The Art of Algorithms: Breaking Down the Algorithms Used on Social Media Platforms
The Art of Algorithms: Breaking Down the Algorithms Used on Social Media Platforms
For years, the algorithms of different social media platforms were shrouded in mystery. It has only been relatively recently that popular social media apps have started divulging some information on how the complex ways online media pushes content for users.
While big platforms still like to keep some trade secrets when it comes to how their sites work, it’s possible to read published research to understand how algorithms work and target people. This creates a better social media experience for the average user and a way for companies to understand how their content is spread.
Facebook maintains that there is no single algorithm they use, but instead, there are “multiple layers of machine learning models and rankings.” Rather than showing posts in order of how recently they were posted and leaving it at that, Facebook predicts which posts users would like to interact with the most.
The Facebook algorithms evaluate every post, score them, and then arrange it in descending order of interest for the app user. Facebook is quoted saying that their algorithms are designed to be “most valuable and meaningful to an individual over the long term.”
The Facebook algorithm focuses on four main ranking signals: recency, popularity, content type, and relationship. Recency refers to the fact that newer posts are always shown first. This follows the fact that social media users prefer to see posts in chronological order or something similar. By popularity, the app refers to how much engagement each post gets, including likes, comments, and shares. Facebook will also purposefully show you types of content it notices you interact more with (such as pictures and videos versus text posts and articles).
The Facebook algorithm also notices what accounts a user typically interacts with and will show their content more than others. Facebook altered its algorithms in 2018 to focus on “meaningful social Interactions”; however, internal research found that engagement with posts doesn’t actually mean that a user wants to see more of something.
Twitter has shared some (but not all) information regarding how they rank content. The main things the Twitter algorithm takes into account are recency, relevance, engagement, content that is rich or valuable, and even the Twitter user’s location.
The Twitter timeline consists of three main sections: ranked tweets that appear on top of the timeline are posts that Twitter thinks are especially relevant to the user. Then there is the “In case you missed it” section, which shows tweets a user may be interested in but might have missed seeing them on the timeline when they were posted. And finally, all the rest of the tweets on a user’s timeline show up in reverse-chronological order.
TikTok
TikTok has also published a blog explaining how their algorithm works. Like other social media platforms, it takes many things into account with their algorithm including hashtags, songs, and even the kind of device the user has. When a video is first published on the platform, it is shown to a small number of users that may or may not already follow the creator. TikTok continues to show the video to people with similar interests until the audience grows in numbers, and the video goes “viral”.
Videos are shown to users that the algorithm thinks they are most likely to interact with, either by watching it in full, liking and commenting, or sending the video to another user. The algorithm often shows users videos with the same sounds and songs in the background, helping to make new auditory memes and launch the careers of new musical artists.
According to Wired, the TikTok algorithm is based on strong and weak signals from the user. Some examples include whether [a user] watched a video to the end, whether they shared it, and if they followed the creator who uploaded it afterward. Other things the algorithm takes into account that users may not think about are the type of device being used, language preferences, and whether they’re located in the same country as the creator of the video.
The TikTok algorithm is also special in that it considers negative feedback rather than reading all engagement as good engagement. Examples of this are if a user clicked “Not Interested” on a video, or if they chose to hide a specific sound or account.
Understanding how different social media algorithms target and read accounts creates responsible social media users who can make the algorithm work for them. It can help companies find their targeted audience.Comprehending social media algorithms brings power to the user experience as well as making it more enjoyable and intentioned. This is why social media apps should continue to be transparent about how their algorithms work, and people that use social media should constantly educate themselves on what data is being used to show them posts on the internet.