Facebook Reveals More Insight Into The News Feed – Full Breakdown

For the first time, in a long time, Facebook has addressed more details regarding the EdgeRank algorithm. Facebook’s news feed product manager Will Cathcart simplified how the news feed works with the following:

1. Prior interaction with an author’s posts: If you like every post by a page that Facebook shows you, it will show you more from that page.

2. Other people’s reactions to a specific post: If others on Facebook are shown a post and ignores it, or complains, it’s less likely to show up on your news feed.

3. Your interaction with posts of the same type in the past: If you always Like photos, there’s a better chance you’ll see a photo posted by a page.

4. If a specific post has received complaints by other users who have seen it, or the page who posted it has received lots of complaints in the past, you’ll be less likely to see that post. This factor became a lot more prevalent with the EdgeRank update in September 2012.

We’re going to break down each component to help provide deeper context into what that means for your brand.

1. Engagement History (Affinity)

Has this fan engaged with your page’s content before? Facebook wants to determine the likelihood that a fan will engage with your future content. They use historical Engagement to understand the affinity between a fan and a page. This is confirming a long standing concept of EdgeRank.

2. Engagement Totals (The Sum of Edges)

Facebook is looking at a few things on this one: first, they specifically mention people who ignore a post. This metric is ultimately counted within Facebook’s definition of Virality. Are the people who are seeing this post taking the time to engage with it? Facebook’s definition of Virality is often questioned, however in this example it makes sense why they report it the way they do. It should be noted however, that we haven’t typically seen a strong correlation between Virality and Reach.

Facebook also confirms they are looking at Negative Feedback to determine the quality of the post. Mathematically, we look at Negative Feedback as simply another Edge, but with a negative value. We hypothesized that the weight of a Negative Feedback Edge was increased significantly with the most recent EdgeRank change in September 2012. Brands that tend to receive a lot of Negative Feedback might have been impacted more drastically because of this.

3. Historical Content Type (Weight)

Facebook implies that the weight of an initial object also carries historical Engagement significance. For brands, it’s imperative to know how your audience likes to consume your content. If links for your brand tend to struggle, try using a link within a photo’s description and monitor the results.

4. Negative Feedback (Weight / Sum of Edges)

This point somewhat mirrors point #2, however, there is one additional piece of information. To paraphrase Cathcart, pages that have received a lot of complaints in the past, will struggle with Reach today. This is an incredibly interesting piece of information. There have been rumors of a Page’s Quality Score or Authority, similar to Google’s PageRank. We often look at PageRank’s evolution as an analogy to EdgeRank’s. It’s difficult to differentiate exactly what’s being said here, but for now it must be assumed that receiving Negative Feedback in the past impacts Affinity with those particular users, thus resulting in the decrease in Reach. However, it could be argued that there may be a Page Quality Score. For the time being, we’re going to chalk it up to decreased Affinity with the users who reported the Negative Feedback, therefore creating a Page Quality-like affect. Either way, Negative Feedback needs to be taken seriously to reduce any damaging affects to your Page’s Reach.

Cathcart went on to explain “we started penalizing things that had an above average rate of complaints, and rewarding things that had a below average rate of complaints. Facebook believes the change was a success because engagement went up” and “complaints went down in the double-digit percentage.”

What Is New About This?

Ultimately, the only new revelation is the increase in the weight of Negative Feedback. Even with the inception of a potential argument regarding a Page Quality Score, the point is moot. Regardless of how it’s currently calculated, the ultimate effect remains the same. Historically receiving high Negative Feedback will result in decreased Reach. Similarly, receiving high Negative Feedback on a current piece of content will also cause poor performance on the news feed.

These insights confirm that Facebook has indeed made a significant change to EdgeRank, and the change is rooted in the impact of Negative Feedback. It’s interesting to see that they’ve achieved the result they’re looking for by reducing complaints within the news feed. The rest of the points are more or less confirmations of previously tested concepts.

We believe that these changes will hopefully continue to make the news feed more relevant to users. This increases the value of the news feed, an also challenges brands to continue to engage their audiences. As long as Facebook is rewarding the brands that engage their audience, these changes will be welcomed.

17 Comments on “Facebook Reveals More Insight Into The News Feed – Full Breakdown

  1. Diane

    That’s all well and good….except that I had good engagement with my fans according to FB’s previous measurements and then overnight, no one was (is) seeing my posts…I use a lot of photos which got comments. But, if I post a photo or link…no one sees it. That just makes no sense at all. And how do we know if we get negative comments?

    • Chad Wittman

      Right now, you have to download your Page Insights. We’re working on something more intensive for EdgeRank Checker.

    • Chris

      You hit a key point about links in posts!

      I have noticed the same thing. When you put up a typical post that doesn’t direct users to an external site, it gets more views. However, when you put up a post with a link in it, the views dramatically drop!

      My conclusion: This has nothing to do with negative feedback; rather, it’s a money thing! Facebook wants page owners to pay them money to get their fans to view posts with links in them leading to external sites.

      The bottom line here is this: Facebook is not fair.

  2. Actually of negative feedback os going to be their matrix. .. why shouldn’t I just report all my competitors pages… of someone doesn’t like something and its to the point where they want to complain… then with the complaint the connection should be broken… unfriend or unlike. .

    At this case… it seems that we as owners need to walk on eggshells to make sure the everyone likes every thing we post or it will eff it up for others. ..

    • sonja

      I was thinking the same thing. Unfortunately there are users out there that will do that. Guess it really boils down to having to pay for your posts to be seen.

    • Virtually every post we create gets two negative feedbacks, and they almost always occur at the same time which makes me think that it is probably one person with two profiles. It never seems to affect the rate of impression growth, so I suspect that Facebook has implemented some anti-sabotage components to EdgeRank. If I were them I surely would: nothing would hurt their brand more than the discovery that one user could sabotage the performance of another user’s experience.

  3. Christina Mejia

    Thanks, this is helpful.

    Any insight into what “an above average rate of complaints” might be?

  4. Dalene Smith

    Why does FB find it necessary to “think” for us? Can’t we just unsubscribe or ignore the news feed we don’t like? This does not seem like a good idea, to me.

    • It seems like a good idea to me. EdgeRank gives me fewer inane Action reports (e.g. “Martha Jones commented on Sarah Jane Smith’s photo”) in my Newsfeed in return for more of my friends Posts and more information about what’s going on in my Groups.

      Also, remember that by taking into account things we like, share, and comment on it is letting US decide what we see instead of trying to predict what me might want to see (based on keyword analysis or site cookies). THAT would be “thinking for us”. EdgeRank is kind of the opposite.

      • Luciano

        I agree re: managing flows of information. Especially given the sheer volume of info out there.

        But let’s get real here: all these metrics (‘liking’ something, ignoring something, giving negative feedback, etc.) have absolutely no bearing on people’s real, actual lived experience of interacting with media. Plenty of images show up on my newsfeed that I enjoy or find valuable, but I don’t necessarily feel compelled to ‘engage’ with… The only *REASON* for these metrics is to lend the illusion of measuring concrete, material actions or ROI where real benefits are kinda immeasurable…

        So this algorithm is designed around a shoddy and artificial set of assumptions about human nature IMO.

        I also worry that we’re heading towards a future where ‘smart indexing’ transforms my online experience into a self-reinforcing bubble where I don’t stumble upon new things or confront different opinions…

        And re: “thinking for ourselves”—wouldn’t it be a billion times easier to give users agency to customize their own newsfeeds? Sort people / pages into groups of priority and adjust these so, say, I get 5% of my uncle’s angry diatribes about immigration and get 85% of all posts from my cousin’s sister’s co-worker? Or whatever I want? You could do this a thousand different ways, really. I don’t really buy your argument that Edgerank is actually enhancing or supplementing my ability to think for myself. Particularly if its ‘rules’ are so mysterious and 80-90% of FB users don’t even know what “Edgerank” is…

  5. John C

    The consequences of changes by FB ultimately mean brands will start to ask challenging questions of FB as to the value versus cost – just as they do with other forms of media. In time FB will be forced to justify and provide evidence of ROI and effectiveness as a marketing channel. Brands in turn will come to realise that FB is just one cog in a very large marketing channel wheel.

  6. Great article. I have seen a substantial impact from the algo update as well.

    P.S. You should consider adding images to your blog, it will help with your EdgeRank in Facebook when I share it :)

  7. Adi

    Hi Chad and thanks for this great post! My question to you is, how one defines negative feedback from fans. I mean, do they (Facebook) calculate negative feedback only when someone actively reports it as negative or do they analyze sentiment over the content that people are writing within the comment itself?

    • Chad Wittman

      I do not believe they are currently doing sentiment analysis. Negative Feedback is defined as hiding the objects or reporting as spam.

      • Luciano

        Yeah, and the article mentions that pages will be ranked down if “someones shown a post and they ignore it”?

        Does “ignoring” simply refer to the # of newsfeeds your post appears in but isn’t engaged with / no action is taken (ie: like, click, share, comment, etc?) or is it somehow more complex than this?

        This seems like an incredibly poor metric. Not everyone actually ‘engages’. I know so many people who have never commented, liked, etc. anything I post yet will approach me in person telling me how much they enjoyed this or that… I mean, you could say there’s such a thing as active engagement and also passive engagement…

        • Chad Wittman

          By ignoring an object in the news feed, you decrease the “Virality” of the object (from Facebook’s perspective). Facebook monitors clicks within an object, hovers of negative feedback indicator, which are also inputs into how they perceive the “engagement” of the content. While it not be 100% accurate (and it never will be), Facebook can statistically gauge the general engagement of the object.

  8. Mike

    We have noticed a significant drop in views when we post content every since the “promote your page” option. We have tried different methods since that change with very mixed results. This was a good article, thank you.

Leave a Reply

Your email address will not be published. Required fields are marked *

9 − = seven

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>