“Tell me when you started tweeting, and I'll tell you who you are.” Jameson Toole, Meeyoung Cha and Marta González don't go quite that far in their conclusions, but their study of when and where people signed up to the microblogging platform Twitter provides a wealth of information about the process of social contagion and technology adoption (PLoS ONE 7, e29528; 2012). In particular, they find quantitative evidence that for 'early adopters' geographic location was a key factor to reach a critical mass, whereas at later stages the influence of mass media took over.

Credit: © D. SHARON PRUITT PINK SHERBET PHOTOGRAPHY/GETTY IMAGES

Toole et al. looked at Twitter's week-by-week user gains between early 2006 and late 2009. From the data, they identified the 408 locations in the United States where more than 1,000 users had signed up during the first three-and-a-half years of Twitter's existence. The Twitter history of each of these places was distinctly different. The first hotspots were, not surprisingly, close to Silicon Valley and in college towns around the United States — places with a young, tech-savvy population. During this early phase, old-fashioned word-of-mouth recommendation within local networks seems to have been mainly responsible for the growing user base.

Later, the Twitter 'virus' spread to major metropolitan areas, and then on to more rural and remote locations. Then another factor came into play: mass media. To measure the effect, Toole et al. drew on data from Google, capturing the relevant news and search volumes. They found, in Twitter's early life, a direct correlation between the number of Twitter sign-ups and the number of search queries and news reports on Google. But some 120 weeks after the launch, the rate of user gain became super-linear. By that time, celebrity endorsements and reports of Twitter's role in the organization of demonstrations and revolutions had created a kind of feedback loop — with the media reporting on the increasing adoption of tweeting while themselves driving it. Toole et al. estimate that mass media has been responsible for a growth of the Twitter user base by a factor of two to four.

This work extends the existing body of literature on the diffusion of innovations, although the case of Twitter may be a special one, as the service comes free of cost and with little risk. How the particular interplay between geographic and media influences captured in this model may relate to the adoption of other innovations and technologies remains to be seen. But the study highlights that traditional contagion and diffusions models need amendment if they are to capture processes on modern information networks.