Consequences of YouTube's algorithm

As a follow-up to the previous post, consider this video by science educator Derek Muller, recently posted to his YouTube channel Veritasium:

In the video, Muller explains why he thinks a previous video, about the use of shade balls in a Los Angeles-area reservoir, went viral. Basically, he explains, when deciding which videos to promote, YouTube’s algorithm seems to be attempting to maximize two outcomes:

  • Watch time: How long the user will spend watching the video
  • Click-through rate: How likely the user will click that particular thumbnail when it’s displayed

As a result, Muller says that going forward, he will:

  1. “Keep making high-quality videos”
  2. “Choose topics that are more clickable”
  3. “Use clickbaity titles and thumbnails”

Because, frankly, if I’m going to work for days and weeks and months on a video, I would like that video to get, say, 10 million views instead of half-a-million.

Now, my heart sunk a little when I first watched this video, because I have very much enjoyed Muller’s previous videos, and I am not a big fan of clickbait. In the past, some of his topics haven’t been as “clickable” but were very interesting (to me) nonetheless. But I absolutely cannot fault him, either, since part of his livelihood depends on people continuing to watch his videos.1

On a related note, recently on YouTube, I have noticed a trend to longer videos, a trend driven by the algorithm’s favoring of videos with longer watch time. I find this a bit unfortunate, because short YouTube videos are great for taking a short break from work, and because longer videos require a larger investment of time. I’m reluctant to watch a video that’s over 10 minutes, unless the video comes highly recommended or the creator has made content I’ve enjoyed before.

  1. For that matter, I absolutely don’t fault news outlets for posting clickbait, either, given the changes to the media landscape wrought by the Internet in general and social media in particular. [return]