The blended rate hides the distribution
Take a company with 1,000 customers that loses 60 over a year. Blended logo churn is 6%. Split by ACV band, the picture changes completely. Seven hundred SMB accounts lose 52, a 7.4% rate. Two hundred fifty mid-market accounts lose 7, a 2.8% rate. Fifty enterprise accounts lose 1, a 2% rate. The 6% figure describes none of these customers, and a retention program aimed at the average would spend most of its effort on a base that is holding fine.
Revenue weighting can flip the conclusion again. If the single enterprise account that left carried more ARR than all 52 SMB losses combined, logo churn points at SMB while revenue churn points at enterprise. Both readings are correct and they call for different work, which is why logo and dollar churn belong side by side inside every segment rather than stacked at the company level. The same discipline applies to net revenue retention.
Cut by what predicts cancellation
Useful segments separate customers whose churn behavior actually differs.
- Contract type: monthly, annual, multi-year - ACV band: the price tier that determines who signs and who can cancel - Acquisition channel: self-serve, inbound, outbound, partner - Onboarding outcome: reached the activation milestone or did not - Signup cohort: the quarter the customer landed
Channel and onboarding cuts tend to be the most actionable, because both point at something the company controls. A partner-sourced base churning at twice the inbound rate is a partner-qualification problem, not a customer success problem.
Keep the denominator fixed
Set the denominator as customers in that segment at the start of the period, decide how mid-period additions are treated, then leave the rule alone. Customers that upgrade across a segment boundary stay in their opening segment for the whole period, otherwise an upgrade can move a churn event between segments and make quarters incomparable.
Segment churn belongs in the revenue plan, not only the customer success dashboard. Renewal and expansion assumptions built on one blended rate misprice every segment at once, which is why a revenue forecast should carry its own retention assumption per segment and per contract type.
Frequently Asked Questions
Why segment churn instead of reporting one rate?
A blended rate averages populations that behave differently, so it describes no actual customer. A company at 6% blended logo churn can be running 7.4% in its SMB base and 2% in enterprise. The blended number tells you nothing about where to spend retention effort.
What segments are worth cutting churn by?
Cuts that predict cancellation rather than cuts that are convenient to pull. Contract type, ACV band, acquisition channel, and whether the account hit its onboarding milestone all separate customers who behave differently. Industry and region usually do not, unless the product is regulated or localized.
How do you handle customers who change segment mid-year?
Assign each customer to the segment it occupied at the start of the measurement period and hold it there for the whole period. Reassigning mid-period lets an upgrade quietly move a churn event from one segment to another and makes rates impossible to compare across quarters.
How small can a segment be before the rate is noise?
Once one cancellation moves the rate by more than a point, the segment is too thin to read month to month. A 50-account enterprise segment moves 2 points on a single loss. Use rolling twelve-month windows for thin segments and report the count of churned accounts next to the percentage.
Put these metrics to work
ORM builds custom revenue forecast models that turn concepts like churn rate by segment into prescriptive action for your team.
Schedule a Demo