How Do Social Media Algorithms Work?

You’ve seen it happen – a post suddenly appears everywhere for a week, and then it seems to disappear just as quickly as it arrived. Or, you compare your feed with a friend who follows the same accounts as you, only to realise you’re seeing completely different things on your “for you” pages.

Algorithms aren’t new; they’ve been used to sort and process information long before Instagram or TikTok existed. What’s changed is the scale and the way social platforms now use them to determine which posts we see, who sees them and when.

How Do Social Media Platforms Use Algorithms?

An algorithm is basically a set of instructions used to make a decision. Think of it like a recipe telling you what to do with an egg and some flour. On a social media platform, the decision is a little more complicated because with so much being posted, the system has to work out what should be shown to you and in what order.

Rather than just displaying every post as it’s published, platforms have to sort through a huge amount of content and rank it based on different signals. Feeds, recommendations and search can also rely on different systems, even within the same app, so what appears in one part of the platform won’t necessarily be selected in the same way somewhere else.

How Does It Decide What You See?

Platforms can build up a picture of what you’re interested in through a long list of signals, which includes what you watch, how long you watch it for, what you like, share and comment on, who you follow and what you search for.

Before platforms started recommending posts based on individual behaviour, feeds could use an algorithm to arrange posts in chronological order. The system still had to follow a set of rules to decide what appeared first, even if that rule was just “show the newest post first”.

YouTube says its recommendation system uses signals including watch history, search history, subscriptions, likes and dislikes to help predict what someone might want to watch next. It also says different parts of the platform rely on different signals – the videos you see on the homepage aren’t ranked in exactly the same way as the recommendations you get while watching a video.

None of these signals necessarily tell the platform much on its own, but when they’re combined, they can help predict what you’re likely to engage with next. It’s basically a prediction rather than judgement, with the system using your previous behaviour to make an educated guess about what might interest you now.

That’s also why your feed can look completely different from your friend’s, even if you follow exactly the same accounts. The platform is using your individual behaviour to work out what you’re more likely to find relevant, rather than simply showing everyone the same posts in the same order.

How Does Something Go Viral?

When a post gets a strong response from the people who initially see it, the platform has a reason to show it to a bigger audience. If that next group responds in a similar way, the process can continue – more views and interactions give the system more information about who else might be interested.

That’s how an account with a few hundred followers can suddenly end up in the feeds of millions of people. There isn’t exactly a formula that can suddenly make this happen; going viral depends on how people respond to a post and how the platform reacts to that response.

Why Doesn’t The Algorithm Always Get It Right?

It’s predicting rather than reading minds. Watch one video about sourdough starters and now all of a sudden your entire feed seems to be convinced that you’ve forgotten about every other interest you’ve ever had.

It can also explain that familiar feeling that “the algorithm has changed”, when usually, it actually hasn’t changed. The platform could just be responding to something you watched, searched for or interacted with recently, thus deciding that you’re interested in seeing more of it.

Why Your Feed Feels So Personal

There isn’t one single algorithm deciding what everyone sees. Each platform has its own systems, and those systems are constantly weighing up what you’re watching, liking, searching for and scrolling past.

That’s why your feed can feel so personal and why something you’ve never heard of can suddenly be everywhere. The algorithm isn’t just sitting there deciding whether a post is good or bad; it’s trying to work out whether you’ll care about it. Sometimes it gets that right, but sometimes, after one too many sourdough videos, it definitely doesn’t.