Saturday, February 5, 2011

The Primal Forces that Drive Social Networks

Social Networks are revolutionizing how we view our world. People are connecting, businesses are being created or transformed, and the world seems like a smaller place. As with any transformation on a grand scale, a plethora of consultants, gurus, blogs, and how-to books have risen to meet the demand for information about the social revolution.

However, it is very rare to hear anything about the underlying forces that actually drive the social network phenomenon.

It’s a shame because the story is a great one that has implications, not only for social media, but for fields as diverse as counter-terrorism, ecology, economics, organizational theory and cancer research. Network Theory has fundamentally changed our understanding about how the world works since its inception a decade ago. Most of all, by understanding how networks form and grow, we can build better ones.

Fireflies and the President of the United States

Our story begins in 1996 at Cornell University in Ithaca, New York, where an adventurous rock climber and former Australian Navy Officer named, Duncan Watts, was thinking about how crickets, frogs, fireflies, and pacemaker cells all seem to be able to synchronize their behavior within large groups.

His mind must have began to wander because he suddenly remembered that his father once told him that everybody is just six relationships away from the President of the United States. The concept had existed in literature early in the 20th century and was documented in Stanley Milgram’s famous “Small World Experiment”.

In a flash of inspiration he went to his PhD thesis advisor, Steven Strogatz, and told him that he wanted to, yet again, change his thesis topic. Watts had a hunch that both phenomena might be related. Strogatz, somewhat used to giving his brilliant student leeway, consented.

The Strength of Weak Ties

As he began his research, Watts came across a highly cited paper written by Mark Granovetter called “The Strength of Weak Ties” about how people find jobs. He found that most people don’t locate employment through their friends, but through friends of friends.

Granovetter dubbed these relationships “weak ties” (after the attraction between water molecules that give the liquid many of its properties). Granovetter surmised that it is through weak ties that information is largely distributed. While we can maintain relationships with relatively few people, the people they know greatly increase our access to facts, knowledge and wisdom.

We have friends from work, school, our neighborhood, etc. While our ties may be strong ties to us, they are weak ties to our friends from separate clusters. For instance, the felon in our neighborhood can be connected to the law professor at our university in only two steps!

Spacemen vs. Cavemen

Watts also began thinking about his youthful love of science fiction and two Isaac Asimov novels in particular; one about spacemen and another about cavemen. The spacemen communicated remotely so that the people they knew didn’t know each other, while the cavemen lived in isolated groups and knew everybody their friends knew. He decided to build a mathematical model that would describe both situations and every possibility in between.

In addition to the “degrees of separation” metric (the average number of links it takes to get from one network member to another), Watts also created a “cluster coefficient,” in effect how tightly clustered communities are within the network.

A good analogy is a school lunchroom. How many people who have close relationships would be calculated by the cluster coefficient while how many introductions one would need, on average, to get to any particular person, would be the degrees of separation (or more technically, path length). This type of calculation has been second nature for poor note-takers and class-cutters alike for ages.

Armed with mathematical representations for both his “spacemen and his “cavemen” he could experiment with different types of networks.

Small World Networks

What he found was startling. In his model, as communities connect to each other, the social distance between people increases – up to a point – and then immediately comes crashing down. It turns out that it takes just a little bit of mixing for the social distance to decrease by an enormous amount, but a lot of mixing to kill communities. Although surprising, the pattern was familiar. Similar “instantaneous phase transitions” have been long known in Physics.

Moreover, he found that in almost all cases, the same result appeared, it was only a matter of time for a network under fairly normal conditions to reach the optimal state. Globally connected networks with strong local cohesion are not only possible, they are the equilibrium case – you just needed a relatively small number of Granovetter’s “weak ties” mixed in to make the whole thing work.

He called the result a “Small World Network” after Milgram’s famous experiment.

Hey! Networks Grow, Don’t They?

Watts published a paper on his findings with Strogatz and it became an immediate success, widely read and cited throughout the scientific community. At Notre Dame University, Albert-Laszlo Barabasi and his student, RĂ©ka Albert, noticed an oversight – networks grow over time and large communities within networks drive the growth. They quickly published their own paper.

What they found was that networks follow a very specific mathematical rule called a “power law” that described well known phenomena such as the “80/20 rule” and Chris Andersen’s now famous long tail. Their findings suggested that even very large networks were driven by relatively few “hubs” around which everything else was organized.

The two teams continued to trade papers back and forth and in a very short time Network Theory had arrived!

Implications of Network Theory for Social Media

Through understanding the forces that drive social networks, we can take some practical steps to improve Social Media performance.

Communities are primary: A network is only as strong as the communities that it contains. A big mistake that many Social Media efforts make is to pursue broad coverage early on. Building enthusiastic, devoted communities requires a local approach (either geographically or in social space). Those local communities have “weak ties” to other communities in other places, even faraway places. So you really can think globally by acting locally.

People want to connect: Connections between communities naturally grow over time for the same reasons that information wants to be free and dictatorships are expensive to maintain. Any opportunity to implement open architecture (while maintaining security protocols for the site core) should be seized upon. Walling off a social network is choosing the path to obscurity (although hardly the one less traveled).

Large clusters drive the network: A small number of extremely active members drive network growth. Mostly, they are driven by reputation and attention so it is crucial to give users every opportunity to be recognized by their peers.

Social Media isn’t successful… until it is: A network doesn’t grow in a linear fashion and it doesn’t grow in just one direction, but two: outward and inward. Watts described a network maturing as an “instantaneous phase transition” similar to a crystal forming. The process moves relatively slowly and then, suddenly, a new global state is achieved. Once a “Small World Network” has formed, the growth becomes exponential.

Social Networks on the web can be extremely powerful. Once you understand the forces that drive them, you can make their horsepower work with you and not against you.

Note: For those of you who are interested in learning more, Watts and Barabasi have both published highly readable and informative accounts of their Network Theory adventure and the friendly rivalry. It’s a lot of fun to read both sides and learn both about their triumphs and their frustration when the other one uncovered something which seemed fairly obvious in retrospect. Besides being brilliant both write well and in friendly and engaging styles. In fact, the books are much more accessible than journalist accounts of the same events.

The titles are “Six Degrees” (Watts) and “Linked” (Barabasi). Steven Strogatz has also published a great book called “Sync” that covers pre-cursor work to Network Theory. All are refreshing counterpoints to “guru books” and offer true insight and wisdom.

a la

By Zak with No comments


Post a Comment

Note: Only a member of this blog may post a comment.

    • Popular
    • Categories
    • Archives