� NYC Specialized HS experience loss of diversity | Main | Natural heuristics �

October 1, 2006

Relevance of Metcalfe's Law in Social Networks

Interesting article regarding the idea that Metcalfe's Law doesn't directly correlate Web2.0 and social networks phenomenon. For those that are not familiar, Bob Metcalfe, author of Metcalfe's Law, and inventer of Ethernet. What I find interesing is that typically people will always attempt to find fault with algorithms well after the mathematician or engineer has long passed.

I haven't seen many theories openly challenged while the author is still among the living. At its premise Metcalfe's Law is the idea that the utility of a network increases as the square of its users. In simple terms, the more people use a particular network, the greater the value of the particular network.

How is this relevent? Good question.. To take the point a step further, Metcalfe's Law was primarily associated with telephone networks. During the glory days of the switching networks, calling exchanges and operator assisted bridging networks, people recognized that as call volumes increased, the value of that switched network increased. Subsequently, Bell could begin charging more for services running out on that copper wire(ie Long distance, call waiting, *69, etc).

Bringing this idea to the 21st Century, we have to understand that true social networks are viral in nature. For instance, take a look at the wayward Napster, at its height it had millions of users sharing data. The numbers grew once people realized that they could find obscure songs and tons of music. The same could be said for the You-Tube, mySpace, and Flickr. Both have perceived value based upon the huge numbers of people that have decided take up residence and share data amongst one another. Essentially, the size of the community network, dictates the perceived value of the data shared.

In contrast, social networks which have failed (ie Friendster), were not able to provide consistent value to its community. Make no mistake, your network must be robust and capable of handling large volumes of data. Interestingly, most users don't care about the inner-workings of the network. They only care when the network is down or simply doesn't work properly and they can't share data with their pseudo-virtual friends.

So the obvious question is how does one correlate network usage to dollars? The monetization of social networks has brought Metcalfe's Law(ML) into question. Here is the debate. Opponents of ML fear is that another Internet bubble is wrongly inflated, and the relationship is not squared (V~N^2) but logarithmic V~N*log(N). Metcalfe admits that unlike Moore's Law, ML was never evaluated numerically. Nonetheless, none of ML's detractors have produced any emperical data to disprove what has largely been considered an ubiquitous concept. According to Metcalfe, ML is best associated with social networks when combined with Moore's Law. There is another aspect of this conversation that is quite interesting. The whole idea of Long Tail (statistics term for the 2% or outliers on a normally distributed population a.k.a Bell Curve), diseconomies of scale and that social networks have automated the whole 'word-of-mouth' concept.

Nobody disputes the relevancy of the work of Bob Metcalfe, in fact I believe it is more relevant today than perhaps a decade ago. Please read his rebuttal to the IEEE Spectrum article.

Om Malik also had Metcalfe on as a guest in a recent podcast .

So, what does all this mean for fledgeling social networks like gather.at and others?
Well, at the core it must be understood that the community owns the network, not the developers or its founder. The perceived value can only be realized when people begin to share and convince others to do the same. Moreover, value is not a defacto artifact of volume. I suppose one could ask, "How valuable is mySpace?" Or for that matter, "How valuable is Linked-In?"

Well, I imagine it depends upon who you ask ;)

Posted by AG at October 1, 2006 8:28 AM