Given the data available, Firefox is a better browser with Pocket than without, on average, for Firefox users. That's justification for bundling it. Perhaps more data will be collected and other addons might be considered later, but again, the bar gets higher and higher since the total number has to remain small.
I think that the wording here is important: Pocket is not an add-on, at least in the sense that I understand it: when I go to Tools > Add-ons, nothing there allows me to disable or remove it. (I can't find it now, but I thought that one used to be able to manage search providers in a similar window.) Add-ons are by their nature trivial to disable or remove, and a solution involving going into about:config, where the user is explicitly warned of the dangers—in language far more frightening to a non-techie than any opt-in could be—is not trivial at all.
You should let users have all the value from new useful features, instead of just keeping the balance barely positive after bundling some bad stuff.
Firefox users like to download videos. A lot. 15 of the top 40 add-ons downloaded this week are video downloaders
So they certainly know what users like... and then ignore that data.
https://mail.mozilla.org/pipermail/firefox-dev/2015-June/002...
(Also, not to be nit-picky but because I think it's an important distinction, I don't think the investigation here "proved its benefit", but rather showed only that those users tested liked it. I suspect every designer would agree that bundling everything a user wants leads, almost paradoxically, to an unuseable mess.)