I took Firebase analytics for a ride — Here’s the conclusion

Update: I’ve added a new review after using firebase on production. You can find the complete analysis here — https://medium.com/@vernekard/im-disappointed-with-firebase-analytics-9ee136c232ec#.qqkn6obe1

Google announced Firebase — a new analytics and PaaS offering that would make your life a bit more easier at IO yesterday. Read on to know what’s new, what’s good & what’s plain ok in the new tool.

To a lot of mobile developers and product managers, Google Analytics for the mobile seems like an after-thought from the web era — a system designed to fit-in with the new needs of the mobile era, rather than tackle them head-on. If you’ve worked on mobile and web applications, you realise that user behaviour is remarkably different on the 2 platforms.

To make your job easier, I decided to take it for a ride.

Getting started is incredibly easy if you’re using the latest version of Android Studio. Just add the downloaded json file, create a project and you’re good to go. Interesting they’ve broken down the different services into their own libraries, so you don’t need to worry about adding extra kbs to your app.

Here’s the complete documentation: https://firebase.google.com/docs/android/setup#add_firebase_to_your_app

The dashboard is pretty neat too. I was most interested in the Analytics & Deep-linking solutions. Here’s the initial dashboard.

Currently there are 3 main areas of focus for an app developer:

1. Attribution: Know where your users came from — typically one uses platforms like Apsalar, Appsflyer, MAT which are partners with Ad platforms, Google & FB.
2. Analytics: Understand what your users do inside your app. Combining this with attribution data tells you which users are the most valuable, giving you a guide to improve your app experience and your marketing efforts.
3. Engagement: Keep the user engaged by pushes/emails/sms etc.

While one can derive great value from merging the 3 data sets, typically for an advanced app, no product meets these three needs sufficiently. Hence one could merge the data sets using BigQuery/Redshift & derive value from the processed data. Google’s foray into this space is interesting for the same reason.

Analytics:

Google Analytics is a fabulous tool. One can get any level of depth into the data if configured correctly. However most developers/PM’s need a certain level of experience to use it correctly. Firebase has recognised this and provided a more varied set of events which are captured by default.

Get Dinesh’s stories in your inbox

Here are the features which were most interesting to me:

1. App open — when the user opens the app for the first time.
2. user_engagement
3. App_clear_data — you can know when the user clears their app data
4. app_remove — UNINSTALL analysis!! Now know which users removed the app, reach out to them for feedback and more.
5. notification_foreground — when a user receives a notification when the app is in the foreground! And a lot more push related data I elaborate on later.

Automatically collected events - Firebase Help

Firebase collects these events by default. Automatically collected events are triggered by basic interactions with your…

Understanding push notifications:

I wrote about how to optimise your push notifications since it’s the best channel for optimising your app retention. I also talked about how one should go beyond immediate conversion metrics & try to dive deeper into the data. Interestingly Google has come up with just the metrics you’re interested in.

Audiences:

Google understands our pain. App growth on the playstore without advanced data segmentation & analysis is impossible. The new segmentation & funnels seem like a step in the right direction. Please note how they have taken ‘custom dimensions’ and re-named it as ‘user properties’ to help app developers tag their user levels, network conditions etc better.

Segment!

I know what you’re saying, our tools already do this. Yes they do, but this level of depth with push & analytics from a free tool is just mind blowing.

Cohort Analysis:

While you can set custom duration for your cohorts, you still can’t cohort on a custom segment. That’s a bit disappointing for me, as we’re still stuck with retention analysis of acquired users only. Retention across channels, retention across versions, anyone?

Deferred Deep-linking:

Providing a consistent deep-linked & deferred deep-linked has been a pain with the lack of a consistent platform.

Experimenting with app links: It’s pretty simple to create links for your promotional offers.

Notifications:

Configuring notifications also was easy. If you want to send messages to a segment of users while the app is in the background, there’s no need to configure the client side handling of the message.

And Send!

Overall it seems that Google has payed attention to the needs of app developers and Firebase is a good mixture of the right ingredients. Future updates should be very exciting. Liked the article, feel free to tell me in the comments or write to me at vernekardinesh@gmail.com

TD:LR-

1. Analytics offering powerful — useful default tracking like first open, notification open, uninstall added
2. Notification: Good to start with, options for key value pairs added
3. Deep linking: Very easy to setup, best part of this whole offering

If you liked this article you’ll like Building for India — a billion opportunities https://growthbug.com/building-for-india-a-billion-opportunities-becfa0572183

If you found the article useful, please feel free to hit the green ‘RECOMMEND’ button.

I would love it if you give it a try. You can follow me on Twitter or Linkedin