Retention
Measure how many users return over time
Retention analysis shows how many users come back to your site over time, revealing whether users find lasting value. On Rybbit Cloud, retention needs a Standard plan or higher.
Understanding Cohorts
A cohort is a group of new users who share a first-visit date. For example, all users who first visited on January 15th form one cohort. A user counts as new if they had no visit in the 90 days before the selected range. Someone returning after a longer gap counts as new again. Filters apply to each user's first visit.
Switch Cohort period between Daily and Weekly. Weekly cohorts start on Monday.
Stats
The band at the top shows the average retention for the first periods (for example Day 1 retention), Avg. cohort size, and Best cohort: the cohort with the highest day 1 or week 1 retention among cohorts at least half the average size. With a comparison period on, each figure shows its change.
The chart shows average retention by period since first visit. Day 0 (or Week 0) is always 100%, so the curve starts at period 1.
Reading the Retention Table
Rows
Each row is a cohort, grouped by first visit date.
Columns
- Day/Week 0 - Starting point (always 100%)
- Day/Week 1, 2, 3... - Users who were active again in that period
Numbers
Use Show cells as to switch between percentages (%) and Users. 40% in the "Day 7" column means 40% of that cohort was active again seven days after their first visit. Click a cell for details on who came back.
Cells for periods that haven't finished yet are marked as in progress, and their figures can still rise.
Color Coding
Stronger color means higher retention. The legend above the table shows the scale.
Click Export CSV to download the table.
Key Insights
- Day 1 retention - Did users find enough value to return the next day?
- Day 7 retention - Did users integrate your product into their routine?
- Day 30 retention - Have users formed a habit?
Common Patterns
- Cliff drop - Sharp decline in first few days indicates onboarding issues
- Steady decline - Gradual drop suggests novelty wearing off
- Plateau - When retention levels off, you've found your core users
Note on User ID Salting
If User ID Salting is enabled, retention beyond 24 hours becomes less accurate since user identifiers reset daily.