This finding seems confounded by biased populations: "engagement (number of page views per unique IP) is half or less of the old desktop or third party app". I

This finding seems confounded by biased populations: "engagement (number of page views per unique IP) is half or less of the old desktop or third party app". I would expect that only the most engaged users would be using old desktop or third party app, since those are much harder to find and require much more effort (such as digging through settings, or investigating other app options). The less engaged, casual users would be using the default options (new web or default mobile app) because they're much easier to find. Therefore I propose the causation goes the opposite direction of what you're suggesting: it's not that old web/3rd party app cause users to be more engaged, but rather that being highly engaged causes users to use old web/3rd party app.