Hidden Networks Matter Most on Twitter

Click on the title to this post to reach an interesting paper at First Monday, Volume 14, no. 1, 5 Jan., 2009, "Social Networks that Matter: Twitter under the microscope," by Bernardo A. Huberman, Daniel M. Romero and Fang Wu. First Monday is a peer-reviewed journal published exclusively on the Internet, sponsored by the University of Illinois at Chicago.

The researchers looked at "... a total of 309,740 [Twitter] users, who on average posted 255 posts, had 85 followers, and followed 80 other users." They wanted to compare the difference between followers and followees compared to real friends, whom they defined as people with whom the users shared a direct message twice or more times. On Twitter, users can either post a general comment that can be read by all who follow that user, or potentially, any Twitter user who checks the "all users" version. Or a user can send a direct message to a single user by addressing that user directly by Twitter name.

As a Twitter user, I will say that I do use direct messages with real friends, people I feel I know better somehow. And I will say things in direct messages that I would not say in the general Twitter stream. Thus, I agree, that the researchers are looking at a real distinction between run-of-the-mill followers, who may be people (or even animals or made up characters! There are horses, dogs, cats, Darth Vader, Ceiling Cat, Basement Cat and more...) with whom I have very little real connection, compared to closer friends.

The researchers actually work through many more pieces of data than I discuss in this blog post, but I was struck most by their final conclusion. They draw a picture of the network showing all links between users. Then they erase the lines between all users who have not shared two or more direct messages. They are revealing the hidden network within the outwardly apparent network of Twitter. Their conclusion is that there are two very different networks in Twitter and other social media. There is a dense network made up of very light social contacts between followers. And there is a much sparser network made up of much more enduring social contacts between "friends" who share two or more direct messages. They may not actually qualify as friends in the real world sense, but they do have more social contact in common, more mutuality than the general follower-followee.

A very interesting analysis. Worth considering -- whether you think of Twitter or Facebook as a marketing tool or a social media phenomenon, there are two different levels of contact there.