YouTube's recommendation system accounts for about 60% of the total time people spend watching videos on the site.

YouTube uses a computer system called a neural network to choose which videos to show you. This system looks at over 80 billion signals every day. These signals include which videos you click, how long you watch them, and what you like or share.The process happens in two main steps. First, the computer looks at millions of videos and picks a few hundred that match your interests. It does this by comparing your habits to other people who watch similar things. This is called collaborative filtering.Second, the system ranks those few hundred videos to put the best ones at the top of your screen. In 2016, YouTube changed the system to focus on watch time instead of just clicks. This was done to stop people from making clickbait videos with misleading titles.The algorithm updates itself many times each day with new data. It learns from new trends and news stories happening around the world. It also looks at your location and language to make sure the videos are relevant to you.