Is peer influence at play in large-scale online social networks? That was a question Professor Ravi Bapna and Assistant Professor Akhmed Umyarov explored in their research into consumer social network behavior.
The pair attempted to uncover whether the primary driver of correlated behavior is peer influence or homophily, which is the tendency of people to associate with like-minded individuals.
"Peer influence is if a friend convinces you to join an action, say buying a product or service, in a social network," says Umyarov. "On the other hand, maybe you and someone else have high incomes, a lot of free time, and other things in common. If you both behave similarly on the same network, that's homophily."
As Umyarov adds, separating the two is not easy. Hence their study, which saw them provide 1,000 gift subscriptions to Last.FM, a freemium-style social network music service that allows you to view the giftee's subscription and its effect on his or her network. Freemium communities allow for free usage, but users have to suffer advertising, as well as paid subscriptions. On Last.Fm subscribers are 36 times more profitable than paid users, but only 1 percent of the 30 million users are subscribers.
"We compared the friends of the group that got our gift against the parallel universe of the friends of the random 'control' group of 1,000 people that did not receive the gift," says Bapna, the Board of Overseers Professor of Information and Decision Sciences. "It was a double-blind, randomized trial. We could attribute any difference between the two groups to our manipulation of the first one."
Their key result: a 50 percent increase in people from the first group's networks paid for subscription for Last.FM which Bapna says is clear evidence of peer influence.
"We also discovered that people with few friends are most susceptible to peer influence," says Umyarov. "If you have few friends, and one of them gets a gift subscription, you're more likely to subscribe too."
These results have implications for companies looking to capitalize on social networks. "If you're a marketer with a limited budget, who would you pick to influence?" says Bapna. "But how do you maximize that influence? That's a different problem."
Both note that their research is only beginning. "We targeted average users," Umyarov says. "What would happen if we could target truly influential users? That's an interesting question, but we haven't answered it yet."
"Do Your Online Friends Make You Pay? A Randomized Field Experiment in an Online Music Social Network" (Bapna, R, Umyarov, A., ) was presented at the National Bureau of Economic Research Summer Institute on the Economics of IT and Digitization, 2012