By Kim Tong-hyung
Staff Reporter
For the users of Twitter (www.twitter.com), the latest global online craze, the number of ``followers'' you have appears to be as important as the size of your bank account in the real world.
However, a recent study questions whether the obsession with the Twitter follower count is reasonable, suggesting that the number of links alone would reveal very little about the influence of the user.
In ``Measuring User Influence in Twitter: The Million Follower Fallacy,'' a team of researchers led by Cha Mee-young of Germany's Max Planck Institute for Software Systems claimed that topological measures such as indegree, the number of people who follow a user, are not always a reliable reflection of user influence.
``Many Twitter users are obsessed with the number of followers. There are competitions going on to see who becomes the first to gather 1 million followers and there are even paid services promising to increase the number of followers for a user,'' Cha said.
``As Twitter and other social media services get increasing attention for their business influence, companies are spending massive amounts of money on viral marketing efforts. However, rather than merely targeting users with the most followers, companies should be smarter in deciding where to spend their money and place their advertisements, focusing more on the quality of indicators, which is represented better by measures such as re-tweets or mentions rather than the number of followers.''
Social media services like Twitter allow users to broadcast their real-time status to an indefinite amount of Internet users on computers and mobile devices. Combining the strengths of blogs and instant-messaging services, these so-called ``micro-blogging'' services are changing the way people consume and produce information, and also garnering interest for their immense business potential as marketing tools.
So, as a result, this had social media and online marketing professionals debating on how to measure the influence of a Twitter user, and the opinions varied depending on how ``influence'' is defined.
Cha's team collected a massive amount of data from Twitter _ covering more than 54.9 million user accounts, 1.96 billion links and 1.75 billion ``tweets,'' or Twitter postings.
The researchers narrowed down the data to focus on ``active users,'' who had more than 10 tweets and active screen names that could be ``re-tweeted,'' or replied, which left them with the accounts of about 6.2 million users.
To measure the influence of the 6 million-plus users, the researchers focused on how the entire set of Twitter users, excluding those with private accounts, interacted with these active users. The information was then broken down to a comparison of three measures of influence ― indegree, re-tweets and mentions.
To put it briefly, the researchers concluded that those with the largest number of followers were not necessarily the most influential ones. Twitter users with a high indegree were not always that much more influential in terms of spawning re-tweets or mentions.
The most influential users can hold significant influence over a variety of topics. However, influence is not gained spontaneously, but through concentrated effort, such as limiting tweets to a single topic.
``We believe that these findings provide new insights for viral marketing and suggest that topological measures such as indegree alone reveals very little about the influence of a user,'' the researchers wrote in the study.
Out of the 6 million-plus active users, the researchers picked the top-100 users in each of the three categories, which left them with 233 users if you excluded those who overlapped.
Although many of these accounts belonged to news organizations or celebrities, there was a visible mix of regular users as well. Among regular users, those who limited their tweets to single topics were often the more influential, social media gurus.
``We believe that online services like Twitter will give sociologists more freedom in pursuing data-driven research, enabling them to question more answers than what was possible from surveys or research from smaller sample sizes. Our study represent one of the first attempts in this step,'' Cha said.