How Facebook's News Feed Picks Stories You'll Like

Facebook's 2012 patent explains how it creates a personalized news feed by showing stories about friends' actions, ordered by your interest, and updating it based on what you click.

What it covers

This patent describes a system for creating a personalized news feed on a social network, like Facebook. It starts by looking at who you're friends with and what they've been doing. Then, it generates 'news stories' about their actions. These stories are put into a news feed, and the system tries to guess how much you'll like each one, putting the ones it thinks you'll like most at the top. Crucially, it watches what you click on or interact with in the feed and uses that information to pick even more stories for you, and to decide the order they appear in. For example, if you click on a friend's photo album, the system might show you more stories about that friend or similar content.

What it doesn't cover

• —News feeds that are not based on a user's connections within a social network
• —News stories that do not describe an action taken by another user
• —News feeds where the order of stories is not influenced by user affinity or interactions
• —Systems that do not monitor user interactions to update the news feed
• —News feeds that do not allow users to change the order of content

The clever bit

The key innovation was using a user's 'affinity' for content and their actual interactions to dynamically rank and update the news feed, moving beyond a simple chronological display to a personalized, engaging experience.

Why it matters

This patent covers the core functionality of Facebook's News Feed, which launched in 2006 and quickly became a defining feature of social media. It explains the algorithmic approach to content delivery that keeps users engaged by prioritizing relevant updates from their social graph.

Real-world examples

• 1.Facebook News Feed (2006 onwards)
• 2.Social media content feeds