The Death of Friendster: Why A Social Network Dies
Long before Facebook grew to popularity, Friendster was the boss of social networks— the real O.G., the veteran, the grand-daddy. At some point of my pre-teenage life, almost all of my friends had an account in the site, and polishing my Friendster profile was a status-defining moment. Hundreds of millions of users were recorded to have registered back in 2002, just a few months after it was created. There was no doubt that Friendster was at its ultimate high, even considered by many as the pioneer of a social media revolution back then.  At that time, Friendster seemed to be in a rollercoaster that only went up.
But then, something went very, very wrong. Users fled one by one; this posed a threat to the company. Friendster turned down a 30-million dollar buyout offer from Google in 2003, hoping for a rebirth. Since then, the site has been rebranded and promoted as a social gaming platform, its fingers still crossed.
By 2009, Friendster announced its death. Their strategy wasn’t working—more users left Friendster for the slightly newer Facebook. This catastrophic decline accelerated the site’s extinction.
This, naturally, led a lot of researchers to investigate on the matter. What went wrong? A group of researchers from the Swiss Federal Institute of Zurich proposes an autopsy that might contain an answer.
David Garcia and his colleagues assert that there are two key indicators of a social network’s health. Together, they have undergone a “digital autopsy” which they called a piece of “internet archaeology”. They have created a mathematical model on the nature of social network deaths.
The first factor at play is called “k-core”. To illustrate this term, imagine that you have five friends. If one were to leave the network, you would immediately notice, right? Since you have such a small pool of friends, you would most likely leave it too. But if you have, say, a thousand friends; one friend leaving would not trigger as big an effect on you. This is the foundation of “K-core”—it is the ratio between the latter and the former. Garcia’s research suggests that when the “k-core” is high, users leaving would create a domino effect, and other users would leave too. Graphically, a user exodus would look like this:
A high k-core distribution leading to a user exodus. Obtained from https://www.technologyreview.com/sites/default/files/images/Death%20of%20Friendster.png
Thus, this concept plays a huge role in determining a network’s vulnerability to user loss.
However, there must be something else that acts as the root cause of early quitters; this is what the researchers refer to as the cost-to-benefit ratio, the second factor. It states that if the time and effort exerted on the network outweighs the benefits incurred by the user, people are more likely to leave it and join a more beneficial network. In the context of Friendster, its cost-to-benefit ratio dropped drastically, evident in its various unsuccessful site redesigns and technical problems. That, added to a high k-core distribution, led to Friendster’s demise.
Facebook and other social networks should have learned a lot about this kind of problem, if they haven’t already.  Nowadays, it’s easier to imagine how privacy intrusions and money-grubbing schemes would interfere with its user traffic, more importantly if there is another emerging network competing for the people’s attention.
In 2009, Facebook is thought to have benefited from Friendster’s collapse. Many have speculated that Friendster’s past would soon become Facebook’s future.  However, it is likely that Facebook will also be the next victim to this set of circumstances.
Garcia, David, Pavlin Mavrodiev, and Frank Schweitzer. Social Resilience in Online Communities: The Autopsy of Friendster. Proceedings of the First ACM Conference on Online Social Networks, 25 Feb. 2013. Web. 10 Nov. 2015. <http://arxiv.org/abs/1302.6109>.
Buckley, Neil. "The Origins of Friendster." Just Friends? 2010. Web. 10 Nov. 2015. <http://social-networking.limewebs.com/friendster-history.htm>.
