Super interesting article. With a system like this you can also get great metrics for product health / tech debt. If you were also looking at key growth metrics for the apps, you'd be able to see which releases improved them, and give credit (in the cases where it makes sense).

Metrics related, here's a book review of Moneyball I just wrote: https://dtrejo.com/moneyball-book-review-and-measuring-reven...

This is an interesting article. Native Apps deployment constraints seem so different than a typical web continuous deployment process.

As an aside, here's a look at Etsy's Continuous deployment process: https://www.slideshare.net/mrtazz/development-deployment-and...

If you like this, Buzzfeed had something similar recently: https://tech.buzzfeed.com/deploy-with-haste-the-story-of-rig...

what is missing in all of this, i think, is the ability to do a/b testing. It'd be nice if app stores provided the ability to repush existing version and new version to a small fraction of users - this way you could compare in a more sane way.

Google Play Store has a rollout feature that makes this possible, but it's not possible on the App Store AFAIK.

i am not sure if that feature is going to have uniform traffic for old and new releases.