AppsFlyer's Retention Report

Retention is a measure of how many existing users are active in your app. The goal is to keep engaged users coming back for more. Retention is one of the most important metrics, allowing you to track app’s open rate over time. This metric can be extremely useful for:

  • Showing which media sources bring the most engaged users
  • Enabling LTV optimizations

  • Measuring product usability after version updates

How Retention is Calculated

For any single day, the retention is calculated as a ratio between the two following numbers:

  • The number of active users (one or more sessions) on that day 

  • The total number of installing users during the selected date range 

Viewing the Retention Report

  1. Click Retention under the Reports section of the Dashboard.

  2. Select the time period for which you want to view the Retention Report and click Go

  3. You can use the filters you want to add to the Retention Report.

  4. You can select the dimensions by which you want to aggregate the report.

Which Dimensions Can be Specified in the AppsFlyer Retention Report?

Cohort Time Frame Selector

Use this window to define the date range for the install date. Every user who launches the app during the designated date range is counted in the “Install Day (for Daily Retention Report)/Install Week (for weekly Retention Report)” column, comprising the base value (100%) for the retention calculation.

 Note

AppsFlyer considers the counting of a day based on the UTC calendar date and not based on the number of hours since install.  For example, if an install occurs on 1 Feb at 22:44 and an app open session takes place on 2 Feb at 01:12 - the app open is shown on Day 1 even though it occurred less than 3 hours after the install..


The selected value represents the time frame units of user activity, since the first launch:
  • For Daily, each column represents a day of activity since the date of the initial installation

  • For Weekly, each column represents a week of activity since the week of the initial installation

Retention Filter and Group By Settings

Allows to drill-down the retention statistics using a variety of dimensions, you can select a maximum of four dimensions:

  • Min. Cohort Size: Only rows with an install number equal to or greater than the selected minimum size are displayed. The default value is 10.

  • Media Source: The networks where you ran your campaign (examples: organic, Facebook, Google AdWords and many others)

  • Campaign: Your campaign names with each network

  • fb_adset and fb_adgroup: For Facebook’s-related segmentations

  • geo, geo_city: The country and the city where an installation occurred

  • af_siteid: The site IDs of where the advertisements appeared

  • sub1-sub5: Custom parameters

  • wifi: Is set to “true” if the installation occurred over WiFi

  • install_os_version: The version of the mobile OS

  • brand: The brand of the mobile device

  • af_prt: The Agency

  • af_adset:  The Ad Set

  • af_ad:  The Ad

  • af_channel:  The Channel

  • af_keywords:  The Keywords

  • app_version:  The App Version number

  • sdk_version:  The version number of the SDK

  • attribution_touch_type (impression / click): The touch type

The results can be grouped by a combination of any of the above dimensions (up to four values can be selected at once).

Examples

Basic Retention Calculation

  • The selected date range is: Jan 1 - Jan 2

  • The selected time interval is Daily

  • There were ten installations during the selected date range (Jan 1 - Jan 2)

  • No-one used the app on the first day since the install date

  • Five users used the app on the second day

The report appears as follows:

Media Source Install Day Day 1 Day 2

XXX

100%

10

0%

0

50%

5

The Retention Report is NOT Cumulative

  • The selected date range is: Jan 1 - Jan 2

  • The selected time interval is Daily

  • There were ten installations during the selected date range (Jan 1 - Jan 2)

  • Two  users used the app on the first day since the install date

  • Five users used the app on the second day

The report appears as follows:

Media Source Install Day Day 1 Day 2

XXX

100%

10

20%

2

50%

5

The report is not cumulative: i.e. users counted in the Day 1 column are NOT counted in the Day 2 column.

The Difference Between Daily and Weekly Aggregations

  • The selected date range is: Jan 1 - Jan 14

  • There is only one user that installed the app on Jan 1

  • Each further day (Jan 2  and beyond) the user used the app

The daily report appears as follows:

Media Source Install Day Day 1 Day 2 ... Day 13

XXX

100%

1

100%

1

100%

1

...

100%

1

In this report each day counts one, and the sum for the seven days is seven.

The weekly report appears as follows:

Media Source Install Week Week 1

XXX

100%

1

100%

1

In the weekly report, the retention count for all the seven days of every week is exactly one, since the app was launched by the same individual user on a daily basis.

Group by Install Date

  • The selected date range is: Jan 05 - Jan 12

  • The selected time interval is Daily

  • Filters are grouped by Install Date. You can add up to four filters to drill-down the retention statistics further

The Daily Report appears as follows:

Retention Data Freshness

Every day at Midnight UTC, AppsFlyer calculates the number of unique sessions of the ended previous day, for all apps, according to their set time zone

For users in time zone UTC- (minus) it may take up to 48 hours, and for users in time zone UTC+ (plus) it may take up to 24 hours, for the report to be populated with data (see example below).

For weekly aggregations of data, the time zone differences are negligible.

 Example

com.noodles is a Chinese foodie app defined in the Beijing time zone UTC-8.
com.burgers is an American foodie app defined in the San-Francisco time zone UTC+8.
On Tuesday midnight UTC (between Monday and Tuesday) AppsFlyer calculates the retention numbers for all apps.
On that time, in Beijing the local time is Tuesday 8:00 AM, and Monday's data is used and aggregated.
In San-Francisco, however, the local time is Monday 4:00 PM, and since Monday hasn't ended yet, Sunday's data is used and aggregated.  

Limitations

  • You can set the Min Cohort Size field to show retention data only for groups above the specified number of users. By default groups with under 10 users are not displayed.

  • Due to UI limitations, only 200 results are displayed.

  • The filter values are not dependent. For example, even when selecting Facebook as a media source, the campaign names from ALL the sources are displayed. 

Was this article helpful?
1 out of 1 found this helpful