Separating truth from fiction about Facebook

March 5, 2013

There’s a lot of inaccurate information out there about the way Facebook is promoting posts from people who pay for it. Some of this misinformation comes from writers using their experiences as an example of how things happen to everyone on Facebook, not realizing they’re different than many other people on the service or other people who use the “follow” option (formerly known as “subscribe.”) There’s also an unfortunate tendency to not check facts with Facebook. 

We should not take Facebook at its word but we should at least give it a chance to explain how it sets the rules. We can judge for ourselves how honest they’re being. It’s far worse to assume. This isn’t an entirely new thing for Facebook, Sponsored Stories were rolled out back in January of 2011. This resurfaced because of an article by Nick Bilton at the New York Times who suddenly seemed to notice his Facebook posts were not seeing the same number of comments, likes and shares as they once did. (Update 3/6: Nick does cite Facebook’s statement that although “people who have more than 10,000 subscribers is up 34 percent from a year ago” they also admitted to him “there has been a 2 percent drop in interaction on the news feed.”)

Misconception #1: Sponsored/Promoted Content is replacing organic content on Facebook I spoke to Vadim Lavrusik, Facebook’s journalism program manager. Here’s what he told me:

“One important thing to understand is that when someone promotes a post in feed and pays to promote it, the stuff that’s getting distribution organically still gets distribution, it doesn’t get replaced from feed. It may get a lower placement, but it doesn’t get replaced. And the placement of the sponsored post or promoted post is also based on the quality of that post (so promoted content still has a quality algorithm attached to it.) If the promoted post isn’t that good, it gets lower placement, but it will get more distribution either way because it’s being paid for, but it’s still takes quality into account.

The claim that I’ve seen explains it as if these paid posts replace organic posts, which isn’t the case. The News Feed algorithm is separate from the advertising algorithm in that we don’t replace the most engaging posts in News Feed with sponsored ones.”

Misconception #2: Quality of Sponsored/Promoted content doesn’t have an bearing on how it is displayed Vadim was quick to point out that even though you’re paying to have your content get better placement, it will still be judged by quality factors similarly to how organic is judged by.

“The other thing is that you have to keep in mind [how the] feed works overall. On average, a Facebook user is eligible to see more than 2,000 pieces of content per day. That’s a lot of content. We’ve tried showing that content in chronological order, and have tested that every time we have, the overall engagement would decrease dramatically, and users would actually see less content because they didn’t find the content they were seeing as interesting and were missing the important stuff that maybe was posted hours ago, so we take each of the 2,000 pieces of content and rank it based on the probability that you’re going to interact with it. On top of this, sponsored content is added in and may bump organic content down, but we’ve minimized the decrease in engagement and make sure to not show too much of this if the user reacts to it negatively. Overall, the feed ranking looks at each piece of content and orders it based on the probability you’re going to interact with that piece of content. This includes things like:

  • Types of content you interact with. (Photos, status updates, links) If you interact with certain post types, you’ll see more of those post types.
  • People/Pages you interact with more often than others (your closeness to those people), so if you interact with someone more often you will see more content from them. This interaction can be anything from going to their timeline/page, interacting with their content, messaging, etc.
  • Other people’s interactions with the piece of content. If a post is getting a lot of engagement based on how many people have seen it in the feed, it will be seen by more people. If your friends interacted with it, this is a stronger signal.
  • Negative feedback: If people have marked it for spam or hidden it, it will be shown to fewer people. This has a bigger weight than other interaction points.”

Misconception #3: Facebook refers to how they judge post quality as “EdgeRank” Facebook certainly has an algorithm for judging post quality, as described above, but nobody at Facebook actually calls anything “EdgeRank,” Vadim told me. It appears to be a term that caught on outside of Facebook.

Taking all of this into account, it’s hard to see how Nick’s drop in engagement would be true for other users, whether they’re “normal” Facebook members or journalists using the “follow” option. There’s a few things that make Nick an edge case, someone who uses and experiences Facebook slightly differently than the broader membership. He was one of the privileged few who were “recommended” members to follow, which allowed him to gain a lot of followers early on. Most members have to scratch and claw to get noticed, a recommended user list gives you an opportunity to catapult your following in a less organic fashion. Just like on Twitter, it creates an inauthentic illusion of “influence,” and as much as I loathe that word and the next one I am going to use, “engagement,” the quality of that “engagement” goes down as your artificial following grows.

I also noticed that Nick tends to post a lot of links, instead of photo posts, which tend to get a lot more “likes” “shares” and comments. If they’re not getting that kind of (ugh) “engagement,” then they’re in turn showing up lower organically in your follower’s newsfeed. This is a feature, not a bug.

No comments so far

We welcome comments that advance the story through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can flag it to our editors by using the report abuse links. Views expressed in the comments do not represent those of Reuters. For more information on our comment policy, see