Addendum (Patreon)
Content
A lot of people responded to my ramble video, re: doing more niche content like horror movies or disney parks, saying to just talk about what I like or to put those in playlists. I would do that except:
My hesitance with subseries on my main channel is based on the sub box algorithm. This is the same reason I worry about staying (relatively) topical. Youtube's programming is intentionally mysterious but the general finding is that if you make a lot of videos that your subscribers don't click on, youtube stops recommending your new videos to them, because the sub box has limited space and the algorithm decides they've lost interest in your channel and pushes others instead. When you hear youtubers saying things like, "my subscribers don't even get notified when I make a new video!" that's what's happening there. It's a common symptom of channels that try to flood let's-play videos in between their regular programming.
I won't name names but some really prominent channels struggle with this, gathering millions of subs, but only getting substantial views on the videos that manage to go independently viral - only because they make other "filler" videos which their subs have failed to want to click. On one hand it's good of youtube to discourage low-appeal content, but it's kind of a one-strike system which can be stressful. And doesn't really encourage risk-taking or variety within a channel. By this system it actually makes the most sense to make a million small channels with hyper-specific content formulas to minimize risk - which is obviously its own burden because then you have to try to gather subscribers independently across all of these.
So, while I don't care on an individual level if some videos have slumps in view counts, I do worry about falling out of favor with the youtube algorithm which would hurt the channel as a whole. That's also why I don't do videos announcing con appearances or collabs or whatever, and just use social media for that. People are more likely to read the title of an announcement video but not click. As a content creator it doesn't matter that the quick 2-minute announcement video isn't getting views, but then when you make your next video, 50% or more of your regular viewers might not even get it on their radar. Then, you're relying on those subs to remember to manually check your channel or potentially lose their viewership forever.
I've been lucky in that the algorithm's really favored me so far. The reason the Troll book review suddenly jumped up so many views in the 48 hours is that it passed whatever arbitrary view threshold it needed to get put in the general recommendations sidebar. So, I still obviously make wild card videos on unappealing topics like the Tomorrowland movie which nobody saw, but when I decide on a topic I also try to think about how much I'm rocking the boat.
Anyway the "science" of youtube is really interesting to me and generally discussed amongst creators so I thought you guys might find it interesting as well.