Facebook calculates the friend ranking scores by analyzing all of a user’s on-site behavior and connections. Other signals in addition to those listed above that likely contribute to these scores include who a user is tagged in photos with, how many mutual friends they, if they Like the same Pages or are members of the same Group, if they check in together, and if they have matching biographical characteristics such as current city, hometown, employers, or education history. The score evolves and becomes more accurate over time.
Facebook uses the friend ranking score to increase the relevancy of its products by making sure users see news feed stories about the people they’re most interested in, can quickly find them in search and Chat, and receive push notifications about them on their mobile devices. Facebook engineer Keith Adams confirms that a user’s browsing behavior only impacts what they themselves see, and that viewing tons of photos of a certain friend won’t make a user appear in that friend’s news feed any more frequently. Users have long been interested in finding out who browses their profile though this information is not available, leading the rise of fake “profile spy” applications that are actually scams.
These scores have been built from years of data and thousands of actions, and the ability to customize the site for relevance is one of Facebook’s greatest strengths. The lack of this data would hamper the functionality of competing social networks such as Google+, even if they had your friend list.
At the top of the friend rankings list users will see their closest friends, people they’ve interacted with a lot recently, and those they frequently visit the profiles of. At the bottom they may see non-friends who they’re connected to through mutual friends or who they used to be friends with.
By using the right-click option “Inspect Element” in Google Chrome or Firebug in Firefox, users can view the file “first_degree.php” which includes the friend ranking data. Facebook pre-loads this file to speed up typeahead search results. In it, users will notice “tokens” associated with some friends. These are alternate names that can be used to initiate a friend tag in a status update or photo or find these friends in Facebook search, and include long and short forms of common names, familial relation types, and old names of users who’ve since changed their Facebook name. For example, a friend named Richard would have the tokens “rick” and “dick”, while a friend that a user has confirmed as their brother would have the token “brother”.
Users who want to view their friend rankings should do so now as Facebook may modify its code to prevent users from seeing this data. This is because knowledge that such a list even exists could scare users into avoiding any embarrassing browsing behavior. It’s these extended viewing sessions of an ex-lover’s photos or the wall of an admired peer that help drive the social network’s massive average time-on-site, so Facebook probably doesn’t want users to think these friend rankings could ever see the light of day.