Retention analysis helps you answer a simple question:
After customers start a subscription, how many are still subscribed months later?
This is one of the clearest ways to understand whether your subscription offer is “sticky” (customers keep it) or “leaky” (customers leave quickly).
In Kaching Subscriptions, Retention Analysis is shown as a table:
Each cell is shown as a percentage (example: 80%), meaning “80% of subscriptions from this cohort are still active at this milestone”.
What is a “cohort”?
A cohort is just a group of subscriptions that started around the same time.
In Kaching Subscriptions retention analysis report, a cohort is all subscriptions created in the same month.
Example:
2025-01 cohort: subscriptions created in January 2025
2025-02 cohort: subscriptions created in February 2025
Why cohorts are useful: they let you compare “apples to apples”. A January cohort has had more time to churn than a December cohort, so we track each cohort separately.
What is a “retention”?
Retention is the percentage of subscriptions in a cohort that are still active after a certain number of months:
Month 0 retention is always 100% because that’s the month the subscription started.
Month 1 retention answers: “After one month, what percent are still active?”
Month 2 retention answers: “After two months, what percent are still active?”
What is a “churn”?
In this report, a subscription is considered churned when it moves from ACTIVE to one of these states:
PAUSED
CANCELLED
EXPIRED
Once a subscription has churned, it will no longer count as retained for future months.
How to read the retention table (step by step)
Let’s use a simple example:
Cohort | Size | Month 0 | Month 1 | Month 2 | Month 3 |
Jan 2025 | 100 | 100% | 80% | 65% | 50% |
How to read it:
Size = 100: 100 subscriptions started in January 2025
Month 1 = 80%: about 80 of those 100 are still active after 1 month
Month 3 = 50%: about 50 of those 100 are still active after 3 months
Why don't I see percentages for for newer cohorts?
If a cohort is very recent, the report can’t measure later months yet.
Example: if it’s early March, the March cohort might show:
Month 0: 100%
Month 1+: “” (because Month 1 hasn’t fully elapsed for that cohort yet)
That’s expected.
What “good” vs “poor” retention looks like
There isn’t one universal “perfect” number - retention depends on your product, price, and subscription cadence.
But these patterns are useful:
Healthy retention: a gradual decline (customers churn slowly over time)
Early drop-off: a big fall from Month 0 → Month 1 (often signals onboarding/expectations/pricing issues)
Sudden change: a cohort that performs much worse than prior months (often tied to a change you made: price, shipping, offer, ads, landing page)
Tip: compare cohorts before and after an important change (new discount, new subscription box, new shipping policy, etc.).
FAQs
Q: Does retention mean “customers” or “subscriptions”?
A: This report is based on subscriptions (subscription contracts), grouped by the month they were created.
Q: Why is Month 0 always 100%?
A: Because Month 0 represents the starting month of the subscription. Every subscription in that cohort exists at the starting point.
Q: Which statuses count as churn?
A: The report treats churn as a move from ACTIVE to PAUSED, CANCELLED, or EXPIRED.
Q: Why isn’t there data for some future months?
A: Those months haven’t happened yet for that cohort. The report shows empty table cells when it can’t measure a month milestone yet.
Q: I just made changes - why didn’t the table update immediately?
A: Retention analysis results may be cached for up to 1 hour, so changes can take a little time to appear.
Need help? Drop us a message in the live chat, and we'll sort it out together.