The statement below indicates that YouTube is not optimizing for watch time and is favoring higher click-through rate (CTR) content. In other words, the YouTube video recommendations just cares about the clicks and is not maximizing watch time. The implications are clear. Enticing thumbnails and titles are the keys to getting YouTube to promote your videos. High quality videos following the thumbnail are not always necessary.
Todd of the YouTube “algorithm” search and discovery team said the following on the Creator Insider channel video:
For example, with watch time, we have come across some changes in the past that we have considered. Like we had this one change recently that significantly increased watch time across YouTube. Actually, if we had rolled out this change, then people would have watched millions of more hours every day on YouTube. So, you might think, “Well, obviously, that is one of the good ones. Let’s roll that one out!”
For the dozens of metrics that we looked at for each change, we found that there were some interesting things that were happening that were not the same across all viewers. What was happening is, one that we were recommending longer videos on average. So, you might see more videos that are an hour long or more, but our hit rate, in terms of like how often a viewer would find a video that they would want to watch was lower. So, you would be less likely to find a video, but when you did find one that you liked, you would watch for longer. So overall watch time was up.
But what was ultimately happening when we dug in, was we were increasing engagement a small heavy segment of the audience and actually reducing engagement from a casual user segment. Because we don’t just look at the total volume of engagement, we also break it out and see how it is affecting different groups of viewers and we care about growing the appeal of YouTube to more and more users over time, we felt this was not representative of the type of change we wanted to make because it was making the experience worse for a significant segment of viewers that we cared about. And so rather than make them pay the price for another segment, we said let’s go back and see if we can find another way to get the improvements here without hurting it for another segment.