Retention

Retention report provides an N-day retention rate for users who have performed a specific event in-service and came back to the service on a specific date and performed a designated event.

From an attribution perspective, you can analyze how loyal users flew in from each ad network channel or ad campaign, and you can optimize the campaign based on the performance of retention reports or use them to settle paid advertisements. 

 

image__5_.png

 

Configuring a Report

mceclip0.png

Event

Choose 'Start Event' and 'Return Event' to check the retention rate of users who performed those events.

Start Event

This is also known as a target event. Airbridge uses Start event to measure Return event retention by date.

  • Install: Number of installs, measured by the first-time app opens. You may select this event to analyze the retention for the User Acquisition campaign.
  • Deeplink Open: Number of App opens through DeepLink(scheme or universal). You may select this event to analyze the retention for the Re-engagement campaign.

Return Event

Return Event is used to determine the retained users among the users who have performed Start Event.

  • Any Event: Any app event performed by the users (at a given period).

You can select each event based on what you'd like to analyze. For instance, to analyze the retention of your UA campaign, select 'Install' and 'Any Event' as Start Event and Return Event respectively.

 

Measurement Option

Measurement Option enables you to analyze retention by Return Event from the User Journey perspective. Retention report measures by Airbridge Device ID(General) by default, but you can also change it to measure by Start Event(Confined).

💡  What is a User Journey?

User Journey refers to a path from which a user installs an app after clicking or viewing an ad, occurs in-app events, and exits the app. The Retention report computes all Return Events existing within a single user journey. 

General: Airbridge Device ID

blobid1.png

  • Airbridge distinguishes the user journey based on the Start event that occurred by the user with the same Airbridge Device ID within the analysis period. 
  • If the user performs another Start event, the new user journey is created, and the existing user journey continues. 
  • Therefore, if the user has the same Airbridge Device ID, the retention of the Return event computes including all occurred Start events even if each Return event was attributed to a different channel. Some data may be duplicated. 

Confined: Start Event

blobid2.png

  • Airbridge distinguishes the User Journey based on the Start event that occurred by the user with the same Airbridge Device ID within the analysis period. 
  • If the user performs another Start event, the new user journey is created, but the existing user journey ceases. 

Use Case

User Timeline

blobid3.png

Mar 1 (Day 0): App Install (Kakao / Start Event)

Mar 2 (Day 1): App Open (Kakao / Return Event)

                 Delete the app

Mar 3 (Day 2): App Re-install (Moloco / Start Event)

Mar 5 (Day 4): App Open (Unattributed / Return Event) 

 

Report Configuration

Analysis Period: 11/1 ~ 11/5
Start Event: Install
Return Event: Total App Open
Measurement Option: (stated as below)
Groupby: Channel

Based on the above report configuration, the retention results according to the measurement option are as follows.

 

Measurement Option: General (by Airbridge Device ID)

The App Open(unattributed / Return event) on Mar 5th was occurred by the same Airbridge Device ID, and thus it counted to both user journeys.

  • App Install event (Kakao / Start event) on Mar 1st is counted for Day 4 retention.
  • App Install event (Moloco / Start event) on Mar 3rd is counted for Day 2 retention.

blobid1.png

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          

 

Measurement Option: Confined (by Start Event)

Another App install event on Mar 3rd creates the new user journey, and the previous user journey ceases the journey. 

App open event (unattributed / Return event) on Mar 5th is counted only for the user journey to which the event belongs.

  • App Install event (Kakao / Start event) on Mar 1st is not counted for the second App Open event on Mar 3rd. 
  • App Re-install event (Moloco / Start event) on Mar 3rd  is counted for Day 2 retention.

blobid2.png

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          

 

 

Group By & Filter

Group Bys and Filters settings allow you to view the retention breakdown using a variety of dimensions such as channel, OS, and more. Note that you can select up to four Group By dimensions at once.

Selected Group Bys and Filters are displayed in the upper table where you can also easily add or delete dimensions.

💡 Group By and Filters you select will only apply to the Start Event.

💡 Sub Publisher and Sub Sub Publisher can’t be exported or downloaded due to the excessive data.

 

Cohort

A cohort refers to the user group satisfying specific criteria. You can define the cohort that you want to analyze the retention.

For example, the cohort allows you to analyze the retention of the 'Any Event' after user engagement through the app Installs over the past 30 days for the users who have performed a purchase event at least once in the past 30 days.

With various criteria such as an event, event property, event count, or date range, you can set your cohort as granular as you want.

 

Event

  • Only 'App Event' is available
  • Any Event refers to all occurred events within the selected period.
  • An event can be selected either 'performed' or 'did not perform'

 

Event Property

  • You can add the event property of the selected event.
    • e.g.) Device Type, Product Price, Is First, etc
  • Airbridge provides comparison operators(>, ≧, <, ≦) for event property with value.
  • Event Property Operator
    • e.g.) Event Action, Device Type, Is First Event per Device ID
      blobid0.png
    • e.g.) Total Revenue, Product Price, CTIT
      blobid1.png

Event Count

  • Enter the number of event occurrence
  • Can be entered more than or equals to 0
  • Event Count Operator
    blobid2.png
  • equals(=) 0 time, less than or equals(≦) 0 time, less than(<) 1 time

    The above conditions are all regarded as 'the number of event occurrence = 0'.

Date Range

  • You can select up to the past 90 days or a period within 90 days based on the audience creation date.
  • Date Range Operator
    • During last: Available from D-N from today's date
    • Between: Start date and end date
    • Since: From selected date to today

 

Retention Report Table

Click an expand button on the leftmost of the table to view retention data either by group by and by day.

blobid3.png

Also, you can directly click the column in the table to delete the column or add filters.

 blobid4.pngblobid5.png

 

Report Copy & Download

You can easily copy or download the data from the reporting table to paste excel or spreadsheet, and download via CSV file or Google Sheets right away.

  • Downloading a report may take a while depending on the amount of data

blobid2.png

 

Frequently Asked Questions

1. The number of installs in the Retention report and Actuals report is different.

As the number of installs in the Retention report may exclude 'duplicate installation' within the analysis period, it may show a lower number than what is shown in the Actuals report.

If an identical 'ADID' installed an app twice or more on the same day, Airbridge takes account of the latest app install into the retention calculation. On the other hand, if that user installed an app multiple times on different days, then Airbridge computes the retention for each date.  

 

E.g.) Option: unique user (based on the ADID), Analysis period: Feb 3rd - Feb 10th

scenario 1)If App installs occurred on Feb 4th at 1 PM, 8 PM respectively, Airbridge takes account of the last app install that occurred at 8 PM only when computing the retention. 

scenario 2) If App installs occurred on Feb 4th and Feb 7th respectively, Airbridge counts the app installs that occurred on each date into the retention.

 

2. What's the definition of 'Unique User'?

Unique User refers to the user who has performed Start Event in a certain period of time.

e.g.) Groupby: Channel, OS Name / Interval: Daily

Analysis Interpretation: Users attributed to unattributed channel from the Android OS in 2021-05-03 equal to 5,168, and the retained users who performed Open Event in the next day equal to 1,687.

blobid7.png

 

3. What's the maximum period I can select for the retention report?

Retention Report uses the app time zone you set when you registered your app in Airbridge Dashboard. The date is available from 1 day ago to the past 1 year. The maximum date range differs per interval as below.

  • Daily: 31 days
  • Weekly: 12 weeks
  • Monthly: 3 months (max. 92 days)

 

4. How does 'Interval' time work for the Start Event and the Return event? As 24-hour windows, or as strict calendar dates?

Airbridge computes intervals by strict calendar dates. For instance, if a user performs their start event on July 1st at 10:00 PM, and performs their return event on July 1st at 11:00 PM and July 2nd at 2:00 AM respectively, Airbridge counts the user performed their Return Event on Day 0 and Day 1 respectively.

 

 

 

 

Was this article helpful?
0 out of 0 found this helpful