Now coming back to the real question, what is meta algorithm and how does it work? In simple words, this is the system that sorts through millions of posts every second and decides what will appear first on your screen. Meta has not one but many algorithms. Different algorithms are used to run Facebook and Instagram. Similarly, separately for posts, reels, explore pages and threads. But the purpose of all these is the same. Showing you only the content you’re most likely to see.
How does meta algorithm work?
Meta’s algorithm is not like magic, but it is a very fast and clever AI based ‘recommendation machine’. It keeps a close eye on every click, every like, every share, every save and every second of watch time. Learns from it and then on that basis decides which post will be shown first on your screen next time.
The help page of the Facebook.com website explains how Meta distributes content. Meta first collects all the posts. Suppose you opened Facebook at 8 in the morning. As soon as the app opens, Meta first lists all the posts that can be shown to you. This list includes-
- your friends’ posts
- Pages you have followed
- Groups you are in
- sponsored post
- Recommended posts suggested by AI
- reels
- viral video
This is called Meta Inventory. Think of it like a hotel prepares its menu. Meta also prepares the entire menu of posts first.
investigation
After creating the inventory, Meta checks each post. This is where the real game of algorithms begins. He sees thousands of signals associated with every post. As-
- Who posted?
- How close is that person to the user?
- How old is the post?
- How many people liked it?
- How many people shared?
- How many people commented?
Meta also sees whether you are using Facebook or Instagram on mobile or laptop. Is it day or night and what is your internet speed? It also keeps an eye on what kind of post you just saw? That means every small and big thing is a signal for the algorithm to guess.
AI guesses
Meta does not try to read your mind through algorithms. He just analyzes your old habits and makes guesses. You like cricket. You watch cricket reels every day. Like and share cricket posts. Watching match videos is also one of your habits. Now if someone has put a new reel of Virat Kohli or Shubman Gill, then the meta will have an idea in advance that perhaps you will also watch it completely. Similarly, if you read stock market news, Meta will add similar content to your feed. If the AI thinks you’ll like a reel, it will send it to your feed, even if you’ve never followed that creator.
AI makes predictions for every post
AI makes predictions for every post. It considers whether you will like, share, or comment on a particular post. Will watch the full video or save the post. Meta itself does not know the answers to all these questions. But AI extracts possibilities.
Relevance Score
Based on these answers, Meta gives a score to each post. This is called Relevance Score. The post which has higher score will appear higher in your feed. Whose score is lower, it will go down or you may not see it at all. He gives score to the post on the basis of some parameters. For example, if the probability of watching any of your videos is 90 percent or the chance of liking is 40 percent, then the score of such a post will increase. That means Meta asks every post whether it is useful for a particular person? If the answer is yes, it is sent up.
Discover more from News Link360
Subscribe to get the latest posts sent to your email.
