EdgeRank is one of the most important algorithms in marketing. Despite this, very few people have heard of it and fewer still can claim that they fully understand it.
EdgeRank is the name of the algorithm which Facebook uses to determine what appears in their users’ news feeds. The news feed is Facebook’s ‘Killer App’. There is a plethora of information available to Facebook users, and the newsfeed is the order in which it appears.
It determines which of your connections is the most important to you and thus appears most frequently, and which kinds of content should appear higher than others. For anyone seeking to market a product or service on Facebook it’s essential you understand how this algorithm works.
Understanding the rules of the algorithm and changing your tactics to reflect the system can make the difference between a business changing campaign and an embarrassing failure. Yet despite this huge importance very little has been written about the algorithm.
Unlike many of the algorithms that are changing marketing, Edgerank is actually not that sophisticated, but don’t let its relative simplicity make you underestimate the influence knowledge of the subject will have on your tactical choices.
Let’s start off with the name. Other than because it sounds cool, why is the News Feed algo known as EdgeRank? This is because every piece of content is known as an “edge”.
So, a status update is an edge; liking a status update, that’s an edge; uploading a photo, that’s an edge, too; or a change in relationship status? That’s also an edge. Basically, every interaction you have with Facebook that creates a piece of content is known as an Edge.
So, the newsfeed isn’t really a feed of news, instead it’s a chart of the most ‘important’ Edges which are determined by the EdgeRank Algorithm. What are the elements that make this algorithm? A combination of three factors: Affinity, Edge weight and Recency.
The EdgeRank formula is based on these three elements. While this does make the algo seem simple, there’s actually a huge amount of complexity behind these three factors.
Affinity
Affinity is a score based on the proximity or how “friendly” you are with someone. You’ve probably seen this in action. Spy on an ex-boyfriend or girlfriend, snoop on their profile and suddenly they’re in your news feed all the time.
Comment on someone’s photos and you’ll find them appearing in your feed more often. This is affinity in action. You’ve sent a proactive signal that you have a ‘close-ness’ to that individual or organisation. The algorithm acknowledges this and begins to order the results in your newsfeed accordingly.
Some people aren’t wholly supportive of Affinity having such a significant role in EdgeRank. The concern is that it becomes a self-fulfilling prophecy, i.e. the more often someone appears in your news feed, the more likely you are to increase you affinity, which in turn increases the likelihood of them appearing in your feed in the future and so on.
But given how people tend to cluster around only a small number of their connections it seems to work well for most people on Facebook.
One of the most important things you need to realise about affinity as a marketer is that affinity is one-way. This means you visiting a forgotten friends profile doesn’t increase the likelihood of you appearing in their newsfeed.
I’m sure for nosey people that is great news. It is less likely to be considered good news for company profiles. For example, if you visit a profile of someone following you, it will have no impact on your Edges appearing in their feed. However you commenting on a photo of theirs which then triggers them to comment back would lead to them having a greater affinity to you.
Edge Weight
Edge Weight is a basic formula which decides that certain pieces of content are more likely to appear in news feeds than others. Photos are more important than someone “liking” a business profile, etc.
There’s no definitive sequence of Edge Weight, but there are certain objects which acquire more EdgeRank than others. This can imply that they tend to have a higher Edge Weight than other types of content.
The three types of content which are widely understood to have the highest Edge Weight are Videos, Photos, and Links. Knowing that these are have the heaviest weighting should alter the way you communicate using Facebook. You should try and incorporate objects with high weight scores into any announcements that you want to reach as many of your followers as possible.
It’s also worth noting that each person’s Edge Weight is different, i.e. someone who likes browsing photographs is more likely to have them in their feed than someone who isn’t as keen on them. It’s not possible to understand which types of objects each of your followers favour, but this should at least deter you from the temptation of adding photos to every status update.
On the one hand, this will seem like a sensible strategy. It will have a high EdgeWeight. On the other hand, for the people who don’t often click on photos you might have more success with a link, a video or an old fashioned text status update.
It’s also worth noting potential political elements to EdgeWeight. At any point in time there will be certain features which strategically Facebook will want to push. Is it that far-fetched to assume that given Facebook’s desire to own Foursquare style logins in they might turn up the weight of Facebook Places Check-Ins?
You should recognise what Facebook are trying to promote, as it’s likely to influence the EdgeRank algo.
So, clearly variety is important, but understanding the relative weights of different types of object will help you increase the potential audiences of any message you’re trying to get across on Facebook.
Recency
The final element of the EdgeRank is related to recency, which is actually known as time decay. For example, regardless of how much the EdgeRank is based on Affinity and Weight, if it’s old news it becomes less likely to appear. This is different to Twitter which relies only on chronological order;
Facebook is still very reliant on the temporal nature of content, which is again pretty self-evident. Put simply, recency is that something newer is more likely to appear than something older.
The implication of this being that you should create objects and edges at the points in time when your audience are most likely to be using Facebook. This decreases the time decay and therefore increases the chance of your content reaching their news feed.
There is also another potential consequence of understanding the time decay factor. When appearing in the news feed you have exceeded the EdgeRank of other items. You might then increase the exposure by creating content when other people are less likely to be creating content, therefore reducing the competition for spots in the news feed.
Things you can do to increase your EdgeRank
Now if you’re a regular user of Facebook, many of these ideas behind EdgeRank are obvious, but with a good appreciation and understanding of how EdgeRank works you will begin to make more intelligent decisions in your Facebook marketing campaign.
The creativity of big campaigns is important. However, from my experience of Facebook Marketing, while it’s these huge efforts that always draw the attention and become the focus of case studies, it’s actually those lower profile campaigns which are built on constant interaction with their fans that really deliver the most value.
Whether you’re managing a campaign to make a big splash or a constant drip, there are plenty of takeaways from understanding more about more about how Edgerank works.
For example if you have an important launch coming up and you really want that content to appear in as many of your fans’ news feeds as possible, what can you do to increase the likelihood of it appearing?
First of all, you want to do all you can to increase the affinity between your followers and you ahead of the announcement. Perhaps you could start a debate on one of your status updates which lots of your followers contribute to.
Not only will the number of comments increase the likelihood of that particular item appearing in the newsfeed , but also anyone who has made a contribution will have increased their affinity to you.
Don’t be afraid to ask for the contribution! Buddy Media has found that using questioning words like Where, When, Why and Would have a dramatic influence on likes and comments, and as these likes and comments increase affinity, this has a positive influence on the reach of any future messages.
So, in the run up to your important announcement, plan a series of updates which have the purpose of trying to increase the affinity between your profile and your followers. Tease the update, ask questions, say something divisive, etc. You need a plan to increase the affinity of as many of your followers as possible.
This should be an on-going plan as affinity is something that will decrease over time, and generally speaking, you want as many of your updates to appear in front of as many of the right people as possible.
We also know that in general, photos tend to appear in news feeds more frequently than normal status update. Well then, perhaps you ought to accompany the release with a series of photographs. Videos and Links also have a higher than average weighting so should also be part of your posting plans.
So, media and assets need to be part of your Facebook strategy. You need to develop a content and asset strategy and a schedule for them. This can be comprehensive or flexible but thinking about what you can do to create updates with higher weight will enable you to interact with a bigger audience. Also, the fact that this type of content is also far more likely to be shared is an added bonus.
If you also know that the more recent your update is, the more likely it is to appear in news feeds, that should have an impact on when your announcement is made to ensure it coincides with your followers logging into their accounts. Explore when you get the most interaction on your posts and try and coincide future updates.
One of the biggest problem I have with most Facebook marketing campaigns is that they look to the number of likes a Facebook Page has the main measure of success.
Although it is very easy to measure, EdgeRank educates us that just because someone ‘Likes’ an account doesn’t mean it will appear in the followers news feed.
This means if you want your messages and conversations to reach your potential customers, you will find a smaller but more engaged audience may lead to more appearances in the feed than a larger but less engaged audience.
Kelvin Newman is SiteVisibility's Creative Director and is the editor of the UK's most listened to Marketing Podcast. He also spends his time at conferences, tweeting too much and working on top secret research and development projects. He's also on Google+.
Recommended reports
How To Create Amazing Facebook Pages
Econsultancy’s How To Create Amazing Facebook Pages Best Practice Guide has been produced specifically with the aim of helping organizations and brands design and implement their Facebook pages for maximum effect, no matter what the objective. This 90+ page document contains sections ranging from building an internal business case for a Facebook presence through to understanding internal strategy and moderation. It also has a comprehensive, 50-point checklist of best practice issues and their associated success factors that you'll need for your Facebook activity.
Alternatively, see our B2B Content Marketing Best Practice Guide or Location-based Marketing Smart Pack.
12:52PM on 16th August 2011
Really interesting guide, where I work we did some data mining to find out a bit more about how our users are engaging with us - might be of use as some of the points you mention (photos etc) came out in our research too: http://bit.ly/r7ea8y
1:05PM on 16th August 2011
I've read a lot of Econsultancy posts, and most of them only have one or two decent gems at best (at worst, they're simply wrong). This is by far the best and most useful post here, and I'll be applying all these considerations to my Facebook marketing campaign. Many thanks.
SEO & Social Analyst at Jellyfish Online Marketing
1:38PM on 16th August 2011
Thanks for the great article. A lot of the information has become apparent to me in the past, especially when I glance over my own 'Top News' feed but it's great to see the algorithm behind it.
Thank you also for the clear and concise considerations that we can take away and start using on our own Facebook campaigns. This information will be of use to not only myself but the clients I work with.
Head of EMEA Marketing at Venda Limited
1:48PM on 16th August 2011
This is first rate information Kelvin - thanks for posting it!
1:56PM on 16th August 2011
I wonder if number of friends/fans has any impact on the Edgerank of a post.
For instance, a popstar with a large number of fans is likely to get a massive number of likes and comments from their fans (seemingly sometimes no matter what they say).
However, those posts don't automatically dominate a feed, maybe because Facebook measures the number of interactions relative to the amount of fans.
Hope that made sense!
Senior PPC Analyst at Epiphany Solutions
1:59PM on 16th August 2011
A brilliant article - thanks for sharing. I've always suspected certain actions I take on Facebook determine the information I see in my newsfeed, but to hear the science behind it is fantastic.
At a personal level I find it very intriguing and interesting - but understanding the importance of Edgerank for brands takes Facebook marketing onto a whole new level.
2:27PM on 16th August 2011
Thank everyone for the kind words, James I don't think the number of followers in itself, would increase the weight, but the greater reach does increase the likelihood of it gaining weight, if that makes sense?
2:35PM on 16th August 2011
Good article Kevin. I wanted to know if this was directly from a Facebook source or if you have created the edge rank elements yourself? I believe you are missing a few elements another point to make is that the mobile Facebook news feed ranks things differently to the main site.
Interesting point James, I don't think Facebook takes the ratio of comments to fans into its equation. Although if you do find some kind of proof it would be interesting to see.
3:39PM on 16th August 2011
Been wondering how this algorithm worked for a while now - thanks for the info!
3:56PM on 16th August 2011
@robert this is all based on Facebook's language, a lot of this post is derived from a talk they gave a while back at their f8 conference about Edgerank
So there's a little conjecture and suggestion on my part but very much based on what they've talked about.
5:37PM on 16th August 2011
Great post & video. Been trying to implement the strategy for political campaigns, because campaigns tend to be - boring. One of the things we do is, we experiment with content. For example, if we throw red-meat out there, the hard core commenters show up and post their comments. So far that works. But again, have to vary the content to keep it exciting..
10:16AM on 17th August 2011
Excellent article. To be honest, I have never even heard about the edgerank algorithm before. Good to know. ;)
10:19AM on 17th August 2011
Excellent article. To be honest, I have never even heard about the edgerank algorithm before. Good to know. ;)
12:36PM on 17th August 2011
Good overview Kelvin.
It's indeed scary how few people know about this. Just won a Facebook-pitch against 3 other agencies & I think we were the only that based our proposal on Edgerank.
Regarding measuring success: Likes are obviously a skewed measure. Engagement Rate would be a good alternative.
And yesterday I discovered http://edgerankchecker.com/. Not sure yet how well it works, but it could be a great way to add another level of benchmarking
12:43PM on 17th August 2011
Exceptional article. Many many thanks! It really helps a lot. I plan to circulate this internally in our agency.
D.
Director at Match Me Now Limited
3:44PM on 17th August 2011
Agree it's an excellent article on a topic which does not get anywhere near the attention it deserves. Robert's point about mobile Facebook news feeds (and news feeds delivered via Apps) being different is defintitely true; does anyone have any ideas how the algorithm is different for the mobile site?
10:01AM on 18th August 2011
Great article, haven't read much about Edgerank algorithm before so it's very intresting to read!
1:21PM on 18th August 2011
tl;dr
1:23PM on 18th August 2011
Excellent article! Are you aware if Facebook can track whether someone watches a video on the newsfeed (such as a link to a video on YouTube)?
Marketing Internship at Currencies Direct
3:01PM on 18th August 2011
That. Was. Brilliant. Thanks for the info!
5:49PM on 18th August 2011
Good overview, Kelvin.
It is indeed scary how few people have heard of Edgerank. We recently pitched against 3 other agencies for a Facebook project & none of the others had mentioned Edgerank at all. We won ;-)
Re measurement. As you say, measuring the amount of Likes is indeed a skewed mechanic.
Taking into account the Engagement Rate is a good addition.
And this week I also found out about http://edgerankchecker.com, a new (?) tool that measures a Page's Edgerank. Not sure yet how good it is, but it could be a very useful 3rd parameter for measuring success. Has anyone had any experience with it?
2:33AM on 19th August 2011
Thanks for the article, great stuff
10:55AM on 20th August 2011
I never knew about edgerank! I'm so glad you explained it! Fantastic video Kelvin.
4:02PM on 21st August 2011
Kevin, thanks for adding Klurig Analytics' EdgeRank video to your post.
The video came out of some work we've done a few years ago, developing plans for an online news and social networking site. We quickly realized that we needed an algorithm that would help important information stand out from the noise, so we developed our own algorithm.
When Facebook's Edgerank algorithm, it was just a quick step to understand how they do things.
It all boils down to the fact that any social networking site is looking to provide the best and most important information to their users. And to do that, the site needs to try to understand each individual user's wants and needs.
In essence, it is very similar to how Google's Page rank algorithm works as well.
So, Kevin, thanks for a great article and thanks for adding our video.
4:21AM on 22nd August 2011
Great insights into the drivers behind a successful social media campaigns. Your recommendations with regards to time decay are particularly insightful. It's what you post and when you post it.
11:46AM on 22nd August 2011
Brilliant article, a great read.
My question is this: do updates to Business Pages from tools such as TweetDeck with links/media in them get included in the edgrank algorithm or is it updates that are made directly from Facebook?
1:24PM on 22nd August 2011
I was aware of edgerank before, spoke about it with many clients but it's great to know that when you preach something it can be backed up by articles like this.
Do you know however if edgerank changes depending on if the default setting on you news feed from 'friends and pages you interact with most' to 'all of your friends and pages'?
8:03PM on 22nd August 2011
Awesome information!
Thanks for sharing!
12:26AM on 23rd August 2011
I have the same question as Dennis. Have been aware of and working with EdgeRank for some time, but am still trying to figure out if the algo affects what a user sees when they have manually selected "all friends" (or "most recent" for a Page) instead of the default view.
(By the way, the comment captchas are ridiculous when they include sequences of upside down numbers and punctuation marks - especially when the instructions say "Type the two words". How is a user supposed to know what to do with those?)
6:49AM on 23rd August 2011
Is there any reason that when we look at the organic impressions in 'Insights' that we could have more Impressions than we have fans?
How could this be? Sometimes it is many thousand more impressions than we have fans, given that we are competing for news feed space and I doubt our fans share our content *that* much
- does anyone have an explanation?
9:05AM on 23rd August 2011
Fantastic article. Only thing that wasn't mentioned was status updates. Surely they have a high edge rank compared to links and photos? Always see them popping up.
Also read in a seperate article about how apps which automatically post (i.e. Twitter app) have an extremely low edgerank.
Main question is really, from a business page, is it more effective to put a status update with an @ link to another page of information on facebook, or is it more effective to put that as a link?
Director at Mediarun Search Ltd
9:08AM on 23rd August 2011
Interesting, insightful, clear and concise. The best econsultancy posting I have seen for a long while.
9:59AM on 23rd August 2011
Michael - you are probably counting impressions, rather than unique impressions. Also unless you have a like gate non fans can also visit.
11:59AM on 23rd August 2011
You cannot comment on other peoples photos/profiles etc as a company page!! I don't understand why so many websites get this information wrong!!
12:20PM on 23rd August 2011
Sorry, I felt compelled to comment again.
This article is not precise whatsoever.
Two pieces of information you have given is wrong. The first being that you can comment on peoples photos as a brand page, which is impossible.
The second being that you say that by posting a status update which encourages interaction means you will gain better affinity with your fans. BUT HOW WILL YOUR FANS EVEN GET TO SEE YOUR STATUS UPDATE IN THE FIRST PLACE IF THEY DON'T HAVE ANY AFFINITY WITH YOU??
It is a catch 22 situation which this article has not address at all.
Creative Director at SiteVisibility
12:47PM on 23rd August 2011
Fair comment on the photo comment element Aimee, I should have made clearer that's more of an approach for a individual rather than a brand.
And with the interactive issue you're right it is a bit of a catch 22, but what I'm getting at is a comment thread with 77 comments is more likely to appear in someone's feed, due to it's additional weight - which attracts further comments which increases their affinity in the future.
Which I'm not sure makes much sense when I explain it in text here...
1:10PM on 23rd August 2011
ye it is hard to explain in writing!
but you did give advice which stated that if you were launching a product and wanted people to see content surrounding that, then people should create a status update which encourages interaction, which will lead to them seeing the content. but that's not entirely true as how will people see that update in the first place....
1:57PM on 23rd August 2011
Great article, well done. Do you think the option at the bottom of your feed over-rides all of this? E.g Where it has Edit options with the drop down "Show posts from all friends and pages" / "Friends and pages I interact with most"? Common sense would say yes but you never exactly know!
2:04PM on 23rd August 2011
Aimee - By doing the other things he mentioned: Including photos/videos/links in post. Varying the "type" of posts you make. Timing the posts to times that people are most active on FB. All those will help contribute to your posts being higher up in users' newsfeeds.
He's not saying that if you have no or low interaction your posts will not be visible at all. If you do all the other things suggested and include a question to engage/encourage discussion you'll be able to help increase your edgerank for future posts. :)
Kelvin - Thanks for the informative post, I've realised our company already does quite a lot of your suggestions, nice to see it actually has a statistical support behind it! :)
Do you have any info on how edgerank applies to "top news" vs "most recent" options in the newsfeed? I know that, by default, FB has "most recent" automatically set to "friends and pages you interact with most" over "all of your friends and pages". Which implies that, despite being called "most recent", there is still some aspect of ranking applying to what actually appears in "most recent" set newsfeeds.
2:21PM on 23rd August 2011
Hi Matt,
Thanks for that but I was just pointing out the bit where he gives advice about the product launch. That's all. It would have definitely been worth mentioning the catch 22 for those who are not so clued up on social media! :)
2:24PM on 23rd August 2011
Tracy, in my experience I have my own personal Facebook set up to show updates from everyone, yet Facebook's Edgerank still seems to be in place as I only see updates from a select few. I am not sure about anyone else though...
4:39PM on 23rd August 2011
I'm of two minds of any discussion of algorithms. While it's great to identify factors on their own, a lack of any real-world examples and data-gathering might make the stated factors be weighted in ways that might not make sense at first blush. You mention the lack of value of a like on its own, but that could be a major push in Edgerank if there's no statistical basis.
The other side to it, though, is that data isn't everything, especially with something as individual as EdgeRank (though a minor comparison can be made to search engine ranks) or Twitter's algorithm. It's I don't think the "Algorithm Marketing" post is particularly apt, since the "Field of Dreams" style "Do good stuff and the EdgeRank will come" mentality is both beneficial to customers/readers and avoidance of clearly manipulative techniques.
Have you found a way to calculate any base values for likes or other interactions?
Digital Marketing Specialist at Livelink New Media
12:43PM on 24th August 2011
A Very interesting article thank you Kelvin, I have some questions...
Do you believe that sentiment also plays a part in this ranking algo? - For example a FB Page could have a large amount of people complaining about a product or service, so would this mean it would then be more likely to appear in the news feed?
However contradicting myself - Users could always just unlike the page, But you would think that sentiment would have some relevance also to keep things relevant?
One last point: Pages which have been reported or other users have complained about to FB - Would these have less of a chance in appearing in users news feeds? Does Facebook "rank" company pages Separately? - Pages with less complaints per followers/visitors is more likely to appear in news feed?
CEO at EngageSciences
2:26PM on 24th August 2011
Perhaps the most important thing in what is a very good article is Kevin's assertion that smaller campaigns on a frequent basis is a far better approach than big splash one-off campaigns which is often promoted by integrated marketing agencies that are being contracted for single marketing initiatives such as product launches.
For anyone that works in social media marketing, it is up to us to provide education to the end clients that social media is something that requires a calendar of activities designed to drive interaction, engagement and affinity. If clients do this then tools such as http://engagesciences.com/ can provide reach datasets on each and every fan and how they are spreading the clients campaigns to their friends.
The true value of social media for marketing purposes is in being able to reach the friends of fans and democratize marketing. Comscore produced a great report entitled the 'Power of Like' which politely points out that measuring the value of social marketing by engagement with fans does not give justice to the real value of social marketing. Rather it is the ability to reach the friends of fans through likes, comments, shares and invites that presents the greatest opportunity. Bing for example has 1.7m fans on Facebook and just the friends of those fans numbers 232m people. Being able to track how each fan is spreading your campaigns to their friends and then rewarding them for that behaviour is the next frontier for social marketing. Full disclosure: I am the the CEO of EngageSciences and a social marketing advocate who works with leading brands such as Nokia, Play.com, Speedo, and Forbes at helping drive earned media marketing strategies.
8:39PM on 24th August 2011
Great article!
Well put together and easy to digest- true, how few marketers realize the ER algorithm determines only about 12% of their audience seeing the news feed!
7:20AM on 26th August 2011
Really useful articlle and incredibly 'easy' to read. Definately things Ill use in the future. Thanks guys
7:28AM on 4th September 2011
Thanks for elaborating the algorithm in simple and easy way and I agree with you that I have never heard ER algorithm before and I am sure if I would not have come here, I would have never come across the EdgeRank algorithm. Thanks for the nice article indeed.
9:49PM on 8th September 2011
Facebook is really unique in how they are collecting user metrics through social media, I feel that is a major reason why Google is chasing after them so hard with Google Plus. They are afraid they'll lose to a FB future, so they need to incorporate social media + search now.
4:27PM on 12th September 2011
Great article, Kelvin! Although only Facebook truly knows how EdgeRanking works, this seems to be as close as we're going to get.
Looking forward to reading more of your articles.
4:41PM on 17th October 2011
Great information on best practices for posting on Facebook. Thank you!
4:31AM on 18th October 2011
I've been trying to optimize our facebook presence for some time now and only recently learned that the process is called EdgeRank and how the algorithms work.I believe this will help us greatly.
1:18AM on 25th October 2011
Thanks, Kevin. I think you're right. I think EdgeRank is almost as important as Google ranking. Of course, it's merely one aspect of a complete, well-rounded business plan with online and offline marketing methods. The future of facebook still looks bright and they seem dedicated to improving their product all the time.
Hearty regards, (c:
-Mark Lipka
4:43PM on 16th November 2011
I agree with you Mark, actually Kevin has provided complete insight of the EdgeRank, I think which is equally important to understand.