YouTube’s algorithm is suggesting longer and more popular videos the more users watch, according to a study from the Pew Research Center.
In the study, researchers underwent what they called “random walks” through the platform, beginning with one randomized video from amongst 14,000 channels, which had more than 250,000 subscribers each.
Researchers would then choose from one of the videos recommended by YouTube in the original video’s sidebar, until they’d watched five videos in total using the same method.
Their study found that 64% of videos recommended to researchers had one million views or more, with only 5% having less than 50,000 views. Additionally, 28% of the videos recommended to researchers were suggested multiple times, with videos increasing in length over time.
The recommended videos increased from a starting time of nine minutes and 31 seconds to an ending record of 14 minutes and 50 seconds. The videos likewise increased in popularity, with beginning videos having around eight million views and ending videos hitting over 40 million views.
The study concluded that YouTube is recommending progressively longer and more popular videos to users the more they use the site’s algorithm.
This could raise concerns for smaller channels, who may not see as many views thanks to this trend. Larger channels have also feared that they could be affected, should they stop frequently uploading.