A/B testing should generally be done within an app, though. If your app already calls home to a server, you can collect usage stats and send feature flags for A

A/B testing should generally be done within an app, though. If your app already calls home to a server, you can collect usage stats and send feature flags for A/B testing. If it doesn't, you probably have to run your tests on in-person groups. The parent here proposes releasing multiple versions of an app and seeing which ones get downloads, which seems much harder and riskier - you end up splitting downloads across multiple apps, and then have to deprecate or maintain the ones that don't work out well.