We at Banyan AI have created this memo for SaaS founders, RevOps, CS, Product, Finance.
Our Goal: Turn churn from a “rate” into a measurable system of drivers you can diagnose, predict, and mitigate.
Why we have created this memo
Over the past months, we analyzed live data from 100+ SaaS companies, combining real billing, product, and CRM signals with trusted industry benchmarks.
The pattern is consistent: churn rarely happens suddenly, it builds quietly through adoption drops, support friction, pricing gaps, and payment failures. In most cases, revenue risk is visible weeks in advance when the right data is unified.
→ If you’d like to see what signals might already exist in your own data, we’re happy to show you how Banyan AI surfaces them in minutes.
SaaS churn is often reduced to one number, but it’s really two distinct mechanisms:
They behave differently and must be measured separately.
Benchmarks confirm this: Recurly reports median total churn around 4.1%, with 1.0% involuntary churn, a 49% dunning recovery rate, and recovery flows saving 72% of at-risk subscribers (median +141 days retained). ProfitWell has also shown that 20–40% of churn can be involuntary.
Our dataset confirms these findings, with median total churn slightly higher (4.4%) and significantly higher involuntary churn (1.5%, equal to 34% of total churn).
Implication: A meaningful share of churn is operational and recoverable, before you even touch product or customer success.
You can’t compare churn numbers without consistent definitions.