Stop tracking blended churn: segment by deal size or miss the pattern
Churn is not a GAAP metric. No universal definition. Public companies and startups all define it differently, and often hide problems by doing so. Jason Lemkin learned this early as a B2B CEO. He compared his company to a public competitor that excluded churn in the first 60 days, calling it a trial period. The real reason: churn was much higher in those first 60 days. ## The three-segment model Lemkin segmented churn into three tiers at his company: **Single-seat and sub-$99/month deals:** 3% monthly churn by revenue. Credit cards expire, jobs change, solopreneurs go under. This segment turns over fast. **$99 to $999/month deals:** 100% net revenue retention after churn. Similar to HubSpot and Zendesk in this zone. **$10k to $100k+ deals:** 120% net revenue retention. Matches patterns at Box, Salesforce, and other enterprise-focused companies. As the company scaled, category sizes expanded (big deals went from $12k to $120k annually), but the segmentation and metrics stayed consistent. ## Why this matters for sales Churn should be naturally higher in smaller business segments than larger ones. If you do not segment it, no one will see that pattern. Blended churn metrics make big customer retention look worse than it is. You end up trying to solve different churn problems with the same answers. Keeping enterprise customers happy is not the same as stopping small businesses from switching to cheaper tools. Track where the money goes. Even if churn looks high by certain metrics, if you retain most of the revenue, that might be fine. Want to exclude trials and POCs from churn? Fine. Just do not call them recurring revenue. Segment them out as POC/trial revenue. Do not count deals as core MRR/ARR until the trial converts. Churn is often much lower when you separate POCs from post-trial revenue. ## Segment NPS and CSAT too NPS and CSAT vary widely by segment. Double down on the happiest segments. Usually the ones growing fastest. You will see new patterns. Maybe churn is not as bad as you think. Or at least different than you think. Big companies stay for years if you provide a real solution. Individual and small customers come and go, even when happy. Blended churn metrics only confuse things. Segment churn and track the money.