Many criteria influence the visibility of your publication in the Facebook news feed. Thanks to machine learning rankings, the most meaningful content for users is selected in just a few seconds when they open the application.
Quick content analysis
Each Facebook user could view more than a thousand posts per day. To sort through and choose the most relevant posts, the social network proceeds in several steps.
First, Facebook performs an initial inventory to gather posts from a friend, group or page that has been published since a user last logged on. The algorithm also takes into account the unread posts from the last sessions and those that have been read but have since led to interactions by friends and family. Then, machine learning models make predictions about which publications to highlight. The list of content to be displayed is then reduced to about 500 posts considered the most relevant for the user.
Then, Facebook algorithm defines the main rating. A score is calculated independently for each publication, which will help define the ranking. And finally, Facebook makes sure to mix the content that appears (photos, videos, articles, etc.).
All these steps of news feed ranking occur at the moment the user opens the Facebook application, in just a few seconds.
The criteria used to classify the contents
The social network relies on a user’s interactions to determine a ranking of publications in his or her news feed. It is based on several criteria, such as the characteristics of a publication. The algorithm will take into account if a person is tagged on a photo as well as the publication date. Likes, comments and all interactions are also analyzed to select the content that will be most likely to appear in the news feed. Engagement by content type is also taken into account: if the person has reacted more often to video content than on photo content, Facebook is likely to favor video content.
