The Customer Who Costs More Than They Pay (And How to Let Them Go)
Photo: MarkChristopherWebster, CC BY-SA 4.0, via Wikimedia Commons
Bad revenue has a way of showing up dressed like good news. The invoice clears. The logo looks nice on your homepage. Your MRR chart ticks up. Then the real bill arrives: a support queue that never sleeps, a roadmap that keeps detouring into custom requests, and a team that starts flinching every time that customer's name appears in Slack.
That's not a customer relationship. That's a subscription to chaos.
To be clear, this isn't about firing anyone who asks for help. Every early customer needs hand-holding. The problem is the bad-fit customer whose demands are structural, not temporary. They bought a product you don't really sell, or they bought the version of your company they wish existed. And every month they stay, they pull you further away from the business you're trying to build.
Revenue Is Not a Moral Victory
Founders love the top line because it's easy to celebrate. But revenue is only good if it survives the cost of earning and keeping it. A $1,000-per-month customer sounds great until you spend 12 hours a month supporting them, build two custom features they need, and lose a week of sales momentum because your lead engineer is stuck on their edge case.
In 2026, capital is not cheap and attention is even less so. You don't need to run a giant spreadsheet to see the problem, but you do need to put a number on it. Otherwise every conversation about firing a customer turns into a vibes debate.
The Lifetime Cost Worksheet
Here's the founder-friendly version. Open a spreadsheet and give each questionable account its own row. You're not trying to be perfect. You're trying to get close enough to make a decision.
Start with monthly revenue. Multiply by expected months retained. That's gross lifetime revenue.
Now subtract direct costs: hosting, third-party fees, payment processing, and any contractor time tied to that account.
Next, add support drag. Estimate hours per month, then multiply by a loaded hourly rate. If a support person costs $75,000 a year, their loaded rate is probably north of $50 per hour. An account that eats 10 hours a month is costing you about $500 before you count the context switching.
Then add roadmap distortion. This is the fuzzy one, so make it concrete. Did you delay a feature for your core market? Did you pay for a one-off integration? Did you spend a week on a security review that only one customer needed? Put a dollar figure on the engineering time, even if it's rough.
Finally, add the morale tax. This isn't soft. If your team dreads the account, if every check-in includes a complaint about them, if your best people are talking about workload because of one logo, that's a real cost. You can estimate it as a percentage of payroll, or you can just note it as a red flag. Either way, stop pretending it doesn't count.
When you stack those numbers, some accounts that looked profitable will be upside down. That's the uncomfortable math.
Warning Signs You're Paying to Keep Them
The math is easier when you know what to look for. Here are the signals that a customer is likely to cost more than they pay:
- They need a feature that is not on your roadmap and never will be.
- Their support volume is rising while their usage is flat or falling.
- They treat your product as a consulting engagement with a login screen.
- They threaten to leave every quarter, but never do because they know you'll bend.
- Your team starts routing around them instead of serving them.
- They're a reference customer in name only, and you'd never actually send a prospect to them.
None of these is a crime. They're just evidence of a mismatch. And mismatches get more expensive the longer you ignore them.
The Graceful Exit Script
Firing a customer doesn't have to mean torching the relationship. In fact, the best exits are boring. You want clarity, a timeline, and a soft landing. Here's a script you can adapt:
'We've been glad to work with you, and I want to be direct about where we are. Our product is built for [core use case], and your team needs [specific need] that we're not the right fit to support. We'd rather be honest now than keep charging you for a workaround. Here's a 60-day transition plan: we'll help you export your data, recommend two alternatives, and keep your account active through [date].'
That last part matters. Give them time, data, and a recommendation. If they're a good-fit customer in a bad moment, this conversation might actually save the relationship. If they're a bad fit, it ends things cleanly. Either way, you're not ghosting them.
One more thing: don't badmouth them internally. The story you tell your team matters. Frame it as a decision about focus, not a judgment about the customer. Otherwise your team learns to hide future bad-fit accounts instead of surfacing them.
What to Do With the Space You Reclaim
Firing a customer only works if you reinvest the oxygen. Put the support hours back into your core accounts. Use the engineering time to fix the thing your best customers actually need. Update your sales qualification so the next version of this customer gets filtered out earlier.
And track it. If you fire a $1,000-a-month account and your support queue drops by 15 hours a month, that's not a loss. That's a raise. If your roadmap gets one week back, that's a shipping cycle. If your team stops flinching, that's a culture win you can't put in a spreadsheet.
Bad revenue is still revenue, but it's not your revenue. The sooner you admit that, the sooner you can build a business that doesn't need to apologize for its customer list.