I took flak recently concerning a blog post that said that Red Hat (RHT) was no different than IBM, Microsoft (MSFT), Oracle (ORCL), and most other enterprise software suppliers in terms of business model, despite statements to the contrary by Red Hat CEO James Whitehurst. My blog post was not a criticism of Red Hat, just an observation.
Most of the flak was knee-jerk open-source-zealot stuff with no IT investment research implications. But the comments got me thinking back to a more than three-year-old IT investment research project in which a client asked me to provide various investment-related characteristics of the then “leading” open source projects. In doing the research I could define “leading” any way I wanted to but I knew the client was looking for ways to cash in on the then imminent boom in “open-source-software vendor” IPOs. Therefore some of the most important characteristics were that the “open source software” was not managed by a company that was already publicly held and that the software participated in a part of the software market characterized by growth and vitality.
I identified about 100 projects to look at as part of a first cut. (During 2007, non-client-protected versions of this research showed up on ebizQ.net, research 2.0, and my own web site and the original sanitized research is still sitting out on the web at slideshare.net.) I had to tell the client that he was out of luck on both major fronts. I hope he took my findings to heart although I never know if clients accept the research except in the general sense that they come back for more.
The problem for so-called open source software vendors was that most of their projects were in un-dynamic, commoditized sectors of the software market and worse still, the leading enterprise software suppliers had already co-opted the concept as of early 2007 (Microsoft had not yet announced its then-new position vis a vis open source terms and conditions but it was clear that they were about to do so). The way I expressed it at the time is “There is no open source market.” I later added “There is no such thing as an open source supplier.”
Of course only a few open source “products” are ever intended to be productized in the commercial sense of that word. Most are one-off activities in support of a specific enterprise’s specific need. Estimates at the time said there were as many as 100,000 open source “products” but Optaros had identified about 300 as enterprise ready. As an indicator of what happened in the intervening three years, Optaros now seems to downplay its open-source-culture roots on its web site although it still maintains the enterprise open source directory which I used to cross-check my work at the time.
As mentioned there were about 100 projects in my original study and the 2007 table referenced above boiled that down to about 50 I saw as having some chance of going public some day. Compiere, Greenplum, Logicblaze (and then Iona, which acquired Logicblaze before itself being acquired by Progress (PRGS)), MySQL, Xensource, and Zimbra have all been acquired. Canonical, Jaspersoft (backed by Red Hat and SAP), OpenBravo, Pentaho, Sugar and a few others are still going through the motions. And I still see possibilities for Alfresco, Acquia, whoever is trying to monetize Ruby, and a few others that were not on my original list because they didn’t exist yet.
But Red Hat -- which of course was already public -- is the exception that proves the rule. In fact it has made a few of its own acquisitions since then, including acquiring closed-source products and re-releasing them under open source terms and conditions. Based on the table that you can look up on slideshare.net, I find that Red Hat is in fact the “last man standing.” One general rule of market dynamics is that if there is only one participant in a market it is not a market. Put the idea of an open source market (or business model or supplier) on the same shelf as the open source desktop.
Disclosure: no financial interest in companies mentioned although I have done recent consulting work for Red Hat.