Users, in both cases. (I've not been privy to the data in a long time and never for Opium, but I suspect the uninstallation surveys showed it quite clearly from what was said.) The original decision to drop it was based on telemetry showing the majority of users never touching bookmarks.
> What could have been done differently in the implementation to have resulted in a different outcome.
From what I saw, the public feedback was mostly pertaining to infrequently accessed sites — which is the one case where $YourFavouriteMagicAddressBar doesn't work well. One could certainly experiment with allowing users to in some way tag or add descriptive text to URLs which showed up in the magic-address-bar, but that may well just not work sufficiently.
> The original decision to drop it was based on telemetry showing the majority of users never touching bookmarks.
Advanced users often deactivate telemetry (for privacy reasons). So you the telemetry statistics cover the normal users. For example: Microsoft made the same mistake and trusted their telemetry statistics too much. That's why they made (in retrospect) stupid and rather unpopular decisions in their recent products.
