How Social Algorithms Work

Do you remember when all the content on your social media feed showed up in chronological order? If you remember that, you were on social media at a time when it was much easier to run a content marketing campaign on there.

Social media algorithms have changed quite a bit since then. They’ve changed for the better in some ways and made it harder in others. But that’s the nature of social media, even at a fundamental level. They’re always changing, and the onus will always be on marketers to keep an eye on trends and adjust accordingly.

So what do you need to know about the most popular social media sites and their algorithms? Let’s take a look.

Understanding the social media network itself

There’s no better place to start than with Facebook. With over 2 billion active monthly users, Facebook is the most popular social media site in the world. Not only that – Facebook also owns Instagram and WhatsApp. While the latter is technically an instant messaging app, Instagram is the fourth most popular social media app in the world with 1.386 billion monthly active users.

Ironically, Statista considers WhatsApp a form of social media, and their numbers report it has 22 billion users. This means that Facebook owns three of the four most used social networks in the world. Its only real competition is YouTube (ranked second, 2.291 billion active monthly users).

So what does this mean for algorithms?

Essentially, Facebook has created its own “network of networks” by buying up two major competitors. None of those platforms are silos. They have independent engineers and designers, but they develop according to the goals of their parent company – Facebook. So you’ll find overlaps there that you might not find on, say, Twitter, YouTube, or LinkedIn.

So how does the Facebook algorithm work, and how does it affect other social media platforms?

Separate personal and platform goals

An algorithm is designed around the goals of the company using it. As complicated as the coding is, Facebook’s algorithm is designed with a simple purpose in mind – use content curation to boost engagement and audience retention.

All social media platforms are obsessed with audience retention. Why? Because that’s where advertising revenue comes from.

Facebook made 28.5 billion dollars in the second quarter of 2021 off advertising revenue. YouTube made 19.77 billion dollars throughout 2020. By comparison, Twitter only brought in 1.053 billion dollars, but much of that down to Twitter not building its business model around ad spend. Still, all three platforms saw a significant revenue increase in ad revenue – and the number is only growing.

How do social media algorithms retain audiences

So, back to the question of curation: how does it work? More importantly, why does it work?

Well, before an algorithm can curate an individual’s feed, it needs to collect data on their behavior. Algorithms are the main drivers behind machine learning. This data is based on certain interactions: who someone follows, the type of content they frequently engage with, even the type of content the people they follow regularly engage with.

That information can give an algorithm an “idea” of what someone is likely to engage with. Now, this isn’t done with 100% certainty – even the most sophisticated code sequences don’t have a 100% success rate. But that’s why machine learning is so important.

An algorithm might determine that a post has a reasonable chance of engagement if it’s seen by a certain user. In that case, it might add the post to the user’s feed. If the user engages, the algorithm makes note of it. If the person simply scrolls by, the algorithm might register that there has been an impression (someone has seen the post) but no actual engagement.

That information is just as valuable. The data is collected, stored, and used in future curation decisions. Social media algorithms monitor, study and adapt to the behaviors of audiences. So what does this mean for marketers?

The algorithms’ goals are not your goals

Content curation on social media is about making a user’s newsfeed as relevant to them as possible. Take a look at your social media profile. How many profiles do you follow? How many of those profiles end up on your feed? If you’re regularly engaging with the ones you like, chances are it’s less than 10%.

But what’s relevant to you as a user isn’t always useful to you as a marketer. That’s because your goal isn’t to curate your audience’s feed – it’s to appear prominently in it. Social media algorithms are happy to facilitate that if you can satisfy its goals.

This means that – if you want a site’s algorithm to work in your favor – you need a way to align your goals. You want high engagement and visibility with your audience – algorithms want that too, but it’s on you to prove your content is worth the risk.

How to work with the algorithm

The curator in an art gallery may be interested in the story behind your art, but their main objective is to create a gallery around the viewing audience. Algorithms are the same. You may measure the relevance of a post on things like branding, clarity of message, and what you know your audience wants. While every algorithm is different, they can only conceptualize those values mathematically.

If you want to keep showing up on people’s feeds, you need to ensure that your posts get regular engagement each time. Machine learning is a self-fulfilling prophecy: the more engagement you get, the more visible the algorithm makes you, which leads to more engagement, etc.

That’s not just a one-off either. You’ll need to consistently post content that gets decent engagement. In many ways, getting organic reach through social media algorithms starts with getting your existing audience to engage more.

Likes, retweets, comments, shares, bookmarks, clicks on links – all of these are engagement metrics. The biggest social media sites in the world all look out for them. You can’t train an algorithm to work for you, but you can teach it to help you. It starts with ensuring that your goals are measurable by the algorithm’s standards.

From there, it’s a matter of repetition, adaptation, and keeping an eye on future large-scale changes.