I was struck by Nicholas Carr’s post about Wikipedia dominating search results. Try it for yourself – google a few topics off the top of your head. Odds are the first page of results will include a link to Wikipedia.

I was surprised by this at first – how has such centralization risen out of the vastness of the Web? But upon reflection, it seems to me a natural consequence of increasing returns.

Diminishing Returns

If you’re not familiar with increasing returns, its an economic theory used to model knowledge based economies. Traditional economics is based on diminishing returns, where each additional unit of a good or service has less value than the one that preceded it. Say you build cars – as you build more and more cars your costs will increase – raw materials will become more expensive, labor costs will go up, you’ll have to buy land to build new factories, etc. At some point its not worth your time to build new cars. Diminishing returns is a powerful model for describing the part of the economy that deals with rival goods – goods that can only be consumed by one person (if I buy the last red car on the lot you cannot).

Increasing Returns and Making Gorillas

In contrast, with increasing returns the value of a good or service increases as more people use it. This causes positive feedback mechanisms to kick-in that reinforce the use of the good or service to the detriment of other goods or services. The end result is that a market becomes dominated by a single good or service.

Increasing returns are used to model knowledge-based markets which are based on non-rivals goods (goods that many people can share). The software industry is full of examples, including Microsoft (Windows and Office), Oracle (databases), SAP (ERP software), etc. Geoffrey Moore noted the dominance of these firms, which he called gorillas, in his 1998 book The Gorilla Game.

The Web is also full of examples – Yahoo, Amazon, Google, etc. And for business that can leverage network effects , where customers are enabled to directly interact with each other, growth can be truly spectacular as witnessed by EBay, CraigsList,
MySpace and Skype. And now Wikipedia.

What is surprising, at least to me, is that open source methodologies can create gorillas. In the software world, you have to look no further than Bind or Apache. But it also happens online – two of the organizations mentioned above, CraigsList and Wikipedia, have been built by the users themselves.

As users entered in more and more content, they were able to attract more and more users. Those users in turn created more content – creating a powerful feedback mechanism that catapulted CraigsList and Wikipedia in the very top tier of web sites.

Dethroning Gorillas

Dethroning a gorilla is hard. To do it you have to offer a product that is vastly superior, otherwise there is no hope in convincing users to pay the costs of changing. And that usually means you have to leverage a technology revolution. For instance, Microsoft beating out IBM by betting on the PC, Microsoft beating out WordPerfect and Lotus by betting on Windows, Google beating out Microsoft (at least online) by betting on the Web, CraigsList beating out newspaper classifieds by using the Web, etc.

But open source provides a second surprise here. If you can’t ride a technology revolution, then open source appears to be the only viable way of attacking a gorilla. The obvious example is Linux versus Windows, but others abound – MySQL/Postgresql versus Oracle, CVS/SVN versus a slew of commercial products,
JBoss versus WebSphere, Open Office versus Microsoft Office, etc. An established
gorilla can crush commercial competitors by any number of means – undercutting them on price, colluding against them, copying their functionality or just buying them outright.

But those techniques don’t work against open source projects. That means an open source project has as much time as it needs to establish itself, get a few users, and start leveraging increasing returns. Which leads to an interesting question – over a long period of time, can a commercial entity compete against open source projects? More concretely, can Microsoft maintain the domination of Windows and Office over the next ten years assuming that some technological revolution doesn’t come along and make the whole experiment moot. If the answer is no, its strikes me that someone has an awfully interesting Economics thesis to write in the future.