A word like “critical mass” is so often used in the figurative sense, it is sometimes hard to remember its original, technical meaning: the smallest amount of fissile material needed for a sustained nuclear reaction. This was first achieved by the great physicist Enrico Fermi, in a “pile” (and that’s the technical term) of Uranium in the unlikely location of a squash court at the University of Chicago.
It turns out that different isotopes require a different mass to go critical. So Fermi needed a pile 8 metres high of unrefined Uranium, but a ball of just 10 cm (the size of a softball) of Plutonium is needed.
Cloud Network pioneers also wonder how long it takes to “go critical”. A common target in the network world is 50pp, 50 percent penetration of the target market. I prefer the term “network effect”, and the idea here is that networks ( just like a pile of uranium-235), when they get to a certain size “go critical”, and the benefits of adding a new node exceed the costs of creating and connecting it.
Let’s look at some Network Effects over the years. They all exhibit the same chicken-and-egg effect in that it makes no sense to buy a telephone (lay rails, string cables) if there are not enough people to connect to.
|Railroads||1820’s||1880’s||70 years||Tycoons, Barons|
|Electricity||1860’s||1930’s||70 years||Local Govts, co-ops|
|Telephones||1870’s||1940’s||70 years||Ma Bell|
|Oil Pipelines||1880’s||1920’s||40 years||Standard Oil|
|Credit cards||1950’s||1980’s||30 years||Visa|
|Internet||1960’s||1990’s||30 years||Fed Govt|
One fact that seems inescapable is that things are speeding up, just as we’ll, as investors are unlikely to want to wait decades for the network effect to take off. Also, the time taken to get to “criticality” depends on the cost of setting up the node and connecting to the network. Stations on a railroad are expensive, so are the tracks between them.
There’s certainly more to a Cloud Network than just scale, such as the classic cloud benefits of service levels and guaranteed delivery but scale matters. And of course there are many examples of networks that fail to go critical at all: Apple’s Ping, the Philips Video Disc, the CueCat, the 3DO games console.
How can you tell if a network has reached criticality? The mathematical definition states baldly that at criticality, the value obtained from the network is greater than or equal to the price paid for access. Not very helpful, but there’s one important thing that is needed: a sustainable business model .
A sustainable business model is in effect when the profits to the network administrator scale in proportion to the value that its subscribers receive. If you have to endlessly pay for more hardware (e.g. phone lines, wireless spectrum, racks at Amazon Data Services) as you get more members, then you don’t have a sustainable business.
There are plenty of Business Networks out there, and when one of them goes critical, it’s going to create a really big noise.