Adam Mosseri, the Head of Instagram, shared an insightful blog post where he sought to better explain how Instagram’s technology works, touching on the all-mighty Instagram algorithm, how Instagram ranks content, and “shadowbanning”.
Here are some highlights from the article.
In the article, “Shedding More Light on How Instagram Works”, the first thing that Adam wanted to clear up was this notion of “the algorithm”. The truth is Instagram doesn’t just have one, single algorithm, but a variety of algorithms, classifiers, and processes. These algorithms are set into place with the ultimate goal of personalizing your Instagram experience with content you care about, in turn keeping you on the application longer.
Ranking the Feed and Stories
Continuing on, Adam states that Instagram relies on thousands of “signals”, and the most important ones are: information about the post, information about the person who posted, the user’s activity, and the user’s history of interacting with someone.
From there, Instagram makes predictions on how likely you are to interact with a post. When it comes to the feed Instagram takes into account the following five (5) interactions: how likely you are to spend a few seconds on a post, commenting, liking, saving, and visiting the profile. Adam did add that the platform tries to avoid showing too many posts in a row from the same person, and used to avoid Stories that were re-shared from feed up until recently.
Ranking the Explore Page
We have noticed quite the evolution of the Explore page as Instagram has continued to refine its algorithms so it was great to learn more about how Instagram ranks Explore. The process is similar to the feed and stories, with Instagram looking at the “signals” of liking, saving, and sharing content you’ve previously seen. Instagram also applies recommendations guidelines to the Explore page in order to help people avoid potentially upsetting or sensitive posts.
Reels are set up in Instagram with the goal of keeping the user entertained. In similar suite, Instagram ranks reels based on predictions on you watching the whole reel, liking it, commenting, or even visiting the audio page associated with it.
A topic that was addressed very diplomatically, Adam addressed the accusatory claims of Instagram shadowbanning people or accounts, admitting that Instagram could do a better job at being clear about why content is taken down and is developing better in-app notifications as a result.
Influencing What You See
Adam concluded the article in providing some insight on how you can improve your Instagram experience by simply interacting with the profiles you enjoy most, picking your “Close Friends” for stories, muting people you are not interested in, and marking recommended posts as “Not Interested.”
This article was just a first of a series of articles Adam plans to publish to be ‘be proactive about explaining our work across all three areas from here on out. Stay tuned.’