How IG Works
Ranking explained
The algorithm explained
feat. @raghavsachar, @ankit_rihal
In a conversation with Raghav Sachar, Ankit Rihal from Meta discusses how we rank things here on Instagram - The "Algorithm” (or actually algorithms) and how we rank Stories, Feed, Reels and Explore.
Key takeaways
- There is no one Instagram algorithm. There are multiple, different algorithms for every part of the app.
- For Stories, we make predictions on things you might do with a Story such as tap on the Story to open it or reply to the author. To make these predictions, we use signals - information that we know about you and your interests. The most important signals for Stories are your viewing history, interaction history, and closeness to the author.
- For Feed, we make predictions based on how likely we think you are to take certain actions on a post such as like, comment, or share it. The most important signals are your activity, information about the post, information about the author, and how much you've interacted with the author.
- For Reels, we make predictions about each Reel to assess how interesting it might be to you. We look at things such as how likely you are to finish and share the Reel. The signals we use are your activity, information about the Reel, information about the author, and your interaction history with the author.
- For Explore, we make predictions on things such as how likely you are to like, save, and share a post. The signals we use are information about the post, your activity, information about the author, and your interaction history with the author.