K7 Insights

How to Fire a Client (Do the Math First)

By Kyle Christensen · September 8, 2026

Part of San Diego Business Coaching

Firing a client is a math decision before it’s a conversation. Before you write the email, you need three numbers: what the account earns you after your team’s hours, how much of your capacity it eats compared to what it pays, and what it costs you to replace it. Most of what you’ll find on how to fire a client skips all of that and hands you a script. The script is the easy part. I work with San Diego business owners who have a client they complain about every week, and when we run the numbers, the answer is a price change about as often as it’s a goodbye.

One more thing before we start. Before you fire the client, admit you hired them. We’ll come back to that.

When should you fire a client?

Fire a client when the account loses money or capacity after you’ve priced it properly and they won’t pay the real number. That’s the whole test. Annoying is a feeling. The numbers are a decision, and they’re usually more honest than the feeling, in one direction or the other.

Three numbers, one afternoon:

1. Margin on this account. Revenue minus every hour your team spends on them at a loaded cost, minus any direct expenses. Your company-wide margin doesn’t count here. This account’s does.
2. Share of capacity against share of revenue. If they’re 8% of your revenue and 25% of your team’s hours (or your inbox, or your phone at 9 pm), that gap is the number.
3. Replacement cost and replacement time. What one new client costs you to land and get up to speed, and how long your pipeline takes to produce one.

With those in hand you have four options: keep them, reprice them, rescope them, or fire them. Firing is the one most owners jump to and the one the math picks least often.

What does a bad client actually cost you?

Usually more than you think in hours and less than you think in revenue, which is why the feeling and the numbers disagree. Here’s what the math looks like on a real shape of account.

Say you run a 14-person marketing agency in Carlsbad. The account everyone dreads is a multi-location dental group paying $5,000 a month. Pull the time logs for the last 90 days and you find your team averages 58 hours a month on them: the account manager, the designer, the revisions, the “quick call” that never is. Your loaded cost is about $70 an hour. A typical account of yours at the same fee takes 34 hours.

Same fee, and one account earns you $1,680 a month less than the other. Add the six hours a month of your own time on their calls, which nobody logs, and the fact that they’re 9% of your revenue and 21% of your billable hours. The 24 extra hours a month they consume is most of a week of a designer’s time, and it’s exactly the capacity a normal account would sit in.

So the real cost of keeping them is roughly $1,700 a month in margin you’d earn with a client who took normal hours, plus your own evenings, plus whatever your team says about the account when you leave the room. Over a year that’s about $20,000 and a designer who’s quietly job-hunting.

Now the question changes. It stops being “can I stand them” and becomes “what’s a fair price for 58 hours a month,” and that’s a question you can put in writing.

If you got to the end of that table with a name already in your head, you’re most of the way to the decision. Grab 30 minutes with me and bring the hours. We’ll price the account together, and you’ll know which row you’re in before we hang up.

Before you fire them, admit you hired them

The client didn’t show up bad. You hired them, priced them, scoped them, and taught them what they could get away with, and every one of those was your call. Owning that is what keeps the next one from turning out the same way.

Four ways owners hire the client they end up firing:

• You sold them a package instead of asking what they needed. You had an offer, they didn’t quite fit it, and you took the money anyway. I wrote about this from the other side once: “The shortcut allows you to ignore the need to iterate your services to fit the demand.” A client buying the wrong-shaped service is going to need more of you than the price covers. That’s on the offer.
• You discounted to win them. A discount feels like goodwill on the day you give it. Later it feels like the client who asks for the most and pays the least, because that’s what you set up. I’ve said it before about price cuts: you have to have a conversation with someone on your team about why they don’t get their raise this year. San Diego owners cave on price more than they admit, and the client they resent most is usually the one they discounted.
• You never told them the rules. No scope in writing, no turnaround times, no definition of what a revision is. They didn’t become a bad client. They became the client you trained.
• You kept them because you had nothing else. We’ll get to this one, because it’s the one that decides whether you can fire anybody.

I did the first one for years. In a company I ran as its operator, we had strict rules about who we’d take on and what they had to buy on day one. Anyone who didn’t fit got squeezed into the offer or turned away, and the ones we squeezed became the “difficult accounts.” When we loosened the rules and let clients grow into us instead, sales went up, and a good share of the difficult accounts stopped being difficult, because they were finally buying something sized for them. It put more pressure on our account managers for a while. That was the trade, and it was the right one.

What does replacing them cost, and can you actually do it?

Replacing a client costs you the sales time to land one, whatever you spend on marketing to find one, and the first month or two of the relationship where the work runs slow. The honest number for most service businesses is one to three months of that account’s fee. That’s affordable if you have a pipeline. If you don’t, the bad client is a symptom, and firing them treats the symptom.

Here’s the line I keep coming back to: “The same reason you didn’t hit that new customer quota this year is probably the same reason you kept that bad-attitude employee around for another year.” You keep the bad client for the same reason. You have no options, so the account you’d never sell today stays because it’s the account you have.

The fix is upstream of the firing. The number one KPI I give owners is new friends per week. Coffee with the CPA in Kearny Mesa, the contractor you met at the chamber thing, the owner who sat next to you at the Padres game. Do that every week for a quarter and you have a stable of options, and with a stable of options a bad client is one phone call from replaceable. Without it, you’ll do what everybody does. You’ll wait, and complain, and renew, and blame the market for the client you chose to keep.

Put the two numbers together and the decision writes itself:

Common mistakes when firing a client

The mistakes come from doing the conversation before the math, or instead of it:

• Firing on the worst day. The Tuesday they yelled at your project manager is the day you decide to run the numbers. It’s the wrong day to send the email.
• Fading instead of ending. Slower replies, the B team, a hope that they’ll leave first. They notice, and San Diego is a small town for referrals. The owner who ends it cleanly gets recommended by the client he fired. It happens more than you’d think.
• Skipping the price increase. The real number is the most honest offer you can put in front of them. Most owners never offer it because they’ve already decided the client will say no. Let the client say no.
• Firing the whole category. One bad account and suddenly “we don’t take anyone under 50 employees.” That rule turns away most of the businesses in the county to avoid one relationship you could have priced correctly. Sounds like you’re just waiting for perfect.
• Writing the novel. Two paragraphs. No history lesson, no list of grievances. Anything longer is you arguing for a verdict they don’t get to appeal.
• Leaving them stranded. Finish what’s in flight, give real notice, hand them two names you’d trust. Read your agreement before the call, and if your work is regulated, check what your licensing body says about ending an engagement.

How do you actually fire a client?

Call them, then confirm it in writing. The call is short, and it’s the last step, which is why it’s the last thing in this article. Everything above it is the decision. This is the delivery.

The script, if the math said go:

“[Name], I want to be direct with you. We’ve looked hard at how we work together, and we’re not the right firm for what you need going forward. We’ll finish [the work in flight] and support you through [date]. I’ve got two firms I’d trust with this and I’m happy to make the introduction. Thank you for the business.”

Then an email the same day that restates the date, what you’ll deliver before it, and the handoff. If they ask why, tell them the truth in one sentence: the way you need this done costs more than we charge for it, and I’d rather end it well than deliver it resentfully. Don’t reopen it after that. Don’t itemize the past. If they counter with “what if we paid more,” and some will, you already know the number, because you did the math first.

Do the math this week

As clear cut as I can say it: firing a client is a decision you run, and the numbers are the decision. The script is four sentences. It only works after the math.

So pick the client you complain about most, pull 90 days of hours, and fill in the table above. That part is yours to do, and it takes an afternoon. Where I come in is the harder half: the pipeline that makes you able to say goodbye to anyone, and the pricing that keeps you from hiring the next one. Bring the finished table to a call, or start with how I work with San Diego owners.