The number of downloads in firebase is 30x different from the google console

Question

I have an app in Google Play and firebase analytics on it. I'm trying to track ad sources. And I can't understand where is the right count of downloads, because Google Play Console shows 150 installations, Firebase shows 3000 "first_open" events. I was confused and added appflyer analytics, and appflyer shows like two times less than firebase. The app's Google Play page shows 100+ installations.

2 weeks have passed since the beginning of tracking, so that the data in the Google Play Console was exactly updated.

Syllepsis

Answer 1

Is it possible your firebase analytics is tracking your development activity? And every time you manually install from your IDE it counts as a first open. You could check this by looking at your device distribution in firebase. If they are all exactly the same as your development device then that's your answer.

Basildon

No, this happens when we start traffic. So i cant trigger "first_open" 700 times for 2 hours. We made 2 waves of traffic and in both cases data in firebase and google console was completely different. The only answer what i have for now its that google play consosle count downloads 1 per account when firebase 1 per device. So probabbly traffic is problem, maybe its a China "click farm" where 1k devices registered on 1 google play account or smth like that. I believe more in console than firebase. – Syllepsis

I just can not believe that google play console can have such a big issue. We also created custom tracking to send referral links to track traffic sources, and there is data very similar to Google Play console. Even unistallations in firebase like 1.5k for 3k installation. When in google play console 150 installation and 90 uninstallation. – Syllepsis

Another possibility is sideloaders. Google Play counts installs from Play, firebase might count installs from all sources, including sideloading. So maybe P2P sharing, downloading from the web etc. is boosting the numbers. Is your app one that you would expect to be popular a lot in countries or for people who do a lot of sideloading? – Basildon

PS, you've made me super curious. If you don't mind, could you send me an email with details of your app. My first initial and last name @google.com ? – Basildon

No, the application is online in one copy and it is on google play. – Syllepsis

Answer 2

If you dig into the raw firebase analytics events dataset through BigQuery and DataStudio, you'll be able to see a field named "app_info.install_source" in each first_open event.

You can then see that your first_open events can be split into 3 main groups based on their "app_info.install_source" field, where the values can be 1 of the following:

1. manual_install
2. com.android.vending (which i assume is via Google Play Store)
3. iTunes

This may help you to get a more accurate picture. If you have many that are "manual_install" type, then it could be:

• Click farm
• Android tests on virtual devices during pre-launch crash testing
• During development you or your team could be triggering it somehow
• Side load or direct downloads of the APK
• Other reasons which you may like to share with us if you find out so we can learn what are the other possibilities

Madrigal

This already helps me a lot. But for me there are still 40 users with a Nexus 5X device, where each has a different locale setting. They appear with every upload to the play store. There must be some automated language test, but i don't know where they come from. I already filter everything from the pre launch reports (everything with context "firebase.test.lab". – Hayashi

Answer 3

Here is the description of first_open event of firebase.

the first time a user launches an app after installing or re-installing it.

This event is not triggered when a user downloads the app onto a device, but instead when he or she first uses it. To see raw download numbers, look in Google Play Developer Console or in iTunesConnect.

Additionally, Firebase collects first_open events even when your users open the app after updating new versions. So for example, if your app has been updated 10 times, a user still uses your app in this period, then you will have 11 first_open events in your Firebase console (sometimes it can be more than that, we don't know if Firebase still has bugs or not). I think Google's noticed this issue recently but I don't know exactly when they will fix it.

Hope this helps!

Analytic

the first_open event is triggered on re-installs or installs, not on updates. it gets triggered on an update only if the previous version did not have Firebase. – Melanymelaphyre

Answer 4

You could also enable first_open as conversion-event inside the firebase-console, and inside the conversion report drill-down into source (at least at the time of this writing). Alternatively, try the BigQuery-code mentioned in this thread: Google Play New Users vs Firebase first_open / New Users Comparison

Millsap

Answer 5

Have you tried subtracting the number of app remove events from the first open events? That number should get you closer.

From my understanding, play console only shows you the net of installs vs remove events, and firebase shows you both numbers first open events and app remove events.

Undulant