The social network already uses a technical algorithm to generate a personalized News Feed for each user. The algorithm, which the company explained in a recent announcement at Facebook's Menlo Park headquarters, takes several factors into account to determine its place on the News Feed: the author and number of likes, comments, and shares.
Since News Feed is uniquely populated for each user, Facebook first needed to define "high quality" content, software engineer Varun Kacholia wrote in the company's blog post. Facebook surveyed "thousands" of users on how they deem high quality content, folding the responses into a new machine learning system that was integrated with the master algorithm.
The updated algorithm considers "over a thousand different factors" — including the quality of the page's other content and the level of completion on the profile — when determining whether a post is high quality, according to the blog post.
Facebook says it has tested the new algorithm on "a small segment of users" and reports that those users engaged more with content on their feeds and hid fewer stories. The new algorithm will be rolled out over the next few weeks to users on both desktop and mobile, according to the post.