Retention tells you whether users are returning to your product over time. Our retention report helps you analyze how your product has grown over time and what's making your customers return to your platform.
Note: The retention report presents data for identified users only. Anonymous visitors aren't included.
Interpret the chart
First, choose how to filter your data with the following options:
Source. Choose which activity and app you want to include in your data set.
- Activity. Choose the type of activity that you want to check the retention for—Pages, Features, or Track Events. The default setting is all activity.
- App. This dropdown allows you to filter to a specific group (all web or all mobile apps) or single app.
- Activity Selector. If you select Pages, Features, or Track Events in the main Activity dropdown, then this dropdown appears so you can choose the particular Page, Feature, or Track Event that you want to measure.
- Cohort Type. Review your results in visitors or accounts.
- Cohort Size. Choose if you want to divide your user base by Month or Week based on their first visit date. In retention analytics, a month is defined as 30 days and a week is defined as 7 days.
- Segment. The default segment is everyone who visited your product or app. You can choose your desired target group as needed.
- Date Range. Based on the selected cohort size, data is displayed for the last 6, 9, or 12 months or weeks.
The default view shows you all of your unique visitors in the last six months using a 1 Month cohort size for any activity across all of your apps.
Each row is referred to as a cohort, which is why retention analytics is also known as cohort analytics.
To help illustrate the chart interpretation, use the following assumptions while interpreting the sample retention cohort:
- Today’s date is October 15, 2018
- User A’s first visit recorded by Pendo is on April 15, 2018
- User B’s first visit recorded by Pendo is April 30, 2018
Both Users A and B are included in the April 2018 cohort since their first visit date is in April.
This means each individual visitor or account has its own 30-day or 7-day timeframe that rolls up into a cohort. Using the example assumption above, learn how to interpret parts of the chart in the sections below.
Here's how each user’s time frame is calculated when using a “1 Month” cohort size (30-day increments):
Similarly, here's how each user’s time frame is calculated when using a “1 Week” cohort size (7-day increments):
Within a row, each cell shows you the retention of the cohort—meaning how many users returned to your product or application after time (months/weeks) has passed.
For example, imagine you had a total of 1,521 visitors in May 2018. For Month 1, 93% of your users returned. This means about 1,414 visitors out of 1,521 visitors used your product or application again within a 30-day period.
All cells are independent of each other, so there might be times when you notice an increase in percentage—meaning more people visited your application in that time frame.
Cells with asterisks
You might see an asterisk (*) to the percentage. This means that the number is subject to change because there's still time remaining for the month (30-day increment) for that cohort.
Using the previous example, if you continued the individual timeframe breakdown, Month 5 and Month 6 are still incomplete for both Users A and B. Future timeframes are highlighted in pink to illustrate why Month 5 and Month 6 have an asterisk in the cohort:
So your chart might look like this:
The shade of color in the chart is based on the following percentages:
Interact with the chart
If you select a cell within the chart, a panel on the right-hand side opens and displays the list of the visitors that are included in the calculation. The list displays the total number of users forming that particular cohort percentage and up to 250 Visitor IDs. Selecting a Visitor ID directs you to that visitor’s detail page.
Similarly, you can select the Dropped Visitors tab to see everyone (up to 250 visitors) who was present in the previous period but haven't returned in the current period.
Toward the bottom of the list, you can select Export CSV to download all of the Visitor or Account IDs from the selected cohort with your preferred metadata values in the CSV file.
Segments and activity retention
Choosing a feature in the Activity filter is different than choosing a segment option:
- When using a segment filter, you're measuring the retention of the segment of users who have used the feature ever. The retention percentage still increases if they've used anything in the application.
- When using the Source Activity filter by feature, unlike the segment filter, retention percentage increases only when the users interact with that particular page or feature. This way, you can measure the retention of a Page, Feature, or Track Event for any segment,such as retention of the NPS page for NPS promoters.
Choosing All Activity and Visitor Cohort Type with an account-based segment could pull in different or more visitors than when using the Everyone filter:
- When using the Everyone segment, you're measuring the retention of the first visit date of a visitor attached to any Account ID.
When using a segment with an account-level based rule (for example,
Account Id = Acme Org), you're measuring the retention of a visitor's first visit date under an account that fits the segment rule.
As an example, imagine the retention chart is set for All Activity, Cohort Type Visitors, 1 Week, for last 6 Weeks (April 1–May 13, 2019).
- Visitor A whose first visit date was March 1, 2018 wouldn't show up in the report if the Segment was set to everyone.
- Visitor A would show up in the report if the Segment was set to
Account Id = Acme Org, and the first event that had Visitor A attached to Account Acme Org was May 1.
- As an example, imagine the retention chart is set for All Activity, Cohort Type Visitors, 1 Week, for last 6 Weeks (April 1–May 13, 2019).
Default date ranges for weekly cohort size
When using a cohort size of 1 week, the default date ranges of Last 6 Weeks, Last 9 Weeks, and Last 12 Weeks are adjusted to ensure the cohorts include an entire week of data and each weekly cohort starts on a Sunday.
For example, if today's date is Wednesday, May 6 and for a cohort size of 1 Week and a date range of Last 6 Weeks, the retention chart shows data for the following 6 weekly cohorts.
- Sunday, Mar 22 to Saturday, Mar 28
- Sunday, Mar 29 to Saturday, Apr 4
- Sunday, Apr 5 to Saturday, Apr 11
- Sunday, Apr 12 to Saturday, Apr 18
- Sunday, Apr 19 to Saturday, Apr 25
- Sunday, Apr 26 to Saturday, May 2
Sunday May 3 through Wednesday May 6 isn't included in this default Last 6 Weeks date range since those dates would only create a partial week.