Krebs believes everything is quantifiable as a social network, from steroid use to linked websites to a strand of HIV working its way through the porn industry.
He is at the cutting edge of the growing discipline of social network analysis, and creator of InFlow, one of the most advanced social networking software tools.
The field has exploded recently as social networks, the complex sets of relationships between members of groups, have formed the backbone of popular Web systems like Facebook and Google's search crawler. Social network analysts use software, like Keyhubs and NetMiner, to uncover how the structure of peoples' connections affect their thoughts and actions.
At the annual PopTech conference this week in Camden, Maine, Krebs will present his top ten social network trends including the idea that the web, contrary to popular lore, is making us less diverse and more prone to extreme ideological thinking.
"What you know depends on who you know," Krebs says. "Depending on the network of information you are in, that is the information that you will believe."
Krebs is mining his data from some strange places with surprising results.
He argues that the moment a few key, highly connected ball players started using steroids in the early 1990s, the explosion of drug use in baseball was inevitable. And, he says, if the league was fully aware of the strength of the drug culture network in its early stages, the whole mess could have been avoided.
Steroid use spread because of the wicked combination of a closed network, or cluster, and positive reinforcement in the form of higher pay for better performance. And because the members of the baseball cluster frequently move between teams, the message spread quickly.
The investigators who blamed the individual trainers distributing the drugs and the individual players referring other players got it wrong, Krebs says.
Though he maps out the individual influence of players, he sees an underlying ideological shift at work. The fact that steroid use took hold and spread so quickly in spite of laws against its use, suggests that the closed nature of the cluster was more important than the individual cheaters, he says.
And the same process happened in reverse when a substantial group of steroid cheaters were caught.
"When the network was exposed publicly through a group of cheaters, the clustered community as a whole broke down," Krebs says. It had nothing to do with the drug laws enacted by the league after the scandals.
According to Krebs, this insight that a social network creates a pseudo-truth that overrides real, objective truth, can help explain why pack mentality dominates the web.
Using the current election as a model, Krebs says that the internet does not bring people with different ideas together. Instead, people seek out groups with similar ideologies, which makes them less prone to objective, flexible thinking. And no matter how extreme the idea, there's someone out there on the web who will build a forum around it.
Psychological research has shown that when people find their "political mirrors," they immediately build clusters around their ideas. This is why politicians' use of confrontational language like, "You're either with us, or with the terrorists," seems to work.
But Krebs sees the positive side of social networks as well. He believes that serious analysis of networks can be used constructively from the outside. The key, he says, is identifying the strong individuals or groups that can lead to group-thinking shifts.
For example, analyzing the rise of the iPod can be used by other companies to chip away at Apple's dominance.
When Apple released the iPod, there were other MP3 players with better audio or a cheaper price. But Apple created a network by connecting groups through an easy operating system and with marketing.
It's taken close to seven years for other electronics companies to catch up to Apple, but the aspirational ads, easy-to-use mantra and beautiful hardware design have all crept into rival companies' strategies.
In the immediate future, Krebs sees social networks facing a decidedly human problem. They need to find a compromise between the seemingly infinite number of network connections and the limited interaction capacity of human beings. After all, a person doesn't need to remember every person he's ever met, unless he is Bill Clinton.
Ultimately, the answer could lie in software that analyzes our individual social networks and makes them more efficient and valuable. Perhaps in order to understand our own connected lives, we might all have to become our own personal Valdis Krebs.Original here