glossary · metrics
Churn rate definition for product managers
The percentage of customers (or revenue) lost in a given period, measured against the starting count.
Churn rate is the percentage of customers or revenue lost in a period. It is also the market’s most honest verdict on whether a product is genuinely lovable and whether the value proposition holds at the price charged. Interviewers test churn because it sits at the intersection of product, billing, and positioning problems, and strong candidates separate those layers before prescribing anything.
The formulas
Customer (logo) churn: (Customers lost / Customers at start of period) × 100. For periods with heavy new additions, use the average base: (Churned / ((Opening + Closing) / 2)) × 100.
Revenue (MRR) churn: (MRR lost from churn + downgrades) / MRR at start of period × 100.
Customer lifetime: ~1 / monthly churn rate. At 5% monthly, expected lifespan is about 20 months, and you replace over 60% of revenue annually just to stay flat.
Logo churn, revenue churn, and NRR
- Logo churn: accounts lost as a percentage of total. Shows breadth of the problem.
- Revenue churn: MRR lost weighted by contract size. Low logo churn can mask severe revenue churn if large accounts are downgrading.
- Net Revenue Retention (NRR): includes expansion. NRR above 100% means existing-account growth exceeds losses. Median B2B SaaS NRR is 106%; enterprise best-in-class is 115-130%; SMB typically 90-105%. Sophisticated interviewers reach for NRR first because it captures churn, contraction, and upsell in one number.
Definitional hygiene matters: disagreements on what counts as churn (pauses, partial downgrades, non-renewals at term end) produce 2-5x variance in the reported number. Name the definition before citing any figure.
Voluntary vs. involuntary churn
Voluntary churn is an active cancellation. Root cause sits in one of three layers: activation failure (never got value), engagement failure (no habit formed), or value-delivery failure (product stopped earning its keep). About 40% of churned customers cite price and 30% cite feature gaps, but discounting and shipping features rarely moves the number. Laddering past the first stated reason almost always surfaces an activation or value-delivery problem.
Involuntary churn is a payment failure. Expired cards account for 42% of all payment failures. Smart retry logic recovers 68% of failed payments vs. 23% with a single retry, representing roughly $1.3 billion in recoverable SaaS revenue annually. Zero product changes required: this is the easiest PM win in retention and the most overlooked.
Benchmarks
| Segment | Monthly churn |
|---|---|
| Enterprise | Under 1.5% |
| Mid-market | Around 2.8% |
| SMB | Around 6.4% |
2025 Recurly median B2B SaaS annual churn: 3.5% (2.6% voluntary, 0.8-0.9% involuntary). The gap between 3% and 8% annual logo churn produces a 2-3x valuation multiple difference in the $3-20M ARR range.
The 2026 AI-native churn regime
Budget AI tools (under $50/month) show 23% GRR and 32% NRR, losing 77% of customers annually. Premium AI tools (over $250/month) with deep workflow integration show 70% GRR and 85% NRR. Feasibility being cheap does not protect you: if the product does not embed into how people work and deliver outcomes they would pay to keep, churn tells you before your runway does. Viable and lovable are still the bar.
How to structure a churn diagnosis
strong
"First I'd align on the definition: are we counting downgrades and pauses? A definitional gap can move the figure 2-5x. Then I separate logo from revenue churn, check NRR, and pull payment failure data first: involuntary churn is often 20-30% of the total and fully recoverable with dunning changes. For voluntary churn I run cohort analysis: rising churn in recent cohorts points to onboarding failure; uniform churn across cohort age points to value-delivery failure. I'd ladder qualitative interviews past the first stated reason, because price and features are what customers say, not usually what's true. The layer tells me the fix: activation failure means shortening time-to-value; engagement failure means building toward the sticky action; value-delivery failure is a scope and pricing conversation."
weak
"I'd survey churned customers and improve onboarding." Skips the voluntary/involuntary split, takes survey data at face value, conflates activation with value delivery, and produces no layer-specific prescription. Would not clear the bar on a metrics question at any tier-1 company.
For related concepts, see retention, cohort analysis, and LTV.