Most founders will never fire a customer, and that is usually a mistake. Every early-stage company has one or two accounts that consume ten times their share of support, engineering, and executive attention while paying an average or below-average price. Left in place, these accounts warp the roadmap, burn out the team, and quietly become the reason your best engineer quits.
Firing a customer is one of the highest-ROI decisions a founder can make. Done cleanly, it recovers team capacity, sharpens focus, and often improves the NPS of your remaining accounts. Done badly, it turns into a public complaint that damages your brand.
A customer is a candidate for offboarding when three or more of these are true: they submit more than 5x the average support ticket volume, they demand custom work that no other customer would use, they escalate to the CEO monthly, they pay below-market pricing on a legacy plan, they use your product in a way that violates your terms, or they treat your team disrespectfully.
Before the conversation, add up the real cost: support hours, engineering hours on custom work, executive time on escalations, and the opportunity cost of the roadmap items you did not ship because you were serving them. Divide that cost by their ARR. If the ratio is worse than 1:1, you are paying them to be a customer.
Offboarding a customer is a 30-minute meeting, ideally with their most senior stakeholder, framed as a partnership decision rather than a punishment:
1. Thank them for being an early customer. 2. State clearly that your product roadmap and their needs have diverged. 3. Offer a specific alternative (a competitor, a service partner, an in-house build) and, where possible, an introduction. 4. Propose a wind-down timeline (typically 60-90 days) with data export support. 5. Offer a partial refund on any prepaid term.
Never cite the real reason if it is about their behavior. Cite product-market fit.
Write the offboarding email as if it will be posted on LinkedIn, because it might be. Warm, professional, and specific. Do not send anything you would not want quoted.
Most founders who fire a bad-fit customer report the same thing three months later: the team is happier, the roadmap is clearer, and the revenue impact was smaller than they feared.
