The Facebook News Feed
Unearthing (at least some of) the reasoning behind content curation on social media
Small business owners use social media for a variety of reasons — from growing your brand awareness, to interacting with potential customers, or just being a member of various social media communities. No matter how you spend your time on social, it’s wise to know more about the machine learning on the back-end of the platforms you visit most.
Today, we’re talking about the Facebook News Feed, and exploring some of the reasoning behind why users see what they see, when they see it. Once you know how the algorithm works, you’ll be better equipped to make sure your content gets seen.
What, exactly, is the News Feed?
It’s worth mentioning first that while this post talks about Facebook specifically, the same general principles discussed here apply to many other social platforms.
The News Feed is the first thing a user sees when they log into Facebook. Your News Feed is a summary of what’s happened recently among friends and the businesses you’ve liked or followed. Every action a friend or business takes is a potential News Feed story.
First, some history
In the early years of social media, Facebook and other networks showed posts in the News Feed chronologically, so they served up the most recent posts first. Reasonably, this chronological system best served the people and businesses that published often. Marketing teams discovered that they could manipulate this system by what we might affectionately call “spamming.” By creating a large volume of posts, they contributed to across-the-board, low-quality News Feeds for most users.
The first algorithm
In 2010, Facebook decided to fix the News Feed issue through an algorithm called EdgeRank. EdgeRank’s job was to predict how interesting each story would be to each user using the formula *affinity* x *weight* x *time*, and then filter each user’s News Feed to show only the top-ranked stories for that particular user.
Organic reach dropped drastically, and marketing teams had to once again figure out how to get their content seen by as many users as possible within the new system.
The algorithm is dead, long live the algorithm
By 2011, Facebook had ceased using the name EdgeRank, internally at least. By 2013, it announced a more refined algorithm with, at the time, over 100,000 weights in the model that produces the News Feed. The original three weights (affinity, weight, and time) were still present, but they were accompanied by other things that held equal importance. The algorithm gained a sophisticated new method, including categories and subcategories, to determining what was shown in any particular user’s News Feed. This current incarnation is consistently updated and refined — a recent update included a response to fake news, for instance.
Instagram joined the algorithm game in 2016, updating to a “here and popular” feed from a “here and now” feed. And in March 2017, despite the protestations of current users and former employees, Twitter commenced their own algorithm to similarly weight users’ Twitter Feeds.
How to navigate content in the current incarnation
In his talk on Facebook’s News Feed at F8 2016, Adam Mosseri, VP of Product Management of News Feed, walked listeners through four key factors of the Facebook algorithm.
Who posted it?
Facebook’s algorithm asks who posted the content and how often the potential viewer interacts with that publisher.
If a viewer interacts with your posts on a regular basis, they’re more likely to see subsequent posts. If that same viewer starts to find your content boring and stops interacting with it, they’ll see less and less of your content.
Content type
Facebook looks at what type of content it is and how often the potential viewer interacts with that type of content by liking it, commenting on it, or sharing similar content.
Pro-tip from our friends at Black Parasol Consulting: New content types — such as livestream or slideshows — always outperform old content types. New content types are out of the ordinary and, therefore, are eye-catching and engagement worthy. New content types are also experimental, and networks will want to put them in front of larger audiences to gain feedback.
Interactions
Posts with a higher number of interactions are inferred to be of more interest. A post that has many likes, comments, and shares will be surfaced more frequently than one without engagement.
When it was posted
Facebook acknowledges that recency can contribute to relevance (though they admit that it’s not the only contributor). Older content is devalued in lieu of more recent content.
The score, and how to use it
All of the weights for each piece of content are combined into a Relevancy Score, which is unique to each individual user on Facebook. The Relevancy Score is what determines which content pieces are shown closer to the top on each user’s News Feed.
It should be clarified that the News Feed is an ever-changing creature. You can stay updated on News Feed changes that might affect your marketing techniques by regularly checking in on Facebook’s News Feed FYI blog. Your first step could be catching up on 3 important updates to the algorithm from January 2017.
As an entrepreneur using Facebook or other social platforms to garner brand awareness and drive traffic to your store, you can experiment with the various factors that play into the Relevancy Scores of your potential customers. Try tinkering with a new (or new-to-you) content type, or encourage interaction with your posts by asking questions or asking viewers to comment. Think about your Facebook fans, from the types of posts they interact with to the different devices they might be using when they’re on Facebook. Then create and publish a variety of interesting content that will attract shares, comments, likes, and clicks.
Want more examples of ways to use the algorithm to your advantage on social media? Check out the ShopLaunch event presentation video recorded in San Francisco.
