Online, level 8
๐ฒ Why You See What You See
How does a video site decide what to show you next?
- How does a video site decide what to show you next? Most apps and websites that show videos, songs or posts depend on a recommendation algorithm. An algorithm is a set of steps that a computer follows to solve a problem. This kind tries to predict what you will want to see next, and it learns from what you do. It notices which videos you watch, how long you keep watching, what you click, what you like or share and what you skip. Then it uses that information to show you more of whatever held your attention before, one suggestion after another.
- Many of these sites earn money from advertisements, so the longer you stay, the more ads they are able to show you. That is one big reason they are designed to keep you watching. Autoplay starts the next video automatically before you have decided whether you even want it, so the site makes that choice instead of you. Some feeds have no bottom, so new posts keep loading as you scroll, and there is never a natural place to stop. This explains why a feed can feel endless, and why an entire hour can slip by when you only intended to watch for a few minutes.
- Because the algorithm keeps showing you more of what you already watch, your feed can gradually narrow until it displays mostly one kind of idea or opinion. That can make it seem as though everyone agrees, even when many people see things quite differently. Fortunately, you can stay in charge. Choose what to watch on purpose instead of letting autoplay choose for you, and decide on a time to stop before you begin. Take regular breaks, and search for other points of view on purpose. If your feed ever shows you something upsetting, stop watching and tell a trusted grown-up.
Try it at home: Look at a video feed together and guess why each video was picked, then talk about one way to find a different point of view.
Questions and answers
Why You See What You See
- ๐ค What is an algorithm? A set of steps a computer follows
- What does a recommendation algorithm try to do? Guess what you will want to see next
- ๐ Which of these does a recommendation algorithm learn from? How long you keep watching (It notices what you watch, how long you watch, and what you click, like, share or skip.)
- ๐ฐ Why are many video sites designed to keep you watching? The longer you stay, the more ads they show (Many of these sites earn money from advertisements.)
- Why can a feed with no bottom feel endless? New posts keep loading, so there is no natural stop
- ๐ If you only watch one kind of video, what can happen to your feed? It can narrow to mostly one kind of idea (The algorithm shows you more of what you already watch, so a feed can narrow.)