The "Cause of death" of Friendster

Absrtact: After the collapse of the social network Friendster, computer scientists have conducted a digital autopsy to try to find out the cause of death of the once-beautiful social networking giant. Friendster's Google search volume Friendster is a social network founded in 2002, more than

After the Friendster crash, computer scientists have conducted a digital "autopsy" to try to pinpoint the "cause of death" of the once-beautiful social networking giant.

Friendster Google search Volume

Friendster is a social network founded in 2002, a year earlier than MySpace, two years earlier than Facebook. As a result, the site is often viewed as the "originator" of social networking. During the peak period, the social network's number of users has been far more than 100 million people, mainly in the Southeast Asian market.

After some technical problems and a redesign, the site's traffic was in a disastrous decline in July 2009, with users fleeing and turning to other sites such as Facebook. As a social network, Friendster just curled up and died. Previously, the site had done something famous: In 2003, Friendster rejected Google's 30 million-dollar takeover offer. In the days that followed, Friendster was reinvented as a social gaming platform, still a success in the Southeast Asian market.

The question people want to know is where the site is wrong. Today, David Garcia (David Garcia) and several partners of the Swiss Federal Institute of Technology (Swiss Cato of Marvell) gave an answer. The researchers conducted a digital "autopsy" of the data collected before the Friendster "expired".

When people become members of a social network--and the time and effort needed to do so--more than they can gain, the researchers say, the time is ripe for people to evacuate the social network in large numbers. The researchers ' idea is that if a user withdraws from a social network, his or her friends are likely to go away, and the trend may flow down the entire site like a waterfall, leading to a massive loss of users.

But Garcia and his partners also point out that the topology of the network provides some resilience, which depends on the number of friends a single user has. That is, if a large percentage of the users on a site have only two friends, so when a user leaves the site, some other users are left with only one person on the site, and then the person is likely to leave, causing another user to have only one friend, and so on. As a result, the "waterfall" that the user withdraws will be running straight down throughout the website.

However, if a large number of users on the site, such as 10 of friends, then a user's departure triggers the possibility of escaping the "waterfall" is much less. In other words, for social networks, the ratio of users with a certain number of friends is a crucial indicator of whether users will be evacuated in large numbers. Garcia and his partners have studied this ratio, which they call the "K-core distribution" (K-core distribution), whose research goals include Friendster, MySpace and Facebaook, and the results of the study have illustrated everything. "We found that different online communities have different ' K-core distributions '. "They said.

Of course, an online community that is easily compromised will not crash automatically. Before the crash, the ratio of costs to benefits must fall to a certain level, leaving the user most likely to be evacuated. In other words, for social networks, a lower cost-benefit ratio and a "K-core distribution" that easily damages them will be fatal. Garcia and his partners noted that in the months before the Friendster crash, the cost and benefit ratios of the site had fallen sharply due to the design changes and the number of technical problems it had encountered.

In other words, judging from this digital "autopsy report", Friendster's "cause of death" is the decrease in cost and benefit ratio. "This indicator can be seen as a precursor to the imminent collapse of an online community." The researchers concluded. Of course, "K-core distribution" is also one of the factors.

For today's online social communities, this is clearly the lesson they should learn. In fact, the crash of Friendster has a lot in common with the collapse of the social news aggregation service Digg, whose design changes have also changed the cost and benefit ratios of its users. And if Garcia and his partners are right, then the "K-core distribution" is also contributing to the collapse of Digg.

Facebook and other social networks should also be wary of such problems if they are not ready. It's not hard to imagine how a botched design would push users out of the way, especially if there is another emerging social network that is poised to stay.

Facebook benefited a lot from the collapse of Friendster in 2009, but one day Facebook may become a victim of similar circumstances, far from impossible.