Algorithm explains: This is how the Instagram feed works
Two days ago, Instagram gave some journalists insights into how their app works. Because already a year and a half ago, the company said goodbye to the chronological feed and introduced an algorithm. This caused a strong headwind, but, according to Instagram, users previously saw only about 50 percent of their friends’ posts and about 70 percent of all uploaded posts.
With the custom algorithm, that’s a whole different story. The 800 million Instagram users would now get rid of about 90 percent of their friends’ posts in the feed and generally spend more time in the app. How exactly this algorithm works and what parameters it takes into account has been explained by the company to a group of journalists.
So Instagram relies on machine learning based on past user behavior to create an individual feed for everyone. Even following exactly the same accounts as someone else, you would get your own personalized feed based on how you interact with those accounts. However, the algorithm can basically be broken down into three main factors:
On the one hand, it is based on past interests, so that similar content has a higher priority. Of course, then time also plays a role, so how recently a contribution was shared. The third ingredient is the relationship aspect, which considers how often you interact with a person or account. Secondary factors that the algorithm also accommodates, how often you visit Instagram, how long you stay in the app, and how many people you follow.
In addition, the Instagram team answered some of the most common questions about the feed. Thus, the company can not imagine at this time to reintroduce the chronological order. Any conspiracy theories suggests Instagram. For example, you would not hide posts from users or favor video and photo formats differently. The feed would also not favor users who use the app’s story or live feature.
Nice that Instagram is so open and gives us an insight into the algorithm. Of course, you always have to keep in mind that a company can tell us a lot. In this case, however, it is obviously trying to educate the users.