In my last post I wrote about testing this ever-evolving idea I’ve been writing about here by actually running a few people through a weeklong Twitter challenge. I’m running all of it in a Slack workspace, which is all running smoothly so far with four participants of various stripes. Naturally, promising actual “audience-building” is a tall order, so we focus instead on things we can actually control: I like to think of it as helping people build a habit of regular community engagement. It is meant for creator types who need an audience but would rather not have to think about it.
The heart of the project, if it turns out to be viable, is a personalized recommendation system that gives you a daily goal or challenge, along with recommending other users and conversations that a user might be into. I’m doing this manually right now by looking at each profile in advance and trying to get a sense of what each one is about as quickly as I can. I look at who they’ve recently followed, what they’ve recently responded to, etc., and adjust my recommendations accordingly. For example, if you already tweet regularly but don’t reply much, I’ll handpick a few ongoing conversations and make you respond to one or another. Otherwise I stick to a default plan.
I’m trying to recommend profiles that have been active recently, are engaging with other users, and especially those who ask questions in public or invite feedback. I’m also trying to get a good mix of profiles that have roughly a similar follower count, and some larger accounts as well who engage with smaller accounts (this is not all that common). To make things more interesting, I also include a little randomness. Obviously, as a human, I can’t do all of the above very efficiently, but there might be something in a system that can respond to what you’re doing and make very good suggestions.