Guides & Benchmarks
Long-form research on why subscribers leave and what keeps them. Each page names the panel, survey or filing behind every number it quotes.
Churn Rate Benchmarks
What churn runs at by price point, industry, company size and billing period, with the panel behind every figure.
SaaS Churn Rate Benchmarks
SaaS churn by company size, customer size and pricing model, and why surveys and billing data disagree by forty points.
NRR and GRR Benchmarks
What net and gross revenue retention run at, and why the two tell different stories as companies grow.
DTC Subscription Retention
Consumer retention by cohort month, and monthly churn for five physical-goods verticals: most loss lands before month six.
How to Reduce Subscriber Churn
Churn nobody chose and churn that need not be permanent are the two biggest recoverable slices, and what to do about each.
Cancel Flows That Convert
The reasons subscribers actually give for canceling, ranked, and which save offer answers each one.
Increase Subscription LTV
Why LTV overestimates young cohorts and underestimates mature ones, who actually stays, and why it survives cancellation.
Subscription Cancellation Law
What the federal position actually is, what the agencies enforced anyway, and which state rules are stricter.
Run Your Own Numbers
Methodology & Sources
What these pages draw on
Every figure comes from a named publisher: a billing provider's own panel, a survey that states its sample, a regulator's filing, or a company's reported financials. Round-ups that collect other people's statistics do not count. A number republished without its method cannot be verified and often cannot be traced back to whoever first measured it.
How a figure gets onto a page
Each claim is recorded against a stored copy of the document it came from, together with the metric that document reports and the population it covers. Table rows name that stored entry, not a publisher, so a row cannot drift away from the source behind it. Where two publishers disagree, the page shows both figures and explains why they differ.
What gets withdrawn
Sources do not always survive a second look. Some turn out to be republishing data they never collected; others credit research they do not link to. When that happens, the figures resting on them come off the page. Keeping the number and dropping only the name would leave it uncited while still looking sourced.
Who compiled this
Researched and compiled by the SubJolt team. Spot something outdated or incorrect? Tell us and we will fix it.
Put These Numbers to Work
Build cancel flows that actually address your subscribers' concerns.