Rolling Retention¶
Rolling Retention provides you with a measure of user churn. Of all the users who came in once over a specified period of time, how many are still around three months later? Understanding how quickly users drop off is crucial for you as you assess the vitality of the product.
Rolling Retention displays the percentage of users still active N or more days after they first installed and launched your app. It is calculated as the ratio of the number of users whose last day of activity is past day N to the number of users who could have been active on day N.
Rolling Retention is located in the Retention section in the left navigation menu.
Let’s walk through this by viewing how rolling retention is displayed in a bar chart and heatmap.
Bar Chart Display¶
The bar chart below the rolling retention (viewed by day) for an app.
You’ll first notice that Day 0 Rolling Retention is at 100%.
Day 0 is the day that a user first launched the application on their device. This might be last month for User #1, last week for User #2, and so on. This bar is at 100% because it includes all users from the date range filter selected at the top right of the page, and they have all launched the app at least once.
To understand the mechanics of rolling retention, let’s say User #1 opened the app for the first time on Day 0 and only opened the app for the second time four days later on Day 4. This user would be added to the tally for Day 4 but also to the tallies for Days 1-3. This is because even though it took this user 4 days to come back to the app, it was still part of their consideration set and they had not churned.
If we look at Day 1 Rolling Retention for this app in the chart, it shows as 60%. This means that 60% of users return to the app on Day 1 or after. The converse is that 40% (or 100%-60%) of users do not return to the app at any point after Day 0, the day that they first opened it.
Rolling Retention can thus be helpful in understanding key drop off points. You can further view this data for a particular segment (e.g. Females, In-app Purchasers) by selecting the desired group from the All Users dropdown at the top right of the page.
Rolling Retention becomes even more powerful when filtered by a segment. Do females of a certain age range drop off from using your app faster than males? Do users who make an in-app purchase within my app stick around longer?
Heatmap Display¶
In this display, each row represents a single cohort (daily, weekly or monthly) and each column represents the Rolling Retention for the cohort on the specific period post-install.
Each cell is color-coded based on the Rolling Retention value in the cell. Lighter shade indicates lower retention and darker shade indicates higher retention. All other gradient shades indicate retention somewhere in between the lowest and highest values for this app. This range is represented on the top left of the table.
The last row in this display provides an overall summary of the Rolling Retention for users who installed the app during the date range that you have selected.