Firebase Install Counts in Audience Reporting

Search…

TL;DR: The best way to measure installs with Firebase data is using the first_touch_timestamp user property.

Firebase Analytics itself does not directly report on installs but does capture the first time a user opens the app. This is a property called user_first_touch_timestamp and is set the first time a user initializes the Firebase Analytics SDK. At this point, a first_open event is fired and automatically collected.
However, the first_open event is a misnomer in Firebase. The event first_open is actually tracked each time a user installs, reinstalls, or updates the app, meaning a single unique user can have multiple first_open events. AdLibertas builds audiences with this in mind and uses the user_first_touch_timestamp property to build retention curves and calculate new user counts. The user_first_touch_timestamp is only set on the first first_open event and does not change as long as the user has the app installed. A few things to note when looking at an Audience:
A user may download the app but never have an active session. This would show as an install event in the app store’s reporting or your MMP, but as it won’t have a corresponding,first_open won’t show in Audience Reporting.
A user may download the app, and open it on a different calendar day. The timestamp of the install event and the first_open event will fall on different days.
A user may have multiple first_open events if they uninstall, reinstall, or update the app.
A user may download and fire a first_open event, but “bounce” meaning they did not fire an engagement event or stay long enough in the app to be counted as an active session. These users are not counted towards daily active user counts.
A user may have multiple first_touch_timestamps, if they have re-installed. This can create confusing results where users have activity, prior to their first_touch_timestamp. To combat this, we recommend you exclude first_touch_timestamps that are outside of your install windows:
In Audience Reporting, we include an “Installs During Period %” metric to help you understand the percentage of users whose first touch timestamp occurs within the query’s date range.
A low percentage of “Installs During Period %” means the majority of users are reinstalling the app during the query’s date range.
A high percentage of “Installs During Period %” means the majority of users are installing the app for the first time.
Please refer to Firebase’s documentation on first_open and other automatically collected events.

Firebase Best Practices: keeping Google BigQuery Costs Down

Setting User Campaigns in Firebase

Last modified 4mo ago

Copy link