client for a decision the founder made.

This also depends enormously on what you sell. If you're a painter finishing a kitchen, dining room and living room and the client asks whether you could quickly do the powder room, that's not much of an ask, and you'd probably just do it. If they ask about the sunroom, that's a real conversation, and everyone in the room knows it. Materials and hours make the line visible. Construction is relatively easy to say no in. I suspect law firms are too.

Soft work is where this gets hard. In marketing, strategy, design, or advisory work, nobody can see the sunroom. An extra deliverable looks like an email. There's no lumber to point at, so the boundary has to be written down in advance or it doesn't exist at all.

And when you do hold the line: don't be a jerk about it. Most people know when they're asking for something extra. A calm "happy to do that, it's outside what we agreed so let me price it" is normal, professional, and lands far better than founders expect. The relationships that break are usually the ones where somebody silently absorbed twenty asks and then blew up on the twenty-first.

Does the cheap client actually cost you more?

Sometimes, and it's worth doing the arithmetic, though I wouldn't make it the test.

Say a client pays you $1,800 a year. Nice person, been with you since the early days. Now count what they consume: onboarding that took the same hours as your best client's, a compliance or insurance obligation that doesn't shrink, three or four emails a month, a chase for documents every quarter, a renewal conversation, a software seat, and forty minutes of attention every time their name appears and you decide to deal with it later.

Price that honestly and the account is losing money. Most owners already suspect this. What stops them is that the loss never appears as a line item. It appears as a Tuesday that disappeared.

Useful information. Just not sufficient on its own, because plenty of small accounts are fine and some large ones are the problem.

Is any of this a positioning decision?

Yes, and it's the part that gets noticed least.

Every client you keep is a quiet claim about who your business is for. If a prospect could see your whole roster, what would it tell them? A firm whose list is forty small accounts and three good ones is a firm that says yes to everyone, and everyone is not a market. Your client list is a strategy document that most founders never wrote on purpose.

Positioning is mostly a series of refusals: who you don't serve, what work you turn down, which projects you won't take even when the calendar is thin. A firm that refuses nothing has no position, only a schedule. Specific sacrifice is one of the strongest trust signals a service business can send, because it's expensive to fake. When you tell a prospect "we don't do that, and I know two people who do it well," they believe everything else you say a little more.

Look again at my fractional CMO story with that lens. The client wasn't a villain. I took on work whose outcome depended on their project being ready, priced it below what it was worth, and never wrote down what would happen if the timeline slipped. That's a structure problem, and I built it. Which is also the good news, because structure is the one part you can fix before it happens again.

What should you try before you end it?

Ending a relationship is the last move, not the first. The research on unprofitable clients is unusually practical here: Harvard Business Review's work on customer divestment lays out a sequence for restoring an account before ending it, and notes that the decision touches morale, capacity and strategy as much as the P&L. Reflexive firing costs more than founders think, including with the clients who are watching.

Three cheaper moves, in order.

Say the thing. The conversation you've been avoiding is usually the whole fix, and it's usually about scope. "The work has grown past what we agreed. Let me show you what it looks like now and what it costs." Most clients aren't trying to get away with anything. They genuinely don't know where the line is, because nobody drew one.

Reprice or rewrite. New rate, new scope, in writing, with a date. Some clients say yes immediately, because they had no idea they were on a number you set three years ago. Some leave. Both outcomes are better than the current arrangement.

Change the shape. Quarterly instead of monthly. A defined deliverable instead of open availability. Much of what makes an account exhausting is that it's being served like a bigger one out of habit and guilt.

If all three fail, end it. With notice, in writing, without a manufactured excuse, and with a referral to someone genuinely better suited if you have one. "I don't think this is going the way it should" is honest and complete. People take it better than you expect.

Reading the relationship this week

Start with the one you already thought of when you began reading.

1. Name what actually happens. Write down, specifically, what goes wrong: the requests outside scope, the approvals that never come, the meetings that produce nothing. Feelings are the signal. Specifics are what you can address.

2. Go read the agreement. Does it say what's included, what isn't, and what an addition costs? If it doesn't, you've found the root, and it's yours to fix rather than theirs.

3. Ask what you built. Did you discount it, take it on before it was ready, or leave the outcome depending on something you don't control? Own your half honestly. It's the only half you can change.

4. Have the conversation before the decision. One conversation about scope and price, this month. If it goes nowhere, you'll have a much clearer conscience about what comes next, and you'll have learned how to write the next agreement.

A painter, a bookkeeping practice and a design studio will draw these lines in completely different places, and they should. The discipline is writing the line down before you need it.

Sources

• Mittal, V., Sarkees, M. & Murshed, F., "The Right Way to Manage Unprofitable Customers", Harvard Business Review, April 2008

Frequently asked

Is dread on its own a good enough reason to end a relationship?

It's a good enough reason to investigate. Something real is producing it, and naming that specifically usually points at scope, pace, respect, or an arrangement that was structured badly. Once you know which, you often have a fixable problem. If you've named it, raised it, and nothing changed, then it's a reason.

How do I stop scope creep without seeming difficult?

Write the boundary down before you need it, then hold it warmly. A clear agreement listing what's included and what an addition costs turns every awkward negotiation into a lookup. Most clients respond well to "happy to, that's outside our scope, let me price it," because it's a normal thing professionals say.

What if the client is genuinely fine and I just don't enjoy the work?

That's worth separating out, because the fix is different. Not enjoying a category of work is a positioning question about what you sell and to whom, and firing one client won't resolve it. You'll just sign another one exactly like them within a quarter.

Won't I lose the revenue?

Possibly, and less permanently than you fear. In my own case I'd stalled for six months over money and replaced it in about a week. That won't be everyone's experience, so don't plan around it. But the fear is usually louder than the arithmetic.

Ready to look at the architecture honestly?

If the same kind of client keeps ending up on your books, no single one of them caused that. The pattern is set upstream, in what you sell and who you say yes to, and it's genuinely hard to see from inside. It's hard to read the label when you're inside the jar. We'll look at the four frameworks underneath your business and tell you where the pattern is being set. Have a look at how we work, or just book the call.

Book the conversation →