Tim O'Reilly has an interesting piece about hedge funds as software companies. His post is a riff on Paul Kedrosky's analysis showing that Renaissance Capital has a higher ratio of software developers on staff than Oracle:
It's not too much of a stretch to say that hedge funds are the new software companies. After all, they have more developers per capita than the latter, and they certainly generate more cash flow per capita.
- Oracle (56,000 empl.): 1 - 8 (one developer for every eight employees)
- Renaissance Technologies (178 empl.): 2 -3 (two developers for every three employees)
A better analog might have been an open-source company, since open-source software companies are in the business of writing software first, with an efficient distribution mechanism to get it out to customers. (SugarCRM spends vastly more on research and development, comparatively speaking, than Salesforce.com, as just one example.) But let's just stick with the general theme for a minute....
I think it's fair to say that software development has always been of huge importance to sectors that believe IT is a competitive differentiator. Financial Services is one, and Healthcare (well, the Pharmaceutical side), Aerospace, and others also have traditionally spent heavily on IT for competitive advantage.
Tim thinks this relates to Web 2.0, but I think he overstates his case:
...[K]eep in mind how the definition of a software company is changing. It used to be that a software company was a company that created software for sale. But somewhere along the way, a big part of the web 2.0 revolution was that software companies discovered a better business model: namely to use the software to deliver services that they would monetize in other ways. Renaissance has no intention of selling their software; they can make far more money using it themselves. But this is also true of Google, Amazon, EBay, Facebook, and every other giant of the Web 2.0 era.
It's also true of Wal-Mart, Whole Foods, and most companies on the planet, but this doesn't portend an end to software companies that actually sell software (or services around them, in the case of open-source companies). It has always been the case, as Eric Raymond famously pointed out, that most software is developed for use, not sale. And yet we've managed to create a massive software industry on those "edge cases" where enterprises prefer to buy rather than build. It will likely always be thus.
In fact, we may be seeing a trend that cuts against homegrown software. As more and more software is developed in an open-source fashion, whether by companies or communities, it is likely that tech-heavy industries and tech-light industries alike will turn to open source as a more efficient way to "buy software" (tech-light) and to "borrow software" (tech-heavy) and buy services around it.
In other words, the software industry may finally be providing the efficiency that its buyers have long requested. For the "tech-light," they can continue to buy software as they always have. It's just that they'll be getting the software for free and will instead get higher-value service to make it useful.
For the tech-heavy, open source offers a way to offload some of their load to outside code (like MuleSource (for an enterprise service bus) rather than building their own ESB). Even where they choose to adopt and integrate this open-source software, the likelihood is strong, speaking from experience, that they'll come back to purchase support around that software.
In summary, this means that open source should actually grow the (quasi-traditional) software industry, because it offers a compelling reason for those who prefer to build rather than buy to build (software)...and buy (service). Hedge funds aren't in the software business for the same reason that I'm not in the legal business, despite the fact that I spend a reasonable amount of time with that. It's a matter of determining one's core competence.
The real question for me in all of this is at what point Web 2.0 companies will, just like their Financial Services brethren, start to buy more software/IT? At some point it becomes inefficient to roll your own (of everything). Hence, I expect to see pullback from the trend Tim sees. When, I don't know. But it will come.