Understanding Facebook Engagement with EdgeRank

For a little over a week now, I have been the newest member of the Backbone Media team. As most of you have probably noticed, there have been some recent changes to the layout of your Facebook news feed. Although this may have initially thrown some people off (evident in the flurry of negative feedback seen around the web), the intent was to provide a more personalized experience for users.

Facebook’s continued struggle is essentially blending your real life experiences with relevant content and stories posted from friends and brands – all organized in an easy to digest timeline. The monkey behind all of that is a set of algorithms called EdgeRank. On September 21, Facebook improved its EdgeRank system to display content and interactions from your friends to be more relevant. Meaning, EdgeRank is optimizing Top Stories for users who are more likely to interact, thus dropping less engaging content into your Recent Stories – we’ll get back to this later.

For those of you who find yourself instinctively heading for the “X” at the top of the page, fear not, I promise it’s not as frightening as it looks. Before we get into the actual formula it’s important to understand what an edge is and how they’re created. Edges begin simply as objects, which can be any input (a video, photo, like, status update, etc.). Three factors are then taken into account when determining the edge associated with an object.

The EdgeRank formula is based on three elements: affinity, weight and time decay. While this does make the algo seem simple, there’s actually a huge amount of complexity behind these three factors.

1.) Affinity score: An object’s Affinity score is determined by the amount you interact with a person/brand’s objects. For example a person whose photos you view often will have a high affinity score. Higher Affinity scores drive higher EdgeRank.
2.) Weight: Weight is believed to be determined by the current objectives of the Facebook team when the object is created. The most recent weight hierarchy generally places videos, photos, links, and status updates at the top of the list. Just like with Affinity scores, greater weighted objects receive higher EdgeRank.
3.) Time Decay: The last factor in determining EdgeRank ensures that old “news” does not repeatedly appear in the top news section of your feed, an object must receive constant interaction or its EdgeRank will decline.

These three factors work together to give every object an EdgeRank. The higher the EdgeRank, the more likely the object is to appear in a user’s top stories.

So what can be done to ensure your brand is able to break into a user’s news feeds?

First, brands need to provide continuous, engaging content to ensure affinity scores among users remain at high levels, and that time-decay does not lead to page stagnancy. Asking questions about who, what, when, etc. are great ways to keep users engaged. An emphasis on posting highly weighted objects such as videos, photos, and links will also contribute to a strong EdgeRank. It should be noted that content quality should not deteriorate, as interesting creative (stories, posts, videos) are still the primary drivers of user engagement.

Now with an understanding of this new process of populating the Top Stories section of a news feed, we took a look at a sample set of Backbone clients to determine what effects this has. Four criteria were taken into account in our analysis:

1.) Daily Active Users–The number of people who have interacted with or viewed the brand page or its posts.

2.) Daily Logged in Views–Daily page views from users logged into Facebook.

3.) Daily Likes/Comments– The number of likes and comments made on news feed stories.

4.) Daily News Feed Impressions–The raw number of times people have viewed a story in a news feed.

The Results.

Looking at the percentage change across these four metrics showed substantial increases virtually across the board:

• The increase in daily impressions suggests high EdgeRank scores are associated with these clients. The remaining data implies that users, who are now able to quickly and easily find news from their favorite pages, are more inclined to engage.
• Although this defines a positive correlation between users and established pages, it may pose additional issues for brands new to Facebook by making it more difficult to break into user’s news feeds.
• Brands with a higher post frequency will see higher daily active user scores, but I would be careful of posting too much for the sake of EdgeRank – content is still the biggest game in town.
• Cross referencing page engagement metrics with the Reach Insights will uncover more insight to maximize your content and its virality.

As of this morning, Facebook launched a new set of Page Insights to further glean insight to the types of content users engage with, as well as what people are talking about. More on that later…