Monthly churn is the default churn metric for one reason: it is the fastest signal a subscription business has. Annual churn tells you what happened a year ago. Monthly churn tells you whether the fix you shipped two weeks ago is working.
How Joy Subscriptions handles this
Joy Subscriptions, a build-to-order subscription app for Shopify, attacks monthly churn at the cancel click and at the failed charge, separately.
Churn cohorts show when subscribers leave rather than only how many, so the fix targets a cycle instead of the whole base.
Why monthly is the standard
- Operational speed — most subscription billing cycles are monthly, so cancellations are naturally month-aligned.
- Comparable benchmarks — almost every SaaS and subscription benchmark study quotes monthly figures.
- Sensitive to interventions — a save-offer launch can show up in monthly churn within 30–60 days, while annual figures hide the change for a year.
The pitfalls of monthly churn
Monthly numbers swing on small samples. A subscription business with 500 active subscribers and 25 cancellations a month has a 5% monthly churn rate; one extra or one fewer cancellation moves the rate by 0.2 points. Three-month rolling averages smooth this noise out and are the more honest operating metric for small bases.
Seasonality also distorts monthly figures. Holiday signups churn faster than year-round ones. January cancellation spikes (New Year recommitments) and July dips (summer fatigue) are real. Year-over-year monthly comparisons help separate the trend from the season.
What to do with a monthly churn number
- Compare to last month and 3 months ago — directionality matters more than absolute level.
- Slice by signup cohort — month-1 customers churn faster than month-12 customers. Aggregate numbers hide this.
- Split voluntary from involuntary — different fixes, different ownership.
- Compare like seasons year-over-year — last December vs. this December, not December vs. last month.
For the broader analytical frame see churn rate analysis; for the underlying calculation see how to calculate churn.