Facebook has two algorithms that are important to marketers and developers but are largely misunderstood by people across industries.
Graph Rank and EdgeRank are what the social network uses to organize the massive amount of activity generated by people, pages and apps and to decide what stories to show whom. Graph Rank applies to Open Graph applications, not page posts or friend’s status updates. Those are affected by EdgeRank, which determines what shows up in your News Feed. Neither of these is a global score. Every bit of activity on the site has a different rank for different users. And because time and affinity plays a role, the rank of each object is not constant.
EdgeRank is the algorithm that determines what items populate your News Feed. With all the friends people have and pages they like, most users would be overwhelmed to see all of the activity generated by these connections. Facebook, therefore, assigns a value to every possible story that could end up in the feed. This value is based on affinity, weight and time. Affinity is the relationship between the user and the page or friend that created an item. Weight is affected by the type of story, for instance, whether it is a photo upload or a comment on another person’s status. Facebook tends to value rich media content and often when it introduces something new like Questions, it temporarily weights activity from that feature higher. The third factor affecting EdgeRank is how recently an action was taken.
These factors are why you might see every check-in and linked shared by your best friend, but only see whole photo albums from someone else. EdgeRank is also the reason most fans don’t see every post from pages they like. The more users interact with the page, however, the greater the affinity score becomes and the more likely they are to see page posts in the future.
[Update July 25, 2012 – Facebook says it no longer refers to its algorithm as EdgeRank, though many people outside of the company still refer to it that way. Facebook software engineer Phil Zigoris told Inc.com:
”EdgeRank is a term that has been used in the past to describe how we optimize the content of news feeds based on what is most interesting to you [as a user]. We don’t have a product or system called EdgeRank. The news feed algorithm takes many factors into account when it’s deciding what to show, including how often two people interact, how many people have liked, commented on, or shared a story, and the type of content that is being shown.”]
Graph Rank is a new algorithm Facebook is using to determine how Open Graph application activity will be distributed through News Feed, Ticker and Timeline. Graph Rank was introduced at F8 with the announcement of a new type of app that can continuously publish user activity to Ticker and Timeline. Because Facebook expects a proliferation of sharing through these Open Graph apps, it developed a system to manage the amount and type of activity that each user will see. Graph Rank seeks to show users highly relevant application stories based on the other connections they’ve made on Facebook. So if a user plays Words With Friends, they are more likely to see a friend’s story about another word game than a story from an arcade shooter game.
This seems to be a reaction to the negative feedback Facebook received after first allowing third-party apps on the platform. Many users were frustrated with the amount of app activity that filled their feeds. The social network initially responded by cutting off several viral channels that allowed apps to grow organically. This left a sour taste in the mouths of many developers. With Graph Rank, Facebook hopes to strike a balance that helps users discover apps they are likely to enjoy without compromising the site or turning off developers.
Together, EdgeRank and Graph Rank help personalize Facebook for each user. Marketers and developers who understand the way the platform ranks content can find ways to optimize their efforts there.
Images from http://www.livestream.com/f8conference and http://www.livestream.com/f8live.
See also: People Talking About This Defined and Understanding the Difference Between Sponsored Stories, Page Post Ads, Promoted Posts and Marketplace Ads