How Facebook and Instagram Algorithms Work

In 2014, Mark Zuckerberg, CEO of Facebook, said that the main focus of the network is to become a "personalized newspaper for each person." Therefore, Facebook is constantly learning about the behavior of its users, seeking to empower them to come up with the definition of what actions define whether content is interesting or not. The goal is to keep visitors for longer on the social network, interacting with as many publications as possible.

The “like” button, implemented in 2007 - three years after the launch of the social network - was the first significant change in the algorithm aimed at mapping the level of relevance of content to users. The strategy developed until, in 2016, the reactions were launched, other buttons that aim to go beyond the approval of a publication: they seek to measure the engagement it generates, receiving positive and negative feedback.

Two of the aspects that guide this analysis are:

1. Origin of the publication

By 2015, after concluding that its users were very concerned about losing updates from their closest friends through news feed information, Facebook's algorithm allowed them to choose which pages or people they'd like to see the updates first. Before that there were only filters for what people did not want to see, with the option to hide posts and unfollow. Relevance of content and engagement with other users or pages was measured by tastings, shares and comments. At this point, the News Feed Preferences were launched to empower them in this choice.

2. Relevance of the publication

Based on the history of interactions, the Facebook algorithm can predict when a publication will be tanned, clicked, commented or marked as spam. The combination of variables creates the relevance score, which organizes the order in which the posts will appear for each user. The first that appears will therefore be the most capable of engaging at that moment. This calculation is always fed by positive and negative feedbacks. The same logic is used with Facebook Ads. If ads are expected to improve users' social network experience, the relevance score contributes to the identification of the target audience for companies that use that attraction strategy.

The Instagram Feed Algorithm

By the year 2014, Instagram had no algorithm that individually monitors users' activities for suggestion of photos in its "explore" tab. There, the most popular publications were shown among all users, without any specific guidance. In addition, the timeline showed the publications in chronological order, regardless of their relevancy. However, when acquired by Facebook, Instagram started to focus on improving the user experience, since it was estimated that they did not see 70% of the photos of their feeds in their visits to the social network. From the new algorithm, the order of publications in the Instagram feed is defined by the following factors:

- Number of likes and comments to measure engagement;

- Level of user interaction of the post source;

- Publication time, so the social network does not lose the chronological characteristic of the visualization;

- To whom the user sends direct messages and what type of content shared with these people.