What is retention rate and how to measure it
Understand retention rate: measure how many users return to your product, identify churn patterns, and improve long-term engagement.
What is retention rate and how to measure it
Quick answers
What is user retention rate? Retention rate is the percentage of users from a given cohort who are still active N periods later. It's always a cohort metric — you measure it for a specific group of users (e.g. everyone who signed up in January) and track how many are still active in month 2, month 3, and so on. Aggregate retention averages across cohorts and loses all the signal.
What is the difference between retention rate and churn rate? They're mathematical inverses — 85% retention = 15% churn. But they're used differently. Churn rate is typically a point-in-time measure (how many customers cancelled this month). Retention rate is a cohort curve (what percentage of January signups are still active in month 6). Cohort curves are more actionable because they show when users leave, not just that they do.
What does a healthy retention curve look like? A healthy curve drops steeply in the first weeks then flattens out — early churn is normal, but users who stay tend to keep staying. If the curve never flattens, you have a fundamental product-value problem. If it flattens at a very low level (5% by month 3), the product has a retention problem that messaging alone won't fix.
How do I read a cohort retention table? Each row is a cohort (users who signed up in a given month), each column is a time period (month 1, month 2, etc.), and each cell shows the percentage of that cohort still active at that time. Comparing rows tells you whether newer cohorts retain better than older ones. Improving rows over time means your product changes are working.
What are the three types of retention I should track? N-day retention (what % came back on day 1, 7, 30 — useful for early onboarding), unbounded retention (the monthly cohort curve — useful for long-term health), and feature retention (what % who used feature X in month 1 still use it in month 3 — useful for evaluating specific features).
How do I build a cohort retention table in Metabase? Write the cohort retention SQL in Metabase's SQL editor, save it, and display it as a pivot table — cohorts as rows, months as columns, retention % in cells. The Metabase retention analysis event walks through exactly how to build and read it. Add dashboard filters to slice by plan, channel, or segment without rewriting the query.
How do I know if a product change improved retention? Compare the cohort curve for users who signed up after the change to users who signed up before. If the post-change cohort retains at a higher percentage at month 2 or month 3, the change worked. Metabase's time series analysis and discover trends guide cover how to read these patterns without being fooled by noise.
---
Retention rate is the inverse of churn, but they don't answer the same question. Churn tells you how much you're losing. Retention tells you how much of what you built is sticking.
The distinction matters because they point to different problems — and different fixes.
The definition
User retention rate: The percentage of users from a given period who are still active N days/weeks/months later.
Retention Rate = (Users active at end of period who were also active at start) / Users active at start × 100
This is a cohort metric. You measure retention for a specific group — users who signed up in January — and track how many are still active in February, March, April, and so on. Aggregate retention averages across cohorts and loses the signal.
Retention vs. churn rate
These are inverses: if your monthly retention rate is 85%, your monthly churn rate is 15%.
But they're used differently:
- Churn rate is typically calculated as a point-in-time measure (this month's cancellations)
Cohort retention curves are more actionable. They show you when users leave, not just that they do.
What a retention curve tells you
A healthy retention curve drops steeply at first then flattens out — some users churn immediately, but those who stay tend to stay. If your curve never flattens, you have a fundamental product-value problem. If it flattens but at a very low level (say, 5% retention by month 3), you have a retention problem that the product alone won't fix.
Comparing cohort curves across different time periods tells you whether product changes are working. If January's cohort retains better at month 3 than December's, something you did in January improved retention.
The three types of retention to track
N-day retention: What % of users came back on day N? (Day 1, Day 7, Day 30). Useful for understanding early engagement.
Unbounded retention: What % of users from cohort X are still active after N months? The cohort curve above.
Feature retention: What % of users who used feature X in month 1 are still using it in month 3? Useful for evaluating whether specific features drive long-term engagement.
Putting it in Metabase
Save the cohort retention query in Metabase, display it as a pivot table, and add it to your product health dashboard. Use dashboard filters to slice by acquisition channel, plan, or user segment without rewriting the query.
For trend analysis — how retention is changing over time across cohorts — the time series analysis guide covers how to spot meaningful changes vs. noise. And for teams who want to dig deeper into the data behind retention drops, the discover trends guide is a good next step.
Define retention — which event counts, which time windows — in Data Studio. One definition, consistent across every dashboard that uses it.
---