Two Princeton researchers have forecast that Facebook will lose 80 percent of its users by 2017. Facebook has responded by poking fun at the methodology and making its own forecast that Princeton will disappear by 2021, with the human population lasting only four decades beyond that.
John Cannarella and Joshua A Spechler make their forecast in a paper titled “epidemiological modeling of online social network dynamics.” It’s based on the hypothesis that social networks operate like the spread of a disease: some people join and in turn persuade others to join, and eventually some people leave and in turn other users now have less incentive to stick around.
The researchers then reasoned that if social networking is a disease, they could adapt previous research that shows Google searches for particular symptoms in particular geographic areas can forecast the spread of the flu.
They put together two models of how searches tracked by Google Trends can forecast the rise and decline of a social network. The first model simply copied the pattern that happens with a real disease, while the second model took account of the fact that people usually intend to be on a social network indefinitely whereas there’s much more constraint on how long you can have flu symptoms (to put it bluntly, you either recover or die.)
They then tested their model using the search term “MySpace” and found that although the first model was a bad fit, the second model did largely match the actual pattern of abandonment on the site. They than ran the second model using the search term “Facebook” and concluded that of the number of users at the site’s peak, only 20 percent will remain by 2017.
While some media sources reprinted the claims without explanation, many have pointed out the numerous limitations. For example, the fact that the model worked for MySpace may have been a fluke result rather than a sign the model actually works, let alone that it will work for all social networks.
For another, Facebook’s highest user base to date was around 10 times higher than that of MySpace. That changes the type of people using the site (which may affect their willingness to leave) and means Facebook is much more culturally established. There’s also been a significant rise in the number of mobile devices since the MySpace era, which could also make people more engaged with Facebook and resistant to leaving.
For its part, Facebook decided to poke fun at the study. Mike Develin, a data scientist at Facebook, has written a post where he extrapolates a variety of data about Princeton. He notes that whether you look at Facebook likes for various colleges, publications from Princeton on Google Scholar, or Google Trends mentions of “princeton”, it’s a decline all the way:
This trend suggests that Princeton will have only half its current enrollment by 2018, and by 2021 it will have no students at all, agreeing with the previous graph of scholarly scholarliness. Based on our robust scientific analysis, future generations will only be able to imagine this now-rubble institution that once walked this earth.
And just in case anyone didn’t realize Develin was being somewhat less than serious, he concludes by noting that extrapolating Google Trend figures for “air” shows that there will be none left by 2060.