Saturday, February 13, 2010
Subtitle: Small Worlds and the Groundbreaking Science of Networks
I am fascinated by networks. Certainly the social networking websites currently available are revealing in a less abstract manner how far our connections are really spread. The most rewarding aspect of joining these sites was the initial 30 minutes after having signed-up when you were quickly reconnected with ex-colleagues and long-forgotten high school acquaintances. From there it is mostly downhill and I am still amazed by those who find so many non-commercial ways to occupy their time posting, tweeting and chatting away.
I spend little time on these sites and am becoming more and more skeptical that they will be able to retain their entertainment value (and massive subscriber growth) before being besieged by individuals and companies trying to push their products and services on you. A recent interview from a Pepsi marketing manager I unfortunately saw on CNN should give a clear indication of what is to come. Pepsi's desire to have a "conversation" with its customers does not bode well to those who find a large aspect of Facebook's appeal to be the limited advertisements pasted on its pages.
However, I do confess to having a soft spot for LinkedIn. It has focused on the business community from day and is now evolving into a profitable company with a clear business model. By being a site catering to business users, it is somehow more justified to find ways to drive revenue - a luxury not necessarily granted to Facebook.
LinkedIn is so great because it allows the users to really see the reach of their networks. By using the "people" search function on the site, I discover to have contacts in Madagascar, Saudi Arabia and Suriname and all of them are only one contact removed from me - meaning that they are connected to someone in my own 170-person network. 20 years ago these same networks existed but it was just much harder to visibly construct them.
The premise of Nexus is that there is scientific evidence demonstrating how the network that links the six billion of us is not completely random in nature but actually has a certain structural configuration. The author then demonstrates how such network structures are not just relevant for human acquaintanceship.
Buchanan introduces the argument by referring to a fascinating experiment conducted by the psychologist, Stanley Milgram in the 1960s. I was familiar with Milgram for the famous test that bears his name yet was unaware he was also credited what came to be know the "six-degrees of separation".
In the Six-Degrees test, Milgram mailed at random letters to people living in Omaha, Nebraska and asked them to forward the letter to a stockbroker friend of his in Boston. He provided no address for the stockbroker but instead asked them to send it to someone who they believed to be "socially closer" to him than they were. Most of the letters made it to his friend as each subsequent recipient forwarded the letter to the next person applying the same request. The most surprising aspect was that it did not take hundreds of mailings but six.
It was then the work of two other academics, Watts and Strogatz, who when further studying Milgram's test, found these six-degree connections to be neither orderly nor random but somewhere in between. These patterns took on an even greater significance when they started examining other types of networks such as electric power grids, human brain cells or the world wide web only to discover clear similarities in all of them.
Before returning to the book, let's look at an example given by one of my professors, Dr. Karen Stephenson, during my graduate studies that should help in understanding how these networks are patterned:
Think of the organization you work in. Most likely it has some type of hierarchical structure of management. Yet by looking at this pyramid, does it really show how the organization is networked together? It is true that the person at the top holds a lot of decision making power but it is certainly bound to be the case that individuals scattered throughout the ranks wield a disproportional amount of network power. These linchpins are the types of people you go to for questions, information or advice and most certainly others turn to them as well.
If every person in the organization drew lines to each person they contacted throughout the day for a question or advice and then all of these drawing were compiled together, you would see a web-like structure emerge. In this web certain people would be crucial "hubs" as they would appear as being connected to a lot of other people as well. Interestingly enough, it is probably the people towards the top of the pyramid who have the fewest connections to others in the organization.
This type of network formation exists in all the communities we interact with on a daily basis. Yet how is it that we are connected in such a short number of steps to people on the other side of the globe if we interact only with our local communities? It is by people acting as "bridges" to other groups far beyond the ones we are involved with at a local level. The exchange student you still keep in touch a couple times a year via email serves as a bridge to her entire network of friends and family back home. Lose connection with her and this whole world (even if you are in no way actively apart of it) disappears.
As stated earlier, this network layout applies to other areas of science as well. One of the more vivid examples is the development of the world wide web. Internet pages are being developed randomly and frequently. However if you look at the links between all of the billions of web pages, you find once again Internet pages acting as "hubs" and a certain order to the random development. What it also means based on the study that examined the network of the web is that even as another billion webpages are added, the number of clicks required to navigate will not change by more than one or two clicks.
These "small world networks" thus make information, electricity or gossip travel much faster than a world without them.