Years ago, I remember pontificating that open source would never touch the application market. Narrowly viewing open source through the prism of the day, I said things like this (to John Koenig at IT Managers Journal):

To date, enterprise-focused open source has been best in those areas of the software stack that have a broad (approaching universal) user population. Things like operating systems and browsers. When you have a population with the aptitude and interest in improving something (like the Linux kernel), there's simply no better way to develop it -- let users develop to what they want.

Open source works less well in applications that have a smaller niche of users. It's not that open source development communities can't do applications because we find FireFox, OpenOffice, and others, or that they can't do vertical market applications, because we find an example in Medsphere's OpenVistA. But as a general rule, this is not their strength. Yet.

And this was when I was starting to hedge. Had he asked me a year before this - say, 2004 - I would have denied open source's relevance beyond core infrastructure.

But then a funny thing happened to open source. Capitalism. Money. And suddenly, business models started to be discovered that would allow companies to form communities (or leverage existing ones) to monetize open source. In middleware. In applications. Everywhere.

As I was pedaling home from a bike ride with Bryce tonight (during which we talked about innovative open source companies that I think he should invest in :-), it struck me that open source is quickly entering the next phase of commercial adoption. The phases roughly look like this:

1. Phase 1: The Red Hat years. Characterized by taking existing open source projects and building a support business around them. JBoss, Red Hat, SUSE, MySQL, and others fit this mold (though some have changed over the years). Tended to attack an existing, large market.

2. Phase 2: The SugarCRM years. Characterized by building an open source community to serve a preconceived corporate mission. Tended to attack an existing, large market. "The open source alternative for...Big Market X." SugarCRM, MuleSource, JasperSoft, Alfresco, Pentaho, Intalio, etc. fit this mold.

3. Phase 3: The XXXX years. Characterized by building an open source community to serve a preconceived corporate mission, but for new, emerging markets. DimDim, Loopfuse, Digium, SocialText, etc. fit this mold.

It's this last one that I think bodes so well for open source's commercial future. These companies are taking open source into still nascent markets. In fact, that's the most interesting thing in this trend: proprietary vendors are getting a shorter honeymoon before open source becomes a disruptive force in their markets. In fact, I know of a few open source companies that are creating completely new markets (one in particular, but I promised the VC hunting them that I wouldn't mention them...yet :-).

There are only so many large, existing markets to cannibalize. The next decade goes to the innovators in new markets, which will largely be shaped by open source, not traditional, proprietary software.

Not that it's solely about open source, of course. SaaS and Web 2.0 companies are on the rampage, too. Basically, while traditional proprietary software companies will undoubtedly milk their monopoly rents for a long time (my alma mater, Novell, continues to pull in hundreds of millions of dollars each year in maintenance on licenses sold at the dawn of the IT industry), increasing piles of cash are going to the technology and business model innovators.

More than enough to sustain many, many new and innovative open source ventures.

Posted by Matt Asay on April 24, 2007 08:18 PM