Last Modified:                                                                                                

Homepage / Publications & Opinion /

Three clicks away from eight million people
The power of social networking
Compiled at the IoD and various hotels in London over several days, polished on TG409 flying Bangkok to Singapore, and dispatched to via my hotel LAN shortly after arrival.

All civilisations have been critically dependent upon the development of communication skills, language, mathematics and abstract thinking. Ours surged forward with the arrival of the printing press, the electro-mechanical/electronic age and the information era. But quantifying the impact of networks and networking remains a bit of a challenge.

If we start at 'ground zero' with the tribe, we see a limited communication network of about one degree of freedom. That is, everyone is an equal distance from each other, and able to talk to everyone else. In early civilisations the distance between family, friends and acquaintances was, for the most part, just one step away, and the span of networks limited by the number of people met and remembered.

There was little chance of extending both the reach and the depth of networking further without some base technology beyond the messenger. The first of these was the printing press, followed later by the telegraph, telephone, radio and TV broadcast networks. However, broadcasting is heavily biased towards one-way communication of the 'one-to-many' kind - just like posters, newspapers and books.

The biggest breakthrough in one-to-one communications came with the telephone network that now links billions of people and offers a limited conferencing capability that can simultaneously link a handful of people at the same time. The number of pair-connecting combinations (~ N2/2, where n is the number of connected users) is often referred to as Metcalfe's Law.

To my mind the most important breakthrough to date came with the PC and the internet, when at last we could all connect multiple people and sources of information. This was the era of many-to-many communication that has now become mobile with the advent of browser-enabled phones capable of supporting many PC applications.

Almost by chance we stumbled upon 'exponential networking' that brought astonishing capabilities and facilities way beyond all previous networks. Reed's Law goes part way to describing this type of network but only for pairs. What we actually have is something far more powerful - of the form Nn (where n is the number of users). Here the user calls the shots - and breadth and depth of reach is everything!

If there is an Achilles heel here it is infinite growth and serendipity! This is a network out of control. It grows without contest, with information and people adding to the total without challenge. There are no checks and balances, or any measures, just unbridled growth, which we often find daunting and impossible to navigate or manage.

Perhaps it is not surprising then that a further form of network should spring out of this technology that is based on exponentiation controlled by pre-selection. Surprisingly these have arrived in the form of online social networks of the professional kind. Here the users select what and who is added when, and so growth is, in some way, controlled and the quality of the network assured.

To date all of the mathematical estimates of a network's power have assumed an egalitarian worth for every additional connection. They have concentrated on combinatorial power instead of some measure of usefulness and utility. This is not a criticism, merely an observation on the state of play of our level of understanding and sophistication of analysis.

Since 2005/6 I have been monitoring the growth and usefulness of my primary social network and I have observed some interesting features. With just 1,690 primary contacts added by me one at a time over four years, I now have access to more than eight million professionals spanning 150 countries and numerous technology, engineering and scientific disciplines. This is powerful stuff!

I am also a member of a number of special interest groups with access to online debates, conferences, tutorials, briefings and news feeds that I can observe or contribute to at will. Again, all of this has been within my evaluation, choice and control.

So how does this type of network grow? Well, it looks like a mini-internet that is only three clicks deep but with more than eight million selected contacts and a total (potential) population of more than 40 million, it could be close to exponential. The really interesting feature is that the logistics growth curve leads to a plateau of limiting interest. But as the total population of candidates grows, so do the members, and this leads to a distortion of the original static assumption model.

The accuracy of the growth model can be judged by the points logged from my own social network account, which looks like this:

What is the takeaway here? Networks and connectivity models are most powerful when we consider the weighting of the individual contacts. And so I would posit that a network of hand-picked and capable people is far more powerful than a network where we have no control of the growth.

To be connected to the entire population of the planet, and the ability to link to every device and website is nothing short of a miracle, but to achieve any level of optimisation of utility seems to be impossible. However, the social networking of people and things is the exception!

Interestingly, it is the concatenation of selection by people that continually optimises social networks and not the technology. For me this is the secret and the big advance we are currently experiencing. The ability to contact more than four billion mobile telephones and more than three billions devices with browsers is impressive - but not as professionally powerful compared to my eight million hand-picked contacts!

A measure of the advantage that social networks bring can be gleaned from the model depicted below, which is a gross simplification by the way!

If I assume a weighting per contact averaging at '1', I suspect that the equivalent weighting per contact for the entire mobile network and the internet approximates to '0'. Anyway, for sure it is extremely small!