The retention curve in Mixpanel visualizes how many users return to your product after a specific period since their initial action (the "A event"). Each point on the curve (e.g., Day 1, Day 2) represents the percentage of users who came back and performed a follow-up action (the "B event") on that specific day after their first event. The curve typically declines over time, showing user drop-off, but there can be exceptions.
Why can Day 2 have a higher value than Day 1?
- The value for each day (e.g., 5.09% for Day 1, 12.5% for Day 2) represents the percentage of users who returned and did the B event exactly on that day after their initial event.
- It is possible for Day 2 to have a higher value than Day 1 if more users returned on Day 2 than on Day 1. This can happen due to user behavior patterns, such as users preferring to return after a longer interval, or due to external factors like marketing campaigns or product updates that encourage users to come back on a specific day.
- The retention curve is calculated based on unique users, not event totals, and each user is only counted once per bucket (day) but can appear in multiple buckets if they return on multiple days. The first column (Date) indicates when the user performed the A event, and subsequent columns show the percentage retained on each day after that event Docs: Retention Basic Features.
Example:
If 100 users signed up on Day 0:
- 5 users returned on Day 1 (5.09% retention)
- 12 users returned on Day 2 (12.5% retention)
This means more users came back on Day 2 than on Day 1, which is reflected in the curve.
Note: The retention curve can show such patterns depending on your user base and product usage habits. It’s not always strictly decreasing, especially if there are periodic events or incentives that drive users back at specific intervals.
For more details on how the retention curve and user cohort buckets work, see the documentation here.
