Have you ever wondered how Facebook presents a customized newsfeed to you? It seems like the platform is good at making a wild guess on our interests and likes. Recently, Facebook explains how the newsfeed algorithm works. Facebook’s tech team explains how they use machine learning to personalized newsfeed for each of their more than 2 billion users. In an animated video, here’s what they say.
When a user interacts with something on Facebook, it lets them know what’s important to the user. These reactions include comments, likes, shares, and tags. From there, Facebook predicts the causes that may be meaningful to the user or what the user might want to see in the newsfeed. These are just some of the many signals that AI uses to personalized it. They may also consider the following:
- Is it a photo or is it a video?
- Who’s tagged in a post?
- How recent is it?
Facebook newsfeed algorithm then calculates the type of content the user wants to see. This is based on the individual score of all the actions made and other signals. Such a motion is also applied to content that users generally want to avoid seeing like click baits or misinformation. The remaining list is then narrowed down to the most relevant posts. These are based on the final score for each post and where real personalization happened. As a finishing touch, Facebook applies processes to achieve a variety of content such as photos, videos, etc.
Facebook explains how the newsfeed algorithm works on 27 January 2021.
Implications for Marketers:
Facebook AI is indeed an accurate way to personalized newsfeeds. For marketers, there are no cheats and tricks to alter the algorithm. What brands need to do is to post content that delivers value to their audience. As such, their content will rank as one of the most reached on Facebook.