Skip to main content

Do you know your algorithms?

From Netflix to Instagram, the coders behind them have carefully predicted what you're going to hit 'like' on. Meet the computer codes that are shaping your life.

First up: what’s an algorithm? For those of us who don’t count themselves as coders (yet…), an algorithm is a formula that a computer can follow to perform certain actions, such as suggesting the next video that you’ll enjoy. It learns from how you behave on an app or website in order to show you the most relevant things possible. For instance, if you leave a comment on a YouTube video of goats climbing onto things, there’s a recipe or algorithm that tells the inner workings of YouTube that you probably love goats, or animals getting into mischief, and so they should show you more of those.  As you’re about to find out, these formulas make the online world go round…


In an ideal world, Netflix wants everything you see on that first page to be something you will fall in love with, whether that’s an American sitcom with a corrupt cop or a reality TV show featuring love-based drama. So how do they do it?

The first way is obvious – they collect data on what you’ve already watched, and suggest more of the same. Drag Race fan? Have a load more RuPaul.

The second is the reason you see super-specific categories, like that ‘American sitcom with a corrupt cop’ one. Somewhere right now, there’s a gang of Netflix employees watching every moment of every show, and tagging it, with things like whether it’s set in New York, features a strong female lead, is LGBTQ+ friendly or stars Zendaya (can we say “dream job”?). That data is used to create categories of similar shows.

The third stage is where the machine-learning algorithms come in. These work out how much weight to give to each data point produced. For instance, if you binged something over one weekend, that programme’s going to be given more importance than if you dipped in and out over a few months.

By combining these three things, you’re matched to thousands of ‘taste communities’, who share similar likes and dislikes to you. And that’s how Netflix hopes to help you find your next binge-worthy boxset.


While your Instagram feed is full of people you’ve chosen to follow, the order in which you see them is dictated by Instagram, not by the order they were posted. Kind of.

Like most platforms, Instagram has toyed with showing you the posts that are popular, rather than those that are recent. For a while, the more likes and comments a post had, the more likely you were to see it.

But having received feedback that said this was… well, just really annoying, they’ve adapted the algorithm, so that now it’s a top secret combo of the new and the popular that dictates the order. Love it or hate it, it’s a good example of a big company listening to their users and switching things up as a result. Power to the people!


How often do you find yourself tapping on a Recommended Video and accidentally spending four hours down a YouTube wormhole? All. The. Time? You’re not alone, and YouTube has written an awful lot of code to make it happen.

The sheer quantity of video uploaded makes it tricky to make personalised recommendations, and unlike Netflix, there’s no army of YouTube employees carefully watching and tagging the contents of each upload.

YouTube has two things to examine to try and recommend stuff you’ll enjoy: the metadata the uploader adds – such as the title, tags (‘beauty haul’, ‘RPG gameplay’ etc) and description – and the data collected from you about how much of a video you watch, what else you watch, whether you give it a thumbs-up or comment etc.  They’re super-secretive about exactly how important each piece of data is, but the idea is that the good videos will naturally appear, and anyone trying to game the system won’t be rewarded with a Recommended Videos spot.

So next time you’re on YouTube, be aware that if you fall down a video wormhole of  kittens meeting puppies for the first time,  you’ll be seeing a lot of cute animal videos in your future.  In fact, no matter what platform you’re on, it’s worth remembering that every minute you watch and every ‘like’ you award is being carefully monitored and added to what the algorithm knows about you.

And while Netflix, YouTube and Instagram make it obvious that they’re using what they know about you, you’ll actually find that same technology all over the web. Personalisation is everywhere – because companies know that you’re more likely to part with your hard earned cash if you immediately see something you love.  In fact, there’s every chance that when you next shop on ASOS, you’re seeing a totally different homepage to your mate, based on what ASOS already knows about your love (or hatred) of leopard print and chunky boots.

Algorithms: they know more than you think!



the MEI logo.
the Casio logo.
the Institute and Faculty of Actuaries logo.

More information


Resources and tools designed to help you guide your children through GCSE's and A-Levels.


Resources, tools and talking points designed to help you improve STEM subject participation and to engage more girls in the classroom.


We are always looking for new partners to help our campaign. Please get in touch!