So, thought it was time to update this thread. My "experiment" of deleting and re-creating my YouTube channels at the end of Q3/beginning of Q4 2023 has pretty much confirmed the Byzantine and capricious nature of the YouTube algorithm. It's nigh impossible to predict if or when a particular video will go "viral," or to reliably "design" a video to do so (at least, if you're a good-faith creator producing original content and not AI slop).
On my railfanning channel, Badgerland Rail Videos I have been regularly uploading a mixture of new content and re-edits (generally with improved horizon leveling, stabilization, etc) of old footage (to the extent possible, on the 5, 10, and sometimes 15-year anniversaries of when it was shot).
The channel has been plodding along for the last few months with each new upload generally garnering one to three hundred views and 8-10 "likes" before kind of petering out; maybe 300-500 views and 15-20 "likes" if I'm lucky. The channel has been gaining about 4-5 new subscribers per week on a good week, but occasionally has gone a week or even two where the subscriber count just stalls.
Then the Sunday before last, I visited a small railroad museum in Oelwein, IA. They have an old F-unit locomotive that is not fully operational (cannot move due to lacking several critical parts), but the prime mover (diesel engine) does run, and the horn can blow. I happened to get there at the right time to be in the engine compartment when they started it up.
While interesting for a railfan like myself (and presumably, the audience for that channel) to see, I didn't consider it an exceptional video. Certainly not worthy of 100 times the exposure of nearly all of my other content. However I uploaded the video last Thursday and it exploded with views, likes and comments. Over 40,000 views, 1,100 "likes" and has garnered me over 250 new subscribers in six days, vaulting me over the subscriber and watch time threshold for full monetization.
Meanwhile, over on my personal or "Everything else, mostly storms and vintage ceiling fans with a smattering of live music and cute animals" channel, back in late February this year I uploaded a compilation of what I otherwise considered "throwaway" footage from chases past where I had a minute or three of photogenic storm structure, a couple of decent lightning bolts, etc. It unexpectedly took off and vaulted up to over 55,000 views before petering out, gaining me about 200 new subscribers in the process.
On the flip side, in May I was on Arnett and got what I consider tied with Keota for my best tornado footage ever, and it kind of sticks in my craw that the footage appears to have sunk without a trace. Three months later and less than 200 views on the "main" video from that chase, and less than 100 on the timelapse compliation that I was particularly proud of. This is the kind of content I really want people to see when I happen to capture it, and for some reason in this case YouTube decided that they just...wouldn't.
This was the footage I was sure would put me over the 1,000 subscriber threshold on this channel. Not so much. It was at around 920 then, now it's been hovering between 941-947 for at least the last six weeks. Frustratingly, for every 2-3 new subscribers, it seems to always lose one. The analytics confirms that this is often from people actually unsubscribing and not closed accounts. I very rarely unsubscribe from a channel after subscribing to it. The only reasons I do are (1.) If I accidentally subscribe to a channel with the "wrong" one of my own channels or (2.) I subscribe and then eventually find the channel uploads *so much* content that it overwhelms my subscriptions feed. It's not like I upload that much nor do I upload any political or potentially controversial content (at least, in my opinion) so I can't help but wonder every time I lose a subscriber what was their reasoning for subscribing and then unsubscribing.
