Youtube Recommender System and Network Effects
Source: https://www.scientificamerican.com/article/youtubes-recommendation-algorithm-has-a-dark-side/
The article revolves around the important connection between Youtube’s recommender system and how they make revenue. The article discusses how Youtube’s revenue directly relates to the amount of time the users spend on watching targeted ads. In order to maximize the views on an advertisement, Youtube has to retain its viewers through having powerful recommender system that would keep the users to watch Youtube.
In class, we learned about network effect and the rich-get-richer phenomena. We now clearly know that we are more likely to come across content that is popular rather than content that is fairly unknown. However, Youtube doesn’t just recommend you because the video is popular. Instead, they gear towards recommending the video that you would like the most. So, how much does the network effect come into play in Youtube’s recommender system?
Youtube actually doesn’t rely much on network effects. It mostly relies on the channel that you have subscribed to as well as the genre of the video that you are watching. The most likely thing that would pop up on your youtube recommendation would be the newly uploaded video from the channel that you have subscribed to. Other recommendations that there might exist would be videos that are in a similar genre. Therefore, you can clearly see that the rich-get-richer somehow does not entirely play a role in the system of Youtube.
However, Youtube certainly relies on network effects to a certain extent. Youtube would generally gear towards the channel that you subscribed to and the genre of the video that you like to watch. However, it is more likely to recommend videos with plenty of views and responses. It is less likely to recommend videos of 0 views. Therefore, there is a complicated relationship between network effects and preference that comes in to play in Youtube’s recommender algorithm.