Discovery on youtube is pretty poor, imo.<p>I suppose it's a lot better than FB (we decide what you see), and the navigation model (search & browse a database) is what I want. But overall, it's worse than most porn sites.<p>These things have feedback loops and 3rd order effects. If discovery worked differently, it's not just what people see that gets affected. Pretty soon, what gets created is affected as the market reacts.<p>This problem is kind of reminiscent of Google SEO/spam problems, but youtube seems a lot worse than search ever was. SEO-spam, clickbait & stuff that hovers just over the spam threshold.. these are pretty dominant.<p>"Monetization" was another example of market poorly handled. It just did not reward the kind of thing youtube should want to reward.
That site's ads were . . . surprisingly tasteful. Got around my adblocker because they were hosted by the advertisers, not an ad company, and, while I wasn't a fan of the Samsung auto-playing video, the Mailchip ad was novel. It felt like reading a classier newspaper, which I can't say I dislike, personally. Plain-CSS ads are the way to go.
The far bigger problem from my POV which is not described in the article is manipulation of YouTube trends. This is to some extent related to buying views but needs a special kind of attention.<p>A few weeks back I noticed weird videos in the German YouTube trends, things like "cheap iPhone cases buy now" with channels that had zero subs. A day later there were 3 of these videos in the trends, the next day 6, then 50% of the trends were full of this. A guy with 0 subs even made a video explaining that he does this kind of manipulation for a living...and unsurprisingly the video made it into the trends the same day.<p>Simply buying views has very little effect (except on maybe advertisers that pay you) but gaining massive exposure through being featured in the trending videos is big.
I'm curious about how the viewbotting is actually carried out and whether this impacts inline video ads on Youtube. I experimented with these video ads and did some very narrow targeting for our audience (searches related to dyslexia or ADHD). I know that a sizable percentage of people with dyslexia or ADHD love our tools (and our video ad contained a demo), but the click-through rate was next to nothing.<p>I wondered at the time if this was an issue with how Youtube does their targeting, or if it was something else. Seeing how popular viewbotting is, I now wonder if the bots end up getting categorized as relevant to different types of ads and are subsequently shown them. Of course, this would depress the click-through rate and the overall efficacy of Youtube video advertising.<p>Would be interested in others' thoughts.
I'm honestly not sure I can trust anything this author says.<p>In this article he links to another he's written [0] which posits that Professor Jordan Peterson is recruiting depressed men for the alt-right. It's full of nothing but false emotionally-driven insinuations, and is especially ridiculous considering Peterson's very explicit stance against extremist ideologies on either side, including the alt-right.<p>[0]: <a href="https://theoutline.com/post/3537/alt-right-recruiters-have-infiltrated-the-online-depression-community" rel="nofollow">https://theoutline.com/post/3537/alt-right-recruiters-have-i...</a>
Social proof is so important, even as I am aware of it I still fall for it, for example Github stars and contributors.
I kinda feel special when I am watching a youtube video of less then 100 views though.
Past youtubers were able to buy views in order to generate social proof. Any new youtuber won't be able to take this same shortcut. Maybe youtube should consider hiding the viewcount of videos under a certain amount, say <5000, in order to give people a shot again.
The article starts with putting Jordan Peterson in the same cohort as Alex Jones and crisis actors conspiracy peddlers. I’m not a Peterson fan, but those kind of comparisons strike me as extremely misinformed and dishonest, and undermines the whole credibility of the article by a lot