The Retention Report shows how many users returned to the app a certain number of days after installing the app or opening the app through a deep link.
You can identify the channels and campaigns that attract the most active app users. The retention data can be used for campaign optimization and ad billings.
Set the Start Event and Return Event to see the retention by different campaign types.
Start Event
The Start Event determines the type of campaign that acquired the user.
Installs refer to the events when users install the app and open it for the first time. Set the Start Event to Installs if you want to see the retention by UA campaigns.
Deeplink Opens refer to the events when users open the app through a deep link. Set the Start Event to Deeplink Opens if you want to see the retention by re-engagement campaigns.
Deeplink Pageview (App)
Add custom events
Some custom events can be added as Start Event. For more details, contact your CSM.
Return Event
The Return Event is an in-app event that indicates the user’s re-engagement with the app.
Any Event refers to any in-app event performed by the user.
Example) To check the retention rate of users acquired by a UA campaign, set the Start Event to ‘Installs’ and the Return Event to ‘Any Event.’
By setting the Measurement Option, you can decide whether to view the retention by unique users or by user journeys that are separated by Start Events.
What is a User Journey?
A User Journey is a path that a user takes from viewing or clicking an ad and installing the app, performing various in-app events, to eventually leaving or deleting the app. All in-app events are identified as Return Events contributing to retention counts.
A User Journey of a unique user is initiated by the Start Event performed by that user. The Airbridge Device ID is used to identify unique users.
If the unique user performs a new Start Event after the first Start Event, a new User Journey is initiated, and the first User Journey is considered to be terminated.
Any in-app event is identified as a Return Event for the User Journey initiated by a Start Event right before that in-app event.
Airbridge Device ID
Airbridge Device ID is the UUID used to identify unique devices. By default, GAID (Android) and IDFV (iOS) are used as the Airbridge Device ID. In case GAID or IDFV could not be collected or is unavailable due to the restrictions of LAT or ATT, a UUID randomly generated by the Airbridge SDK is used as the Airbridge Device ID.
A User Journey of a unique user is initiated by the Start Event performed by that user. The Airbridge Device ID is used to identify unique users.
If the unique user performs a new Start Event after the first Start Event, a new User Journey is initiated, and the first User Journey is considered to be terminated.
Any in-app event is identified as a Return Event for the User Journey initiated by a Start Event right before that in-app event.
User Event History of ‘User X’
<User Journey A>
2022/03/01 (Day 0): App Install (attributed to Kakao)
2022/03/02 (Day 1): App Open (attributed to Kakao)
2022/03/02 (Day 1): App Delete
<User Journey B>
2022/03/03 (Day 2): App Re-install (attributed to Moloco ads)
2022/03/05 (Day 4): App Open (unattributed)
Report Configuration
Analysis Period: 2022/03/1 ~ 2022/03/05
Start Event: Installs (App)
Return Event: Any Event (App)
GroupBy: Channel
Report Result per Measurement Option
1. When the Measurement Option is set as ‘General’
→ The retention rate is calculated per unique user distinguished by the Airbridge Device ID.
The App Open event on 2022/03/05 is counted as
User Journey A’s Return Event on Day 4, and also as
User Journey B’s Return Event on Day 2
Channel | Date | Total Install | Day 0 | Day 1 | Day 2 | Day 3 | Day 4 |
|---|---|---|---|---|---|---|---|
Kakao | 3/1 | 1 | 1 | 1 | 1 | 0 | 1 |
| 3/2 | 0 | 0 | 0 | 0 | 0 |
|
| 3/3 | 0 | 0 | 0 | 0 |
|
|
| 3/4 | 0 | 0 | 0 |
|
|
|
| 3/5 | 0 | 0 |
|
|
|
|
Moloco | 3/1 | 0 | 0 | 0 | 0 | 0 | 0 |
| 3/2 | 0 | 0 | 0 | 0 | 0 |
|
| 3/3 | 1 | 1 | 0 | 1 |
|
|
| 3/4 | 0 | 0 |
|
|
|
|
| 3/5 | 0 |
|
|
|
|
|
2. When the Measurement Option is set as ‘Confined’
→ The retention rate is calculated per User Journey distinguished by Start Events.
The App Open event on 2022/03/05 is not identified as User Journey A’s Return Event and does not contribute to any retention counts.
The App Open event on 2022/03/05 is identified as User Journey B’s Return Event contributing to Day 2 retention counts.
Channel | Date | Total Install | Day 0 | Day 1 | Day 2 | Day 3 | Day 4 |
|---|---|---|---|---|---|---|---|
Kakao | 3/1 | 1 | 1 | 1 | 0 | 0 | 0 |
| 3/2 | 0 | 0 | 0 | 0 | 0 |
|
| 3/3 | 0 | 0 | 0 | 0 |
|
|
| 3/4 | 0 | 0 | 0 |
|
|
|
| 3/5 | 0 | 0 |
|
|
|
|
Moloco | 3/1 | 0 | 0 | 0 | 0 | 0 | 0 |
| 3/2 | 0 | 0 | 0 | 0 | 0 |
|
| 3/3 | 1 | 1 | 0 | 1 |
|
|
| 3/4 | 0 | 0 |
|
|
|
|
| 3/5 | 0 |
|
|
|
|
|
Set the GroupBy and Filter options to see the most relevant data for you.
The GroupBy and Filter options are in the configuration box, where you can easily add or delete fields to customize your view.
Attention
All values displayed in the Revenue Report with the GroupBy and Filter options applied are based on users who performed the Start Event.
A cohort is a group of users that can be grouped together by their behaviors.
By using the cohort filter, you can segment users by events they performed or didn’t perform and get a more granular view of retention.
Example) By setting the configuration box as below, you can see the retention rate of users who installed the app between 2022-05-01 to 2022-05-31 and completed more than one order in the past 30 days.
Select Event
Only in-app events can be selected.
Any Eventrefers to all events performed within the selected period.
You can set a cohort of users who ‘performed’ or ‘didn’t perform’ an in-app event.
Select Event Property
You can add Event Properties (e.g., Device Type, Product Price, etc.) by clicking the “+ had Property” button.
Comparison operators (
>, ≧, <, ≦) are available for setting Event Property fields with numeric values.equals(=) 0 time(s), less than or equals(≦) 0 time(s), less than(<) 1 time(s)The above Event Property setting examples for Event Counts indicate a zero number of occurrences of the Event, and therefore, the cohort is defined as a group of users who didn’t perform the Event.
The analysis period can be set using three options:
Between: You can select the start and end date to define the analysis period. The start and end date must be within 174 days before the current date.
Since: The end date is fixed to the current date. The start date can be selected within 30 days before the current date.
Last: You can enter the number of days from 0 to 174. If you enter 30, the analysis period is set from 30 days prior to the current date until the day before the current date. When checking the ‘Include Today’ checkbox, the analysis period is set from 30 days prior to the current date to the current date.
Once you customize your Retention Report following the instruction above, you can see how many users performed the Start Event and how many of them performed the Return Event on Day N.
In the screenshot example below, “1,147” users were acquired on 2022/11/25 through google.adwords. On Day 1, which is 2022/11/26, “490” users came back to the app and performed the Return Event.
To get more granular, you can create a cohort of users who did or didn’t perform a certain event and view the retention of that particular cohort.
In the screenshot example below, a cohort has been defined as a group of users who performed the Install event and performed the Add To Cart event but didn’t perform the Order Complete event during the analysis period of 2022/11/24 to 2022/11/30. The Retention Report shows the N-Day retention by media channels.
When you click the ‘>’ button on the leftmost side of the table, you can see the daily breakdown of the retention data. The shape of the button changes to ‘∧’ when you click on it.
The cells in each column represent different dates depending on the dates displayed in the individual rows. For example, if the analysis period is set from 2022-11-24 to 2022-11-30, the Day 0 column holds the data of the dates as indicated in the below screenshot example.
The report can be downloaded as a CSV file or google sheet file. The download time may vary depending on the amount of data.
If you want to integrate the report into an existing sheet, you can copy the report with one click and paste it into any spreadsheet.
Q. The install counts in the Retention Report and the ones in the Actuals Report don’t match. Why is there a discrepancy?
The install counts in the Retention Report tracked within the analysis period don’t include multiple installs performed by a unique Airbridge Device ID, and therefore the install counts in the Retention Report can match or be less than the install counts in the Actuals Report. If two install events occurred by a unique Airbridge Device ID on a single day within the analysis period, only the last install is counted for retention measurement. If two installs occurred by a unique Airbridge Device ID on separate days within the analysis period, both installs are counted separately for retention measurement.
Example)
Analysis period: 2022/02/03 ~ 2022/02/10
If multiple installs are performed by a unique Airbridge Device ID:
Scenario 1: If two installs occurred at 13:00 and 20:00 on 2022/02/04, only the install that occurred at 20:00 is counted for retention measurement.
Scenario 2: If two installs occurred on 2022/02/04 and 2022/02/07, both installs are counted for retention measurement.
Day 0 is the start date of the analysis period, which includes the Start Event performed by the users. The Start Event is always included in the Return Event, which is set to Any Event (App) by default. The Retention Report shows the retention of the users who performed the Return Event, and therefore, the Day 0 retention rate is always displayed as 100%.
Q. The Total Install count does not match the sum of the install counts of individual dates. Why is there a discrepancy?
The Total Install count may differ from the sum of the install counts of individual dates. The discrepancy occurs because different rules are applied for counting unique users.
Unique Users
The number of unique users is different from the number of events that occurred. Even though an event occurs 10 times during the analysis period, if all 10 events were performed by a single user, the number of unique users is counted as 1. Unique users are determined by the ADID (GAID, IDFA) or IDFV and randomly generated IDs when ADID is unable to be collected.
The Total row displays the number of unique users who performed an event throughout the analysis period. Even though a single user performed the Start Event several times on different dates during the analysis period, that user contributes as one unique user to the user count displayed in the Total row.
The install counts of individual dates indicate the number of unique users who performed the Start Event on the individual dates. If a single user performed the Start Event several times on different dates during the analysis period, this user contributes as one unique user to the user count displayed in each of the rows of individual dates.
Therefore, the Total Install count may differ from the sum of the install counts of individual dates. The following is an example to help your understanding.
Example) A single user performing the Start Event multiple times during the analysis period
Analysis period
2022/01/01 ~ 2022/ 01/01/02
User behavior
1. Installs app after viewing an ad in media channel A on 2022/01/01 and deletes app
2. Re-installs app after viewing another ad in media channel A on 2022/01/02
Analysis result
1. Number of users displayed in the Total row of media channel A for 2022/01/01 ~ 2022/01/02: 1
2. Number of users displayed in the row for 2022/01/01: 1
3. Number of users displayed in the row for 2022/01/02: 1
The Retention Report is synced with the app timezone that was set at the time you registered your app to Airbridge. The analysis period can be set up to 174 days.
The analysis period you can set depends on the interval options you select.
Daily: Up to 31 days
Weekly: Up to 21 weeks
Monthly: Up to 3 months (Up to 92 days)
Intervals are based on strict calendar dates.
Example)
Start Event: ‘Installs’
Return Event: ‘Any Event’
If an install event occurred at 22:00, 2022/07/01, the in-app event at 23:00, 2022/07/01 is counted as the Return Event for Day 0 retention, and the in-app event at 2:00, 2022/07/02 is counted as the Return Event for Day 1 retention.
Was this page helpful?