I don't have any insights into how Facebook screwed up here, so speculating is dicey. That said, this pattern of BigCorp doing product "B" which is supposed to the the same as product "A" which is popular and from someone else, and failing miserably at it. That is a fairly common pattern. I have seen it happen at two separate companies where I worked and in both cases it was an edict from "on high" that there be a competitor product created, but nobody on the management team (and in one case the engineering team) had ever used the competitor's product, or understood why someone would use the product. The result was a product that checked all the same boxes but without any of the execution touches that made the product successful.

I could imagine someone at Facebook said, "Hey, look how much screen time Twitch gets with our users, we should do something to make that Facebook screen time instead. Let's make a Twitch clone, how hard can it be?" And gave it to a development team that had no idea why someone would spend time on Twitch. Google was pretty infamous for doing this (how many IM systems have they built now? 8?)