Instagram and Its Experiment in Reducing Network Effects
Link:
One of the critical features for measuring the success of content on Instagram is how many likes a post receives. The number of likes determines how many users will see the post as their algorithm takes the number of likes in account to determine the most relevant media to display to users. However, there is a flaw within this structure of prioritizing media.
Network effects occur when individuals are influenced by other nodes within the network to engage in a particular behavior. Before recent changes, Instagram displayed how many users liked a specific post and even displayed the usernames of the closest friends of a user which also liked to post. Due to conformity principles and the power of numbers users faced two external influences to like various content. First, the power of strong links can influence the individual to believe content is positive and they should like the post. This power of strong links comes into play when the user sees that their close friend has liked a post. The second influence is the power of numbers. Accounts who had massive amounts of followers would naturally get more baseline likes on their post than an account with around a hundred followers, as there is a portion of the audience with a very low satisfaction threshold and thus will almost always like a post. When a user sees thousands of other people are engaging with and liking a post, they are also more likely to engage with and like the post in belief that if lots of other people find this content as good, I most likely will as well. This relates directly to the “rich get richer” phenomena which we discussed in class as the users with the larger amounts of followers get greater numbers of likes and thus get higher engagement rates due to the influence of crowds and strong ties on users’ behavior. The users with more followers get more chances at getting more likes and once they have a large number of likes amassed, many people will like the post no matter the quality of the content of the post.
Thus, Instagram has hidden the number of likes from the view of users so as to make users’ decisions truer to the quality of the user’s media. This is a good step forward to eliminating network effects. However, as I previously mentioned, there is still the influence of strong ties on people’s engagement rates; with users being more likely to believe that content is good if their close friends think it is good since Instagram still displays close friends of the user which have also liked each post.
Although, by removing the like count from these posts, Instagram will have successfully removed the influence of crowds caused by like count, dampening the rich get richer effect. Since users can no longer see how many other users have liked a post, they are more likely to make personal judgments to whether the content is worthy of a like or not. However, Instagram should be careful as, in displaying the name of the closest user who has also liked the content, the rich get richer effect may still be at play. Instagram knows which users each member is closest to due to their embeddedness within social circles and their personal engagement rates with other users’ content. Accounts who have larger amounts of followers are more likely to have their content seen by a user’s close friends, which means such content is also more likely to be liked by close friends. Thus, the more followers a user has, the more likely their posts will receive more likes, as users face a higher probability of having close friends who also follow larger accounts. Thus, it is more likely that a close friend likes will be displayed to another user on a large account with lots of followers which will influence their engagement behavior. Additionally, an account with a lot of followers will display other users’ names that are increasingly close to other users as the post gains more and more likes. Thus, by displaying users’ names, we see the effects of strong ties and the “rich get richer” effect still at play, though more indirectly than through the display of the number of users who liked the post.