2/4/19: Choosing the best way to collect app usage data within Ellie Calendar
For this weekly update I’ll unpack a bit about how Ellie Calendar is going to handle measuring app usage going forward. First it starts with admitting that I can’t build a robust analytics platform by myself because I’m not that great of an engineer and there is only one of me.
Also, Apple’s App Analytics platform is borderline useless. It can’t measure in-app events, see which screens are viewed the most and measure performance. I need this information in the near future to keep learning from the people using Ellie every day.
So I tried out a few analytics platforms such as Google Firebase, Microsoft App Center, AWS Pinpoint and Facebook Analytics. Below is my quick summary of how I chose the best platform for my needs.
Spoiler: I went with Microsoft App Center because it fit my requirements. They are all great, but read below for my analysis of each platform. Next week I’ll go over how I integrate analytics in a privacy focused way.
Google Firebase
Firebase is by far the more robust platform, has really great integration with iOS and has super detailed crash analytics. Firebase analytics is built off of the old analytics platform Crashlytics from Twitter’s old Fabric suite, so it’s had some time to mature.
My only issue was that I had to install Google’s database management, push notifications, login support and more into Ellie which made the app size double. They offer Firebase in an all of nothing model which wasn’t really appealing to me. Also, I didn’t like how it shared a common user agreement to the rest of Google for privacy reasons.
Microsoft App Center
Honestly, App Center doesn’t have a ton of features but the positive is that it’s extremely lean. It meets all the check boxes I needed and allows for flexibility in the future. The magic with App Center is that the rest of Azure is behind it. I can query my app data and create dashboards with Azure Log Analytics. It’s a great dumb pipe into the actual useful parts of Azure’s ecosystem and I liked the strict privacy policy.
AWS Pinpoint
I really wanted to love AWS Pinpoint because I really love AWS for pretty much anything in the cloud. The package required to install into Ellie was fairly lean and it was similar to Azure in that I could export all of the data into storage and manipulate it there. My issue was more around how to query the data. It could be my lack of knowledge or that I couldn’t find clear documentation but I had no clue what to do with the data after it got into S3 or DocumentDB.
Facebook Analytics
Fairly robust software but the real perks of this come from being integrated with Facebook, which Ellie doesn’t currently do so it was kind of a none starter for me. Similar privacy concerns to Google’s terms of use. I also couldn’t find enough examples of other companies using Facebook analytics exclusively.
---
Weekly Updates: This is part of my weekly updates on making a modern calendar called Ellie. It might include super technical stuff, product decisions or design challenges.
Cloud Engineer w/ 18 certs: CISSP | GCIH | CCNA | CASP+| CySA+| Server+| Cloud+| Linux+| Security+| Network+| A+
6 年Tried it. Didn’t work. Why go back to a system we know has failed in the past?
??? Cyber & Physical Defense Engineering @ Microsoft Security | Keeping people safe by building transformative & creative modern solutions!
6 年How does your decision impact a business end user's need for ETL operations into data analytics tools like Tableau or PowerBI? Are you building dashboards directly within the Analytics Platform and would those be shareable with a business stakeholder or analyst?