If you are a Facebook denizen, or a Twitter aficionado, you are in for some surprises. Gender and age dynamics run deep, perhaps deeper than you thought, on social networks.
If you are a guy, take heart: Men are 49 percent more influential than women.
Women, take comfort in the fact that you are able to influence men more than your fellow ladies. To be precise, women are 12 percent less susceptible to influence than men.
Married individuals are the toughest to influence, that is, they are least susceptible to getting influenced when it comes to adopting a product.
Single or married people are more influential than those whose status says, “It’s complicated” or “They are in a relationship”! Older users – age 30 and above – are more influential than younger users. (From real life experience, I’d say that’s common sense.)
It’s now understood that there’s an influence and susceptibility trade off – that people who are more influential tend to not be susceptible and people who are susceptible tend to not be influential. In fact, almost no one is both highly influential and highly susceptible.
Not only are some people clearly more influential than others, some of these highly influential people are themselves connected to other highly influential people, giving them the potential to be super-spreaders.
These are findings from an interesting (even intriguing I’d say) study of over 1 million Facebook users by two professors from Stern School of Business at New York University which is being published in today’s edition of the journal Science.
Sinan Aral and Dylan Walker studied this problem by tracking how an app for rating movies and actors spread through a group of Facebook users. As each user downloaded and used the app, a random group of his or her Facebook friends were notified about the activity and offered their own chance to start using the app. The researchers measured influence by looking at whether a user’s peers adopted the app after receiving the message.
Finding influentials is a big rage today. Companies like Klout are trying to measure “influence scores” for people in social media networks like Facebook and Twitter, so that brands can target them with advertising to help spread their products. So, what was the primary objective of this study given that in the last few years, scholars from disciplines as diverse as finance, sociology, and economics have been interested in knowing how social networks or peers influence behaviour or decision making?
The goal of bringing rigorous science to this task is actually much broader and farther reaching, says Sinan Aral, one of the co-authors. Understanding social contagion – how behaviors spread in a social network – is critical to designing successful intervention strategies to promote or contain any behavior in a population.
Policy makers, parents and managers have been interested in whether children’s peers influence their education outcomes, whether workers’ colleagues influence their productivity, whether happiness, obesity and smoking are ‘contagious’ and whether risky behaviors, like drug abuse, spread as a result of peer-to-peer influence. This study, he says, helps us understand these social behaviors much more precisely.
Several studies have been published in recent times, including “The Spread of Behavior in an Online Social Network Experiment from MIT, and Dynamic Social Networks Promote Cooperation in Experiments with Humans from Harvard. However, a key differentiator of Aral’s and Walker’s study is that it avoids known biases in current methods. The jargon for this is ‘Homophily bias’. In other words, the bias that emerges from birds of a feather flocking together – that we tend to make friends with people like ourselves. So, if two friends adopt a product or behavior one after the other, current methods have a hard time distinguishing whether it’s because of peer influence, or rather, that the friends simply have similar preferences and thus behave similarly. “Our method solves this problem and other estimation challenges using randomization,” says Aral.
The authors say this research also helps resolve recent debates about the “Influentials Hypothesis.” Some, like Malcolm Gladwell, author of The Tipping Point, argue that influential individuals catalyze the diffusion of opinions, behaviors, and products in society. But others contend that it’s not a small number of influential individuals, but rather the prevalence of susceptible individuals that catalyze social contagions. This work shows that it’s actually the joint distribution of both influence and susceptibility in the network which together determine the pattern of the contagion.
At a time when social networks are crowding the Internet, finding a rigorous, scientific way to understand how information spreads rapidly or how it affects behaviours offline, has many uses. For instance, other than using it for marketing products, it can be used for social good. Just as the authors are now trying to apply the same science to promote HIV testing in Africa and other positive behaviours like exercise and political awareness.
In 2009, when the avian flu pandemic broke out, varied sentiments were vented about the novel influenza vaccine. Researchers then studied publicly available data from Twitter to measure the evolution and distribution of those sentiments. They found that projected vaccination rates based on Twitter feedback, tallied with vaccination rates estimated by the Centres of Disease Control, US, using traditional phone surveys.
That said, we also see a data mining, or refining (if you accept the term that Acxiom, a US company, uses for decoding the consumer DNA using data from Facebook and other sources) explosion out there, and all this is certainly meant for targeted advertising and viral marketing.
In that case, how does this classification – of influentials and susceptibles – compare with the brick and mortar marketing world when it comes to pushing a product, I asked Aral.
Digital marketing has been shown to be potentially very effective and it is not clear yet how influence translates over different channels of communication, he says. The science of social influence is only now beginning to blossom.
“More research is needed for example on whether influence is a generalized characteristic of individuals or rather product and behavior specific (my intuition is that it varies by product, behavior and person) and whether it varies by channel – for example whether its different face-to-face or digitally (we simply lack hard scientific evidence on these questions at this point),” he says.
Do these findings hold true for other products or behaviours, say supporting a political candidate or donating for particular cause or voting Indian idol? We need more studies for that, say the authors.