When to Fire a Client (Using Profit Data, Not Feelings)
Freelancers know the feeling. There's a client you dread hearing from. Scope creeps, emails multiply, revisions loop. You've told yourself they "pay okay" and you'd lose too much revenue if you fired them. The math usually disagrees.
This post is the framework for deciding when a client stops being worth it — using profit data, not feelings. And when you decide, there's a right way to actually do it.
The three inputs every fire-or-keep decision needs
- Profit per hour on this specific client over the last 3–6 months. Not billed rate. Profit per hour.
- Your current baseline profit per hour (aggregate across all clients, same period).
- The opportunity cost — realistic replacement. What would another 20 hours/month of your pipeline-building time generate at your baseline rate?
With those three, the keep-or-fire math becomes arithmetic.
The keep-or-fire threshold
If a client's profit per hour is below 60% of your baseline for 3+ consecutive months, and you've already tried one explicit fix (rate raise or scope tightening), they're a fire candidate.
The 60% threshold is where the client stops being a net positive once you factor in:
- The time you could spend on pipeline for higher-paying work (opportunity cost)
- The emotional overhead that doesn't show up in hours but affects your work quality elsewhere
- The risk that a bad client warps your professional standards (you start accepting scope creep that other clients will learn from)
Worked example: Jordan's three clients
Jordan is a freelance content strategist. Baseline profit per hour across all work: $95. Three clients over the last quarter:
| Client | Revenue/qtr | Hours | Profit/hr | % of baseline |
|---|---|---|---|---|
| Acme Co (retainer) | $12,000 | 80 | $137 | 144% |
| Bright Ltd (project) | $8,400 | 95 | $82 | 86% |
| Nightmare Inc (project) | $6,000 | 120 | $45 | 47% |
Nightmare is below 50% of baseline. The apparent reasons: scope creep (original 40-hour engagement grew to 120), unbilled revisions, email thread volume, meetings that expand.
The "but the revenue!" defense: Nightmare brings in $6,000 over the quarter. If Jordan fires them and reallocates those 120 hours to pipeline/client work at the $95 baseline, the potential is $11,400 — nearly 2× the Nightmare revenue.
Before you fire: two explicit fixes to try first
Fix 1: Rate raise with explicit scope
Email the client:
Hi {Client},
Reviewing our last quarter — the original scope was {X} and we've been doing closer to {Y}. I want to keep working with you, and I want the structure to reflect reality. For the next quarter, I'd propose:
• {New scope, explicit inclusions + exclusions}
• {New rate or retainer amount}
• {Revision policy — e.g., 2 rounds included, additional at $X/hr}
Let me know if this works or if we should discuss.
{Your name}
If they accept: problem solved. If they negotiate the scope and rate toward something still-profitable: problem solved (compromise). If they reject outright: that's useful data — the client was never going to scale with you.
Fix 2: Tighten scope without changing rate
For clients who balk at rate raises, try constraint instead. "For the next quarter, I can allocate {X hours/week} to your work; anything above that gets scheduled for the following quarter." This forces them to prioritize and cuts your unbilled-hour exposure.
How to fire a client (professionally)
Assume no damage from the relationship end. You don't owe them an explanation. You owe them clarity and reasonable transition support.
Template: Soft exit (keeping the door open)
Hi {Client},
I'm restructuring my client work for next year and scaling down to a smaller roster. After {date — give 30–60 days}, I won't be taking on new {work type} from {Company}. Happy to help you find a replacement — {name a peer who does similar work} is someone I'd vouch for.
Thanks for everything over the last {time period}.
{Your name}
Template: Hard exit (when the relationship is bad)
Hi {Client},
I've decided to end our working relationship. After {date — 30 days minimum}, I won't be taking on new work. I'll complete {list of in-progress deliverables} by {date}and provide a final invoice.
I'd recommend {referral} if you need a replacement quickly. Thanks for the work we did together.
{Your name}
No apologies. No lengthy explanations. Clients sometimes push back looking for a debate — don't take the bait. "I've decided" is a complete reason. If they ask for more, repeat "I've decided to focus my roster on a smaller number of engagements" and end the conversation.
Three clients you should probably fire immediately
- Late payers. Any client consistently past Net 30 on multiple invoices, without a valid structural reason, is signaling that your time isn't worth their attention. The profit per hour is also artificially high because it doesn't count the hours you spend chasing payment.
- Scope-thrashers. Clients who change direction mid-engagement multiple times per quarter. The explicit scope-tightening fix above rarely works with chronic thrashers; the behavior is cultural, not tactical.
- Emotional-overhead clients. Clients whose emails make you avoid opening them, whose meetings drain you, whose work leaks into your non-work time as mental overhead. This one doesn't show up in hours but shows up in your quality on every other engagement.
The post-fire math
After firing, the hours you recover go somewhere. Three patterns for what to do with them, in order of likelihood to actually raise your overall profit per hour:
- Prospecting for better-fit clients — the highest-ceiling use, but highest variance. Invest 30–50% of recovered hours here in the first month.
- Systematizing existing client work — build the templates, automations, and processes that raise profit per hour on every remaining client. Lower variance, immediate impact on existing engagements.
- Unpaid skill-building — appropriate only for specific capability gaps blocking you from a higher band. Easy to over-invest here.
Related reading: profit-per-hour benchmarks by trade for figuring out your baseline, the profit-per-project pillar for the underlying math.