Comment: The Impact of Metcalfe's Law

There is another "law" apart from Moore's Law which has an impact on the progress of computing both in business and in general. This is Metcalfe's Law.

The "law" originated with Bob Metcalfe, the founder of 3Com and the designer of the Ethernet protocol for computer networks ? work that he did while at Xerox Corporation. It is believed that Metcalfe invented his law as a means of encouraging companies to network computers together.

In any event the Law states roughly that:

"The 'value' of a computer network is proportional to the square of the number of computers in the network."

Although it is sometimes quoted as:

"The usefulness of a network is proportional to the square of the number of users."

In a sense Metcalfe wasn't saying anything profound. Mathematically, if you put a number of dots on a piece of paper, then the number of lines you can draw that connect any two of them is proportional to the square of the number.

This had been known for centuries. All that Metcalfe was doing was applying this idea to connected computers (or in the second formulation, users connected through computers).

We can refer to the telephone network as an example. If the network connects 10 people, then 81 different calls can be made (A calls B, B calls A, A calls C, etc.). But if the network connects 100 people, 9801 calls can be made.

As the number of people increases, the number of calls explodes in an exponential fashion. And if the calls are valuable, then the value explodes in a similar manner.

This Law has been used to explain the phenomenal explosion of the internet. There seems to be a phenomenon of critical mass in this. The internet began its rapid growth in the 1993/1994 time period. However, even then there were about 2.5 million host computers on the network.

By 1997 this had multiplied by 10, and now there are estimated to be about 100 million host computers on the network.

Of course, the internet was actually initiated by the United States Defense Advanced Research Project Agency (ARPA) with ARPANET in 1969, so it took 24 years for the critical mass to be achieved that led to a pervasive computer network.

The same was true of the telephone: the first telephone exchange was installed in San Francisco in 1878, but it required operators to make connections. It was not until 1931 that the telephone had a dial added and mechanical switching cut the cost of operators and extended the reach of the system. That was when telephony exploded.

For computer networks to explode, a similar automatic addressing protocol (TCP/IP) and a large routing network were required. By 1993 it was being delivered.

OK, so Metcalfe's Law may be a correct observation, but what insight does it give us? Can we use it? The answer to this is "yes", as long as we look a little deeper than simple numbers.

In terms of simple numbers, for example, in 2000/2001 there were about 120 million new users bringing the total population figures up to 327 million. If the simple statement of Metcalfe's Law holds true, then the value of the internet increased by 250 per cent.

The figures for 2001/2002 are similar, with about 175 million new users (a total of 502 million) implying a value increase of 235 per cent in the year. However, if we look at the figures for ecommerce growth, they are in the 60 to 70 per cent range with a current estimated value of about $600 billion.

This is an impressive growth in value (and an impressive total) but much lower than suggested by Metcalfe's Law. The clear reason for this is that internet users are not all internet shoppers. They use the internet for many reasons, from email and chat to free game playing. The reality is that there are subgroups within the internet user population who contribute to the critical mass of any particular kind of usage.

In some areas, growth is astronomical. For example, the FriendsReunited site attracted about four million new users from scratch in the space of a year. It was an idea whose time had come ? setting up a hub that old school and university friends could use to contact each other. There were enough internet users to support the idea long before anyone had the idea.

But the site did not have a well defined value model, so the revenue it generates is not yet as phenomenal as the traffic.

Other widely applicable ideas, such as the Hotmail idea (free email if you tolerate some adverts) were similarly successful at a faster growth rate than the internet itself.

Many other ideas are still below critical mass. Supply chain integration, for example, still waits for the right kind of application-to-application connection. And this is the big point about Metcalfe's Law. Undoubtedly connectivity multiplies value by the collaboration it enables ? but it only starts to tell when you get effective collaborative activity.

It is the basic value of the collaborative activity (and its cost) that determines when take-off may occur, and it can do so with very small numbers if the value is high enough.

Alternatively, apparently large numbers can be misleading. So the telephone may connect, say, 1,000 people enabling 998,001 different calls (which is a large number), but in reality each person is only likely to phone a handful of others so the number of useful calls it enables is much lower. Although the curve is still exponential, critical mass may take a long time to get to, as indeed it did with telephones.

What Metcalfe's Law is telling us is that there are many good business opportunities still waiting for the right enabler, which might be the right technology or the right business idea or the right combination of the two.

The internet population is now so high that when a good idea is implemented, Metcalfe's Law can make it rise like a rocket.

Robin Bloor is founder and leading light at Bloor Research

Like this story? Spread the news by clicking below:

 del.icio.us     Digg this     reddit!