Iraq and the Long Tail of conflict
If you want the executive summary: the principle behind most of the web's biggest success stories, such as Amazon, Google, Netflix is that real-world ventures must limit their inventory to the top-selling items, and their profit is concentrated there, while online ventures can have vastly larger inventories, and so the bulk of their profit shifts instead to the huge catalog of less-popular products. It springs from the statistical observation of a Pareto Distribution: there is equal volume under, for example, the first 10% of the graph as the remaining 90% of the graph*. Most companies have traditionally concentrated on the first 10%, but new companies are being enormously successful by finding ways to serve the remaining 90%.
One venture that's been pretty successful recently is international terrorism, and it struck me the other day that it has a lot in common with all those trendy web companies out there. In fact: Terrorism is long tail war.
America's army is inflicting an awful lot of damage, it's true, by making enormous expenditures in lives (and also money) all in a few places: Afghanistan, briefly, and then Iraq. Al-Qaeda, on the other hand, is also spending enormous sums of money, but it's doing it by investing in very large numbers of much smaller attacks. September 11th got all the press, but it was the largest thing they've ever done -- and it still involved only 19 people and at most a few hundred thousand dollars compared to the USA's $81 billion spent on Iraq. Everything else has been much smaller: 6 bombers in Madrid, 4 in London (not counting the copycats), a few in Bali, the Nigerian embassy, the USS cole. Small numbers all, but lots of them, adding up to an enormous psychological impact. The Americans won the battle in Iraq, but calling a victor in the War On Terror™ is very tricky -- in terms of hearts and minds, Islamic Extremism and Western Democracy (in as much as either Al-Qaeda or the United States can claim to genuinely represent those groups) seem about evenly matched, globally.
But the parallels betwen Al-Qaeda and the new economy don't end there:
- It's distributed. New York, Florida, London, Madrid, Bali... while definitely they have a concentration of support in the middle east, AQ has repeatedly shown after the destruction of the Taliban and the capture of many of its higher-ranked figures that decapitation strategies are not effective: the network heals itself.
- It's peer-to-peer. Nothing is more peer-to-peer than a radical who recruits his friends to the cause, who recruit their friends, and so on. Even with a very low success rate of converting friends into militant extremists, the geometric nature of the attempts to expand has ensured that support for AQ has grown in recent years.
- It makes heavy use of the Internet. Statements and propaganda are released over the Internet and distributed for AQ worldwide by the unwitting cooperation of millions of curious readers. Videos of beheadings are worming their way through file-sharing networks as we speak. All this at zero cost, while the Voice of America fails to reach an audience a tenth as large.
- It uses network effects. The more outrages are committed in the name of AQ, the greater its reputation. The greater its reputation, the more impressionable, disaffected young people are persuaded that it is a cause worth fighting for. That's an effect eBay would recognize from an altogether less gruesome sphere of operations.
And by extension from these observations, one can conclude the same things about AQ's brand of militancy that we can about other web companies: it will grow, it will prosper, it is impossible to eradicate -- and it is the future. I've never more wished I was wrong, but these conclusions seem inescapable.
Update: there's no such thing as an original thought on the Internet. John Robb has reached similar conclusions, and has some interesting data to back himself up, while Nature has a report on a statistical study that shows wars follow the power law (one of many other names for a Pareto Distribution).
* These percentages are examples; in reality they vary from case to case