Researchers conduct 'autopsy' of social network Friendster (Wired UK)

A team of systems designers conducted an "autopsy" of social networking site Friendster by analysing several online communities.

Researchers at the Swiss Federal Institute of Technology examined Friendster, Livejournal, Facebook, Orkut and Myspace in a bid to identify what makes a social network survive or decline, and what makes them capable of withstanding changes.

As the paper on the topic explains, "changes may cause users to leave, which may trigger further leaves of others who lost connection to their friends. This may lead to cascades of users leaving."

Friendster was founded in 2002 and at its peak had more than 100 million users. In 2009, having undergone a comprehensive redesign and suffered some technical problems, the site started haemorrhaging users, and was eventually closed down in 2011 and reopened as an online gaming portal.

The study seeks to find out what went wrong and takes the form of a "digital autopsy" on Friendster using data collected at the time by the Internet Archive.

The team -- led by David Garcia and co-authored by Pavlin Mavrodiev and Frank Schweitzer -- defines the social resilience of an online community as "the ability of the community to withstand external stresses and disturbances as a result of environmental changes", particularly the user interface of the social network.

They found that when the time and effort (the costs) associated with being a member of a social network outweigh the benefits, then a decline in users becomes likely. If one person leaves, their friends become more likely to leave and as more people leave, this can lead to a cascading collapse in membership.

Each network has some resistance to this decline, depending on how many friends each users has. If a user has a thousand friends, they will hardly notice when a couple leave. But if a user has three friends and one leaves then they are much more likely to leave themselves.

So if a large proportion of people who use the network have a small number of friends, it can be highly vulnerable to mass exodus.

The team used "k-core analysis" to identify the fraction of the network in which all users have at least a certain number (k) of friends. This fraction -- the k-core distribution -- was analysed for each of the five aforementioned networks.

Being vulnerable to mass exodus doesn't mean a network automatically fails. In order for that to happen, the cost-to-benefit ration must drop to a point where individual members choose to leave. So the combination of a vulnerable k-core and a low cost-to-benefit ratio create a recipe for disaster.

Just before Friendster collapsed, the cost-to-benefit ratio fell significantly due to the changes to the user-interface combined with technical issues. "This measure can be seen as a precursor of the later collapse of the community," says the study. This was the ultimate cause of death.

Interestingly, the study found that the topologies of Livejournal and Facebook are less resilient than the unsuccessful networks Friendster and Orkut. "This indicates that the environmental condition of an online social network plays a major role in its success. Thus, we conclude that the topology of the social network alone cannot explain the stories."

The report flags up a comedy video made by The Onion, which sees fictitious "internet archaeologists" commenting on the decay of Friendster.

"While proposed as a satire of the speed of internet culture, this video illustrates the opportunities that a failed social network o?ers for research. The users of such a community leave traces that allow us to investigate its failure. In this sense, we can name our work as Internet Archaeology, because we analyse non-written traces of a disappeared society, aiming at understanding the way it worked and the reasons for its demise."

You can read the full study on Arxiv.org

Image: Shutterstock

Source: http://www.wired.co.uk/news/archive/2013-02/27/autopsy-of-friendster

octomom dan savage new world trade center