October 13, 2010
Dissecting the Science of Social Media: HP's Social Computing Labs Chief Huberman Shows How Twitter Popularity Does Not Equal InfluenceRichard Carufel's exclusive interview this week: Dr. Bernardo Huberman, Senior HP Fellow and Director of the Social Computing Lab at Hewlett-Packard Laboratories
Social media's power may be overrated.
As Twitter's popularity ballooned among tech-embracing young people — with the rest of the population increasingly following suit — leave it to the cold and calculated hand of hard science to shoot a few holes in the genre's mystique. And that's exactly what has happened over the course of several months of work by the Social Computing Lab at Hewlett-Packard, led by noted physicist and Stanford University consulting professor Dr. Bernardo Huberman.
As a senior HP Fellow and director of the Social Computing Lab at HP, Huberman and his team have been running social media through the scientific ringer in order to quantify the genre's true power, which Huberman admits is significant — "After all, there are almost 200 million people with Twitter accounts," he notes — but he goes on to draw a clear distinction between Twitter's key metrics: popularity and influence.
In their vast and comprehensive research — analyzing more than 22 million "tweets" — Huberman and his team decided to further focus on the disconnect between Twitter popularity and influence, and their scientific research led to the creation of an IP Algorithm, which measures the influence and passivity of Twitter users. One key, and very telling, finding shows that, since the majority of Twitter users are passive — i.e., they don't retweet the information they receive via the 140-character messaging vehicle — high numbers of followers doesn't necessarily add up to greater influence. See the empirical evidence in the group's paper "Influence and Passivity in Social Media — HP Labs Research."
The research suggests that, although many tweets reach a high number of original followers, the recipients do not necessarily pass the information along to their own audiences — or even click on the URLs embedded therein. Here, Huberman explains why the accepted logic — that bigger numbers equals greater influence — does not compute.
How did social media come across your scientific radar?
This is part of my study of the economics of social attention. We have access to this information about Twitter and other social media. One of our key findings is that there's a difference between being popular and being influential. Some who are very popular have little influence, and vice versa. We're debunking the notion that popularity means influence.
And when it comes to branding, influence is a strong word. But isn't popularity just as good?
Half the people on Twitter are following brands, like HP and other companies. So some people believe this type of communication is important for PR. As a result, many companies are becoming media companies because people are making many more purchases and decisions via the web. How people perceive you on the web is clearly important to companies.
There is an incredible value to those media — after all, there are almost 200 million people with Twitter accounts. There's no doubt its effective, but does that mean it's influential? That's what we been studying.
How do the findings support the results?
We measured influence very carefully. We've seen that some people retweet everything they get. Others are purely passive. If you can spur to retweet those who are more passive, then you have considerable influence.
But only 1 in 300 users are retweeting. If a lot of people are following you but not retweeting, then your level of influence is clearly limited.
Twitter has certainly hit something in the subconscious of people as a way of communicating things, which is why it's such a success. But whether Twitter is a great means for selling products or [having a marketing impact] is yet to be shown. I can't really say whether Twitter is the great new marketing tool [that many say it is] — that's not part of our mission with this project.
What will come of these studies that might give communicators some actionable options?
We're working now on a tool by which people can measure their level of influence, which we hope to have ready within the next couple of months. Soon you will be able to go to a website and find out just by plugging in your Twitter address. The algorithm we're developing will analyze your number of tweets, retweets, passivity and number of followers, and give you a ranking.
In the meantime, what can PR professionals and other communicators do to take more control of their "influential" status?
You can control your level of influence only by creating things people care about, which is obviously an intangible. Is it because you're more authoritative, more interesting, or send out more information? We can't be sure of that — it's a very complicated psychological phenomenon.
How do you get people to retweet?
It's a matter of approach. If you have information that's interesting and has never been heard before, it will get passed along. If you're sending out lies, then you might have a spurt of popularity, but it probably won't last very long.