How do social media algorithms work?
You don’t get far on the internet without the word algorithm popping up – and social media platforms are no exception. Whether you realize it or not, these algorithms are constantly working in the background. Here’s how they shape the content you see and why it matters.
When we talk about social media algorithms, we are talking about computational systems that decide what recommendations social media platforms offer you in your feeds.
You may be familiar with the Explore page on Instagram, the For You page on X, and the TikTok For You Page (FYP) – feeds that the social media platform automatically generates for you. For each of these pages, the social media platform uses machine learning (ML) systems to predict what kind of posts you would like to see.
You will find that the more engaged you are in posting on your social media platforms, the better referrals you will receive. This is because these ML systems can learn from your data. The more you interact with the content, the better the algorithms predict what you would like to see.
In addition to content recommendations, social media platforms can suggest accounts you might want to follow (e.g. Facebook’s People You May Know feature, which you can easily disable), popular topics you might be interested in, inappropriate content or spam you’d rather not see , must-see viral content, events in your area and much more.
Any recommendation or customized experience like personalized ads is possible because an algorithm is involved.
Social media platforms collect data from your user profile, posts you’ve interacted with, currently popular posts, and sponsored posts to offer you recommendations and a personalized experience.
In other words, social media platforms use the following to offer you recommendations:
Your interaction data |
This includes your likes, reactions, upvotes, comments, replies, shares, retweets, click-through rate (CTR), how long you spend watching a video or other type of content, swiping behavior, story interactions (taps forward and back), and the accounts you interact with the most. |
Your content preferences |
An algorithm may study your interactions and determine that you primarily work with one format, length, or category of content (eg DIY YouTube Shorts). In that case, you will get further recommendations along these lines. |
Your historical data |
This includes factors such as accounts, posts or topics you’ve previously engaged with, your past searches, accounts you unfollow or ignore. |
Social signals |
Social media platforms track emerging trends and topics, as well as posts that have a rapid increase in user interactions. These platforms then recommend these trends and posts to a wider audience. |
Your profile data |
Information associated with your profile, including age, demographics, location and device, may be used by algorithms to target specific content to you. |
Your network |
Platforms may recommend content based on what posts your friends interact with, or accounts based on connections you have with other users (eg LinkedIn profile suggestions). |
Third Party Tracking |
Social media platforms typically use tracking cookies to influence personalized ads and content recommendations. Likewise, platforms owned by the same parent company can share data (especially if your profiles are linked), so if you often work with certain content on Instagram, similar content can be recommended on Facebook. |
Your feedback |
The Platforms allow you to provide feedback on the referrals offered to influence future referrals. |
By combining all of this data, social media algorithms can predict which posts you’re likely to be interested in or which posts you don’t want to see. Combined with their community guidelines, platforms then provide content to users.
Do social media companies really control what posts you see?
With all these sophisticated systems in play, it can be easy to claim that social media companies are controlling the content you see. But it’s not that black and white.
At their core, most social media companies rely on machine learning algorithms to predict what content you will like. The machine learning part is key here – the algorithm studies your data and evolves itself to decide what posts to show you.
However, these companies are not completely hands off. They actually have more of an effect on what you see than you might think at first.
First, each platform has its own set of rules that determine what is allowed. For example, Facebook filters certain types of content from your news feed, such as NSFW posts or content from groups known to promote violence. Similarly, X (formerly Twitter) has been known to block, flag, or demote tweets that violate its policies, including tweets containing misleading information, deception, or intimidation.
Social media companies may also downgrade or block certain posts entirely, especially for sensitive or controversial material. We have seen examples on various platforms. In 2020, Facebook and Twitter (before rebranding to X) limited the visibility of posts about the Hunter Biden laptop story, deeming the material misleading or hacked.
X has seen perhaps the most dramatic shift in content moderation. Under Elon Musk’s leadership, X changed its style of content moderation, so Donald Trump’s account – originally banned – was reinstated. Even when Elon Musk took over, X briefly blocked links to its competitors. These actions show that companies have the freedom to manage certain types of content – going beyond algorithms.
Some companies also decentralize or delegate certain moderation decisions to individuals or communities. For example, on Reddit, the moderators of each subreddit (Reddit community) can create and enforce their own rules. So you’ll find some subreddits with very loose guidelines and others with stricter rules. In this case, Reddit as a company does not directly control what you see beyond certain guidelines for all subreddits.
So while it’s an exaggeration to say that social media companies fully control everything you see, they certainly shape your experience through community/content guidelines and decisions about what to ban or take down. These influences vary across platforms and may change over time in response to leadership changes, societal pressures, or other factors.
In the end, one constant remains – these platforms want to keep you connected. The longer you scroll, the more ads you see and the more revenue the company generates. Both algorithms and human efforts serve this purpose. While they occasionally fail, these systems generally succeed in keeping you scrolling.