It's Who You Know (Or Don't)

Nothing demonstrates the wildfire spread of online social networks such as MySpace and Facebook better than the experience 73 Stanford students had last fall. They were enrolled in the computer science course Creating Engaging Web Applications Using Metrics and Learning on Facebook—and did they ever engage. At a public session held at the Arrillaga Alumni Center toward the end of the quarter, 500 people heard how “10 million [users] in 10 weeks” were expected to have installed the students' free applications, some of which were reportedly generating enough ad revenue to pay tuitions.

Most of the apps resemble party games or amusements—ScribbledPhotos lets users draw graffiti on Facebook photos; with KissMe, members can bestow kisses—although some focus on generating useful information. But co-instructor BJ Fogg, an experimental social psychologist whose Persuasive Technology Lab studies how computer technology changes people's beliefs and behaviors, dismisses critics like one blogger who called the applications “monumental drivel.” The point was for students to learn how to think, using psychology and metrics to make their applications more appealing. And by quarter's end, 10 million proved an underestimate: “What other class in the history of the world created student projects that reached 16 million people in 10 weeks?” asks Fogg, MA '95, PhD '97.

It's that kind of dramatic result that has a growing number of academics starting to look at how online social networks function and what they deliver compared with traditional, offline connections. Fogg can envision revolutions; others are not so sure.

Students and teachers in the Stanford Facebook class share in the credit, of course, but much of their success must be attributed to the intrinsic power of Facebook—which Fogg proclaims “the No. 1 persuasive technology of 2007 and maybe of all time.” Online social networks are inherently “viral,” built through overlapping circles of friends, relatives, colleagues or acquaintances. In most other businesses, referrals and positive word of mouth come from good service. But to make almost any use of a networking site, you first must be connected to people. So new users quickly start “friending,” or inviting, others, in effect shilling for the site before they've really tried it themselves. Because the major sites enable users to import their e-mail contact lists, it's easy to invite everybody you know with one click. True, the invited must accept to join, but joining is free and declining or ignoring a friend request can feel awkward.

As a site reaches critical mass, bandwagon-like “network effects” feed further growth: since new users want to be where others are, numbers can grow with epidemic speed. Today, hundreds of millions of people worldwide belong to at least one online social network, making MySpace (with 110 million members) and Facebook (70 million) among the web's top 10 sites. Beside these behemoths are many others, each with its own slant and many run by Stanford alums (see below). LinkedIn, for example, is generally regarded as the most popular professional network, with 20 million users.

Such hyper-connectedness sets today's social networks apart from the online communities that have been around since the dawn of the Internet, where people gather to share common interests from hobbies to rare medical problems. The newer networks offer enough services—and entertainment—to keep users glued to the screen for hours a day. For some, MySpace and Facebook are integral to their lives and identities: users can endlessly spruce up their profile pages; share photos and videos; play games; compare tastes in books, movies and music; keep up to date on the goings-on of friends (and strangers); buy and sell through classified ads. On Facebook, users can engage in more civic-minded pursuits, too—rallying around political causes, supporting charities, coordinating blood drives and even lending money.

But all these activities aren't Facebook's main draw, Fogg concludes from his survey of Stanford students. “It's mainly about connecting with others in a convenient way,” he says—keeping in touch with old friends, building new relationships and expressing your identity and views. “The games are an excuse to do these other things.” MySpace, Fogg observes, is more about entertainment and playing with identities.

On the all-business LinkedIn, less colorful and time-consuming by design, users can tap into their networks for job leads, employees or vendors; try to get the inside scoop on job candidates or prospective bosses; introduce one person in their network to another; ask questions or build their reputations by answering questions; or just put up their profiles in hopes that opportunity knocks.

How much does all this differ from what people do offline? Not very, some observers say. “Social networking is an intrinsically human activity that goes a long way back and is closely tied to what we consider progress and civilization—as well as nastier things like organized warfare,” says Howard Rheingold, a lecturer in communication and the author of The Virtual Community (Basic Books, 1993) and Smart Mobs (Basic Books, 2002), among other books about the Internet's social effects. He says that social networking technology has merely lowered distance barriers to communication, much the way the printing press did in its time.

Bernardo Huberman, a consulting professor of applied physics, concurs. He also is director of the Social Computing Lab at HP Laboratories and is the author of The Laws of the Web (MIT Press, 2001). “The main effect of the sites has been the immense collapse of distance and cost,” he says, likening the situation to what happened after the cost of telephone calls fell. “Twelve to 15 years ago, a long-distance call was expensive. Even though you may have had friends in London, you didn't call your friends in London every day.”

Online social networks also let you broadcast information to everyone at once. Establishing and perhaps maintaining a relationship is much easier online than in the so-called real world, says anthropology lecturer and academic technology specialist Claudia Engel, who has taught classes on digital ethnography. “The cost-benefit balance changes dramatically.” It's not just about dollars and cents, either. “Sweaty palms, fear of rejection—all that disappears online,” Huberman notes.

Much of this is true of e-mail, too, which—at the worst extreme—lets users spam millions of others at shockingly low cost to the spammer. But computer scientist Ashish Goel, MS '98, PhD '99, associate professor in management science and engineering, says there's something more intriguing about social networks than pushing information out—namely, pulling in information from diverse sources. On LinkedIn, for example, you can quickly find out who in your network knows a particular individual. “It's very hard to call all 250 of your friends and ask, 'Do you know this person?'” Goel observes.

Getting a promising job lead, for example, “is not dependent on your cousin remembering who they might have worked with a couple years ago in Cincinnati,” says Ingrid Erickson, a graduate student at the Center for Work Technology & Organization in the department of management science and engineering. LinkedIn, in which users' public profiles are essentially résumés, provides an up-to-date record of skills, experience and connections. “There's a bit of a knee-jerk reaction that face-to-face is the gold standard,” Erickson says. She argues that online exchanges sometimes can be better. Because your online profile remains accessible, you can make a stronger, longer-lasting impression—and perhaps connect in person later. Conference attendees, for example, can learn about each other beforehand, making their face-to-face networking more targeted and effective than in the days of exchanging business cards around the punch bowl.

No one disputes the exponentially greater accessibility of information and connections online, but opinions vary on how genuine or effective online relationships can be.

Some critics point to the meaninglessness of amassing online “friends” the way you might collect baseball cards. The sites themselves are partly to blame, Fogg suggests. “A thousand connections—on LinkedIn, that's a status marker,” he says, noting that the interface was clearly designed to display everybody's number. “And given that it was used in Silicon Valley first, among competitive people, that number became like a video-game score.”

“There has to be a trade-off between the volume of connections and the quality,” says economics professor Matthew O. Jackson, author of the forthcoming book Social and Economic Networks (Princeton University Press). He points out that having quick, easy access doesn't ensure strong possibilities for trust to develop and valuable transactions to occur. “There's informational value, but ultimately how valuable a particular relationship becomes depends on how much people interact.”

Fogg agrees. “At the beginning, the label 'friend' suggests you're going to watch each other's back,” he says, “but after you get 60 connections or whatever your social capacity is, it doesn't mean you'll travel the world and ring in the new year together.” We have room for only so many relationships in our lives, which is why the costs of real networking—not exchanging virtual gifts on Facebook or inviting your former co-workers to join your LinkedIn network, but helping others as you hope they may one day help you—stay stubbornly high.

There's another way in which online networks are a poor approximation of an offline network: they take a crude, binary view of human connection. On the sites, someone is either your friend or not, and early attempts to let users subdivide friends (such as Facebook's Friend List feature) still seem robotic and inadequate. That lack of nuance leads to a host of problems, Fogg says. “If somebody sees that Felix is my friend, and let's say Felix is [really] an acquaintance or stranger, they're going to make assumptions about Felix that he's a great guy—and that gets into problems of trust and credibility.”

Just as in the offline world, social networks rely on trust—but trust online doesn't always work the way users expect. For example, I asked LinkedIn's co-founder and VP of product strategy Allen Blue, '90, why I might have trouble getting hold of somebody who on LinkedIn is a connection of a connection. “Every person within the system acts as a gatekeeper for everybody else in the system,” he told me. “There's no assurance that any of those gates are going to work at any given time. You rely on the quality of the relationship you have with somebody, and that person relies on the quality of the relationship they have with somebody else.”

Sociology professor Karen Cook, whose current research centers on trust in social relations and networks, says that even if all the relationships in the chain are strong, trust goes only so far. And to be worthy of our trust, someone must be both competent and well-meaning. “I may trust my husband with a child,” explains Cook, '68, PhD '73, “because he's highly motivated to take care of the child, but not trust him with a checkbook—but that's more about competence.”

What's more, trust is not transitive. As Rheingold puts it, borrowing from the writer Clay Shirky, “I may trust my friend's friend, but I don't trust my friend's friend's crack dealer.” Trust decays across the network, Cook explains, because the farther removed someone is, the less we know about why and how much others in the chain trust that person.

Still, there is at least a theoretical basis for thinking that online networks might trump narrower offline connections, especially for business purposes. Sociology professor Mark Granovetter wasn't thinking about online networks when he coined the phrase “the strength of weak ties.” It was the 1970s, and Granovetter was studying how people found employment. The most fruitful connections for job leads, he found, were not close friends and family, but more tenuous contacts—weak ties. That's because people close to you are privy to more or less the same information you are; those outside your inner circle will know of opportunities that you and your friends don't. With the rise of online social networks, where people can maintain many more weak ties, Granovetter's idea took on new legs.

So did the even older theory, popularly known as “six degrees of separation,” that any two people in the world are connected by at most a few intermediaries. The idea gained traction following Stanley Milgram's 1967 Small World Experiment. Milgram aimed to discover how many steps it would take for a package—sent by one person to an acquaintance who in turn would send it to a friend, and so on—to reach the hands of a total stranger.

Over the years, the popular imagination has distorted Milgram's results, as psychologist Judith Kleinfeld uncovered more than 30 years later. The endpoints of Milgram's original experiment were not random strangers around the globe, but a small sample of Americans in Massachusetts and Nebraska. Moreover, Milgram's six degrees was an average, not a maximum. In fact, the overwhelming majority of packets never reached their targets.

Yet the legend of six degrees persisted, and online social networks added credibility. We're comforted to know that it's a small world after all, and it's nice to think we can connect to opportunities through people we already know. The defunct SixDegrees.com made this promise explicit, but other networks at least hint at the same possibility.

Certainly, social networks exhibit the “small-world property.” Some individuals know many more people than the rest—think politicians, agents, hairstylists—and because of these well-connected hubs, it takes far fewer jumps from node to node to connect any two individuals on a social graph than if social ties were distributed evenly. But mere connection isn't enough: many of Milgram's subjects didn't so much as mail a packet forward in the chain. The mathematical sociologist Duncan Watts, author of Six Degrees: the Science of a Connected Age (Norton, 2003), has been replicating Milgram's study on a large scale, using e-mail, and has found the same problem with messages not reaching their targets.

Software like LinkedIn's can lessen that problem by revealing the shortest path between you and another individual in the network. But there's a greater obstacle that databases can't overcome. Small-world networks work through hubs, but most of the hubs' ties are, almost by definition, weak. And, as Graduate School of Business professor of organizational behavior Hayagreeva Rao puts it, “Typically, you get information from weak ties, but you can't count on them to get anything done.” This is true online and off. “It's costless to add a connection,” Rao says, “but when the person asks for an actual favor that implies trust or some hassle, you're going to say, 'Wait a minute.'”

Granovetter's research suggests the same thing. People he studied didn't find jobs through chains longer than one or two—either a friend or a friend of a friend. “Beyond that point, it's no longer more effective than the newspaper,” he says. Interestingly, a quarter of LinkedIn's revenue comes from newspaper-style job listings.

In 2004, three years before Facebook became a household word, Peter Thiel invested $500,000 in the fledgling social networking site. Thiel, '89, JD '92, learned about the investment opportunity from his close friend and former PayPal colleague Reid Hoffman, '89, himself a co-founder of LinkdIn. Hoffman had found out about the Facebook opportunity almost by chance. Its founder, Mark Zuckerberg, had moved from Boston to Palo Alto and became housemates with the Internet entrepreneur Sean Parker. “I knew Sean, and I liked Sean a lot,” Hoffman told a San Jose Mercury News reporter. Hoffman thought highly of Facebook, too, so he introduced Zuckerberg to Thiel. Today, Thiel's stake in Facebook, at least on paper, is worth about $1 billion, and everyone else in this chain has reaped benefits, too.

Yet online networking played no role in matching up these online networking insiders. “The vast majority of economically significant networks are not online, they're offline,” says Graduate School of Business assistant marketing professor Harikesh Nair, who studies economic implications of social networks.

Granovetter, who has been studying Silicon Valley networks, says that high-level VCs “are going to avoid those social networking sites like the plague. Their whole stock-in-trade is their network, but they're never going to manage those networks online: it's too democratic.” For them, the costs of belonging—notably loss of privacy—presumably do not justify any possible benefits. (However, several online social networks restricted to VIPs have sprung up; time will tell their effect on offline life.)

The economic benefits of online networking are hard to determine—in part because any tangible payoff occurs offline. “I don't think we have good studies of it,” Granovetter says, “but as far as I can tell, online social networking hasn't made any substantial difference in how people find jobs.” That could change, as job search sites like Jobster and CareerBuilder link up with Facebook and members increasingly use the site to network for jobs instead of purely socializing.

But for BJ Fogg, the important thing is that technology can lead to interpersonal persuasion on a massive scale, and his lab's next initiative is a lofty one. Peace Innovation (peace.stanford.edu) aims to inspire international peacemaking experiments using YouTube, Flickr or other second-generation web services. The ultimate goal: world peace in 30 years. Fogg knows that sounds ambitious, even a bit crazy. “But I think it's crazier to not try,” he says. “The world's in pretty bad shape, but we have some amazing tools to connect and change people. We're not going to sell laundry detergent with them.”

Some Alums in Social Networking

MARINA KRAKOVSKY, '92, is a writer in San Mateo.