@jon This kind of optimization work is interesting, in part because of what you can't see. Algorithmic timelines drove virality. They made the site more useful for the people who didn't contribute to conversations, because they got a feed of today's cool shit, and broke it for people who did. So, you selected for cool shit and pushed people towards promotion and takes over conversation. I know, because I watched what it did to me, and it's why I mostly stopped posting, despite a lot of followers.

I'd really strongly suggest spending some serious time thinking about how you knew, from the data you had, the things you think you knew. Yes, if your only goal was MAU and ad impressions, you were doing the right thing. However, if that's your only goal, why not just go work for some place that's just a money farm? I'm not saying mastodon is better — I'm still rather skeptical — but it's definitely doing some things that are less bad. The public square nature of Twitter didn't need the algorithmic boost — it was doing just fine with old manual RT — and for a lot of the rest, you could have done a lot more with curation instead of killing conversation.

I'm sure these conversations have been done to death internally, and you can and do claim the authority of the data, but I think the path-dependency that generated that data blinded you to the other options.

I have no idea who's on what app anymore but weekends are still when we help Afghanistan so here is a campaign to help people in Herat Province:
https://aseelapp.com/en_us/support-afghan-families-of-herat-provience-in-winter-season.html

Attached is a map of Afghanistan. In green are provinces my Twitter followers have closed out campaigns for previously, Herat is the large western province in red.
#SaveAfghanistan #afghanistan #HeratProvince #AseelApp #ChaosGoblinArmy