Mysteries of the YouTube algorithm

The other day I saw that one of my recent Danger Zone videos had over 300 views - which is unusual as most of my videos get around 100 views.  The video wasn't anything special, just one of my games.  So I compared the analytics of that video to the next, which got a more typical 100 views.

The big difference was the number of YouTube Search impressions: 1582 versus 196, which yielded 196 and 21 view respectively.  In both cases the click-through was a little over 9%, but this was definitely one of the big reasons for the higher number of views.

But looking at the search keywords there isn't any difference.  This isn't surprising as all of my videos have basically the same text and metadata.  The search terms also weren't that specific.  (As opposed to my "new map" videos which get a lot more views because people are searching for that keyword.)  So there must have been another reason why the one video showed up in a YouTube search eight times more than the other.

Looking at the graphs for impressions over time the difference becomes obvious.  At around 20 hours the 300 view video starts getting a lot more YouTube search impressions.  The 100 view video received 37 search impressions on the second day while the 300 view videos received 317 impressions.  This looks like the "YouTube algorithm" which people talk about.

The one difference I can see is a higher click-through rate for the Browse features, i.e. people clicking on the video from their YouTube home page.  YouTube must have decided that the video was "interesting" and therefore pushed it up the search results.


Comments

Post a Comment

Popular posts from this blog

My Crosshair - why it doesn't suck

 The other night I was playing a game of Competitive mode.  While I don't queue up for Competitive, occasionally I'll accept an invite to join a lower ranked team.  My team was solidly trounced 1:15 as our opponents had several higher skilled players.  On the last round I was the last survivor as I was the only member at the other bombsite - and one of my teammates mocked my crosshair (not the first time it's happened). Unlike the vast majority of "pro" players (and wanna-be pro players), I use the Classic Dynamic crosshair (with a center dot) rather than a static crosshair.  It's also bright pink and thick.  And each of those characteristics is by choice and for a specific reason. For the longest while I was using the default crosshair - why change?  Sure I could copy some pro's settings, but I'm no pro.  But then I started really watching my games (one of the advantages of recording everything) and studying what I was doing.  What I realize...
Read more

Buying a new PC

I started playing Danger Zone on a Mac, specifically a late 2013 27" iMac with a 3.5GHz Intel Core i7-4771 (Haswell), an Nvidia GeForce GTX 780M and a beautiful 2560x1400 27" screen, or as I referred to it: a 6 year old gaming laptop with a really nice screen.  Unfortunately, it really didn't have the horsepower to drive all 3.6 million pixels - even on lowest quality settings.  But it's what I had, it worked, and so I used it - although I received many comments about how low my fps was. Then AMD announced their Zen 2 processor, which seemed to be like the original Athlon processors - both better and cheaper than mainstream Intel processors.  So I started looking, planning and dreaming.  Ideally I wanted a computer which could play Danger Zone at 1440p @ 144Hz at top quality settings (aka "full pretty") - but I couldn't get good, hard information about the minimum requirements for that goal.  So I fell back on my "thrifty" nature and instead tr...
Read more