March 18, 2006

...Learn TDD with Codemanship

Lighting Fires: The Science Of Megatrends

After perhaps one beer too many in my local pub last night, my non-programming neighour and I got on to the topic of why some things become worldwide trends, and other perfectly good things are doomed to obscurity. Take Riche Kotzen, for example. Richie Kotzen is disgustingly talented, his material is catchy and commercial, and he's nowhere near as bad-looking as Phil Collins or Eric Clapton. So why is it odds on that you've heard of them, but you've never heard of Richie Kotzen? Why didn't his "ship come in" like theirs did?

I bet you can think of hundreds of examples of things that probably deserved to become megatrends, but fizzled out while seemingly random choices went on to global dominance. We see it in films, TV, music, food, fashion, drink, politics, software - you name it, you see the same thing happening.

The mechanism behind megatrends is the same mechanism behind forest fires. Imagine this simplistic simulation of a forest fire:

Our virtual forest is a grid of squares. Trees are randomly dispersed among these squares. At random, we drop one match onto one square. If that square contains a tree, the tree will burn. There is a chance that the fire will spread to the adjacent squares, but only if there is a tree in them. Let's make it really simple and just flip a coin. If there's a tree in a square next to a burning tree, then there's a 50:50 chance that the fire will spread to that square.

The chances of the fire spreading to all four adjoining trees as in the above scenario are 1/(2 * 2 * 2 * 2) = 1/16.

In the next scenario, the chances of the fire spreading this far are 1/512. The chances of the whole forest burning are about 1:30,000,000 - vanishingly small, I'm sure you'd agree. and that's just a tiny toy forest, with pretty good odds on a fire spreading from tree to tree.

Now imagine instead of a forest it is a community. Instead of trees standing next to each other, we have people who communicate with each other (a network of friends and family). We play a new pop record to one random person in the network. That person buys the record. There's now a chance that they will tell people they're directly connected to about the record. What are the odds that every person in their network of friends and family will buy the record? What are the chances of their friends' friends all buying the record? What are the chances of the whole community buying the record - just from playing the record to one person.

As we've seen, even with a tiny toy network with less than 64 people, the chances of everybody in that network buying the record are vanishingly small. Of course, you could play the record to more than one person, and that would increase your chances. But in a network of 60 million people, or 600 million, or 6 billion, that will have an enormous cost. And there's still no guarantee that the record will spread and become a megatrend.
Posted 15 years, 6 months ago on March 18, 2006