Boosting Facebook Post Visibility
What exactly gets stories put into a news feed on Facebook? This is a particularly important question for people running pages or communities and/or advertising on Facebook. Last week we listened to a guest lecture from a Facebook employee about what exactly goes into Facebook’s algorithm for sorting stories, but I’d like to analyze a few facets of this extremely complex process more specifically in the context of this class.
One of the most obvious factors is likes and comments on a post – in particular likes and comments by friends. While the the reasoning behind ranking based on these factors is fairly intuitive, we also have a more formal way to consider it: strong triadic closure. Suppose your good friend often likes and comments on posts from another source. That makes you more likely to want to make a tie with that source yourself, so if Facebook can push a link to it onto your news feed, you’ll be much more likely to click that link. This is a big deal for Facebook. More clicking is more engagement, which is more ad revenue.
Notably, the “source” mentioned above doesn’t necessarily have to be another person. It could be a page or community as well. This is significant because Facebook also favors factors like tagging other pages in a post. This is the same principle as with liking and commenting. Suppose you follow a page very closely. Any pages that that page is closely related to (denoted in this greatly simplified analysis via direct tags/mentions) is probably a page you will have some interest in as well. As with liking and commenting, the principle of strong triadic closure means that you are more likely to make a tie with that page, which, once again, means more ad revenue for Facebook.
Of course, even after all this analysis, there are bound to be some interesting exceptions which require a bit more work to filter out. For instance, consider posts commonly referred to as “bait posts” which use a picture split into four sections that say “Like for….Share for….Comment for….Ignore for….” These pictures often have a very high level of engagement because of how they encourage users to interact with the post. By a model based on the factors listed above, such a post should be pushed very high in a news feed. However, these posts are rarely quality content. To this end Facebook actually employs some image recognition to identify this format of picture post and push it farther down in your news feed. Once again this serves to help increase engagement by removing posts that won’t help make long term connections/sustained interactions which are a better long term bonus for Facebook’s ad revenue.
Of course, it’s not all about revenue. There is certainly something to be said in favor of Facebook analyzing and sorting posts like this. Their sorting algorithm keeps our news feed clean of spam and full of relevant content. Imagine if we just got posts based on the time they were posted. Pages could easily abuse this, and our news feed would just be people spamming “Like for cats, Share for dogs, Ignore if you like Justin Bieber” posts. In this way (among others), Facebook’s algorithm is actually very beneficial to the end user.
It just also happens to make them more money in the mean time.
Source for elements of Facebook’s algorithm (this source sites updates from Facebook itself):
http://socialmediaweek.org/blog/2014/11/inside-facebook-news-feed-list-algorithm-factors/