Facebook has long used an algorithm to filter your feed, and an old link resurfaced Monday sheds some light on roughly what it cuts. The link, http://www.facebook.com/?sk=nf_all, was posted by Scottish developer Tom Waddington and displays your feed under an older and less restrictive version of Edgerank, Facebook’s ranking algorithm. While even that doesn’t show every status and photo your friends post, it reveals considerably more than Facebook’s current algorithm — which, according to TechCrunch, cuts as much as 85 percent of your feed.
What, exactly, does the Facebook filter cut? According to a presentation Facebook engineers gave in 2010, the algorithm accounts for three things: the type of post, the age of the post, and how much a user interacts with the person or brand behind the post.
(Cool trick: A browser extension by developer Jeremy Keeshin claims to show you your algorithmic “affinity” for other Facebook users, or the likelihood that you’ll see their posts.)
Edgerank should prevent you from seeing updates you don’t care about but for me, at least, that isn’t quite true. In my filtered feed: Gawker’s review of Lincoln, a news article about cats, some whiny poetry from a high school acquaintance, a photo set of some classmates at a party. In my less-filtered, all-link feed: all the same things, plus a few more posts from companies I follow and a couple updates from people I forgot I knew. There’s no saying which feed is actually more relevant or useful. In fact, sometimes the filtered feed feels totally arbitrary, showing a friend’s photo one minute and hiding another photo from the same friend the next.
Regardless of how much or what it filters out, the mere existence of the filter angers marketers and journalists, who have seen impressions plummet in recent Edgerank iterations. Users seem none too happy, either; a Facebook page for Edgerank bears only comments like “I do not like your page” and “you are really missing the mark,” and a month-old petition on Change.org asks Facebook to make the filter optional.
Edgerank also raises the specter of the “filter bubble,” the theory that personalization algorithms are increasingly limiting our information and points of view. Such algorithms are hardly unique to Facebook, of course. Netflix, Amazon and Google personalize what they show you, too. But while that type of filtering may be convenient for some — does anyone really want to see a thousand photos of his ex-classmate’s kids? – critics have also argued that they make for a less-informed public.
In his 2011 best seller on the topic, Internet activist Eli Pariser took a particularly condemnatory stance. “Personalization filters serve up a kind of invisible autopropaganda,” he wrote, “indoctrinating us with our own ideas, amplifying our desire for things that are familiar and leaving us oblivious to the dangers lurking in the dark territory of the unknown.”
On Facebook, at least, the unknown is now slightly more knowable. But you’ll need that special link to see it.