To address long-standing curiosity and dispel misconceptions surrounding its algorithm, Meta-owned Instagram has published a blog post explaining how posts and stories recommendations work for its users.
Adam Mosseri, the Chief of Instagram, explains that each section of the app has its own algorithm, and the ranking factors vary accordingly. In the case of the Feed, the algorithm considers five key interactions: the likelihood of spending time on a post, commenting on it, liking it, sharing it, and tapping on the profile photo. The algorithm weighs these actions, giving a higher weight to actions that users are likelier to take.
Furthermore, Instagram emphasizes that each tool within the app, such as Feed, Stories, Reels, and Explore, has its own algorithm, tailored based on how people use those features. For instance, users tend to seek their closest friends’ stories, utilize Explore for content discovery, and find entertainment in Reels. Consequently, the ranking of content varies across these different sections, and Instagram has introduced features like Close Friends, Favorites, and Following to provide users with greater control over customizing their experiences.
Let’s take a closer look at how the feed ranking works on Instagram. The feed is a blend of content from the accounts users follow and posts from suggested accounts that Instagram believes might pique users’ interest. To determine what users might find interesting, Instagram considers several factors, including whom they have followed, liked, or engaged with recently. These factors, referred to as “signals,” encompass a wide range of information related to user activity, post details, and the person who posted it.
User activity plays a crucial role in shaping the feed. Actions such as liking, sharing, saving, or commenting on posts provide Instagram with insights into users’ interests. Additionally, information about the post itself, including its popularity and engagement metrics such as likes, comments, shares, and saves, contribute to the algorithm’s predictions. Moreover, details about the person who posted the content, such as the frequency of interactions with that individual over the past few weeks, help Instagram gauge the user’s potential interest in their posts.
Instagram stores all this information, creating a comprehensive set of signals. These signals encompass various aspects, including when the post was shared, the platform being used (phone or web), and users’ preferences such as the frequency of liking videos. By analyzing these signals, Instagram’s algorithm determines the ranking of content in users’ feeds, aiming to surface what they are most interested in.
For Stories, the algorithm ranks them based on factors such as the frequency of users viewing an account’s stories, engagement with those stories, and the user’s proximity to the author. By analyzing these signals, Instagram makes predictions about which stories users are more likely to find relevant and valuable, including the likelihood of tapping into a story, replying to an account, or moving on to the next one. These predictions determine the order in which stories are shown in the users’ stories tray.
In the case of Reels, the algorithm considers user activity, such as the reels they have liked, saved, reshared, commented on, and engaged with. Additionally, factors such as the user’s history of interacting with the person who posted the reel and information about the reel itself and the person who posted it are taken into account.
