Do any of us have influential friends? Photo: Flickr/Auntie P

Do your friends influence your taste? Not so much, according to a Harvard University study.

The Harvard researchers tracked college students’ Facebook relationships and measured how taste in music, movies and books spreads through social networks. It turns out that the degree to which your friends’ tastes and yours are connected has more to do with how you became friends in the first place than the force of that allegiance later on.

‘The message has overwhelmingly been one of how easily ideas, beliefs and behaviors spread through social ties’

When it comes to taste, “peer influence is virtually nonexistent,” said Kevin Lewis, a Harvard sociology graduate student who co-authored the study. Lewis cautioned that the experiences of college students on Facebook may not apply to everyone in all circumstances, but the results offer a sobering counterpoint to the conventional wisdom on the ubiquity of taste diffusion. “The extent to which friends’ preferences actually rub off on each other is minimal,” he said.

The one exception the study found is fans of jazz and classical music.

Our lack of influence on each other calls into question a large body of conventional wisdom on social networks, social media and viral marketing. A Google search on the phrases “social media” and “viral marketing” returned about 16,700,000 results. If we don’t influence each other, does that means viral marketing is a bogus concept? And what does it say about the business value of social media?

Many recent reports in academic literature and popular media tout the concepts of social contagion, social diffusion and social epidemics, said Lewis. “The message has overwhelmingly been one of how easily ideas, beliefs and behaviors spread through social ties,” he said.

But just because we can all think of a few instances of liking a certain band, movie or book because a friend liked it doesn’t mean that peer influence is a robust social phenomenon, said Lewis. “How many of our close friends’ tastes don’t we share? How many friends of ours have tastes that we don’t know about or pay attention to at all? And of those tastes that we do share with our friends, how much of that is a consequence of diffusion and how much of that is because those shared tastes were part of the reason we became friends in the first place?” he said.

Some evidence in reports of peer influence is anecdotal, some is derived via questionable methods and some can’t be generalized, said Lewis. “For all that has been published on diffusion, we still have a surprisingly poor idea of how diffusion dynamics vary across different people, places and behaviors,” he said.

The Harvard team used the most statistically rigorous tool available — stochastic actor-based modeling — and applied it to a readily measurable population: college students on Facebook, said Lewis. The study began with 1,600 students and followed their Facebook activity over four years. The study’s final models of selection and influence were derived from the activities of 200 students whose friendship and taste data the researchers were able to collect over all four years.

Where’s the value?

The study’s findings suggest that it would be much more worthwhile to invest in understanding how and when friendships are a conduit for preferences, rather than assuming that they are and planning marketing strategies accordingly. “They clearly are under some circumstances, but we still don’t know whether those circumstances are common or important enough to warrant the time and money of business strategies,” said Lewis.

Advertising giant Ogilvy & Mather published a study in October that looked at social media and sales and brand perception. “What we see is a more complex relationship between friends in the context of influence,” said Irfan Kamal, the firm’s senior vice president of digital/social strategy. The agency studied the media touchpoints of five fast food brands on 400 people. They found that people were two to seven times more likely to increase their purchases from restaurants when they were exposed to those restaurants on social media, said Kamal. Some of that exposure is branded content, he said, but “we do know that social content by friends was influential in impacting purchase.”

In general, it’s difficult to distinguish which preferences have been influenced, said David Armano, Executive Vice President of Global Innovation & Integration at advertising agency Edelman Digital. “What we look to measure in social media marketing are things like social sharing,” he said. “You track what’s being passed around social networks and measure surface indicators such as likes and retweets.”

It is who you know

One of the most valuable aspects of social media is who you know. It’s easy to glean information about members of social networks. This focuses sales, marketing and product development efforts. Knowing something about one person gives you insights into the people that person knows.

The Harvard study affirmed that, as in other aspects of life, people’s social media relationships tend to be with people who are like them. Gender, race and socioeconomic similarities govern friendship evolution on Facebook. Friendships also form when people frequent the same places and have friends in common. “From a business perspective social media are obviously tremendously valuable, in particular for understanding who befriends whom and who likes what,” said Lewis.

Who you know is arguably a more valuable aspect of social media than who you might be influenced by. Even in a world where friends don’t influence each other’s tastes, businesses can gain valuable information from mining social media ecosystem for who knows whom, said Armano. “That data can help a business make better decisions versus merely attempting to influence choice.”