As NewsForge's in-house business writer I've had a problem for over a
year -- what to say about the biggest business story in Open Source. The biggest story wasn't that Red Hat got a little traction in the market and was being treated like a distinct OS by most third-party developers. It wasn't that Caldera, the largest Linux distributor in terms of revenue, was headed by an Open Source agnostic. Although the news that IBM
had undergone a Linux conversion was a big story, it wasn't nearly as compelling as news coming from a little closer to home. Outside of NewsForge and other OSDN sites, anyone would tell you that the most compelling business story in Linux, if not all of IT, has been the incredible shrinking market cap of NewsForge's (and OSDN's) corporate owner.
VA Linux, the greatest opening hit in stock market history, is now worth one half of one percent of its peak market capitalization. I have been looking for comparisons for the last couple of weeks, but as far as I can determine, this may be the largest value crash of any publicly traded company with a continuing business since 1929. That makes it a legitimate news story, if not exactly a comfortable one to cover.
The litany of VA's problems may be familiar to some of you. The company is getting out of its original core business. While other Linux-centered hardware companies like Cobalt were snapped up by major hardware developers for 40 to 50 times revenue, VA couldn't find a buyer for its equipment business. While other floundering Linux companies were showing revenue growth, VA's declined
precipitously, even before the company finally abandoned its hardware business. Analyst Amit Chopra, of Credit Suisse First Boston, said that VA had simply been "overwhelmed by traditional server vendors." Even before the company
wrote off an astounding quarter billion dollars in special charges and goodwill against revenues of $16 million, VA's $0.64 loss per share was double the street estimates among the few brokerage houses still following the stock.
It could have been worse, had some wire service stories been accurate. Investors who subscribed to Factiva were told that VA's loss was $ 2.9 billion, before the wire service corrected its release 17 minutes later.
There was more confusion. Reporters who had managed to get into the announcement teleconference reported that VA would rely on sales of proprietary software in the future. So much for Open Source. The next day Eric S. Raymond added his own spin, as he distributed a message suggesting that VA's commitment to proprietary enhancements amounted to little more than an interim marketing tactic. Raymond said, "There is less here than meets the eye. This is a change in tactics, not strategy."
Raymond suggested that VA's problems were due in part to conservative IT managers, and that the only way to get these managers to buy software instead of simply downloading it and using it required "hanging some proprietary tinsel off the product." Raymond suggested strongly that VA's strategy and value proposition continued to be focused on "the service contract."
Unfortunately, there is little evidence to suggest that "conservatism" or a corporate reluctance to embrace Linux had much to do with VA's problems. On the contrary, Linux is rapidly being brought into enterprise environments, at least experimentally. According to Carl Howe of Forrester Research, better than half of the largest U.S. corporations
have adopted Open Source for one or more projects in the last year, an adoption rate that exceeds the PC revolution in its earliest days.
The second problem with Raymond's analysis is that it assumes enterprise adopters are being "penny wise and pound foolish" as they spend resources training their staffs rather than renting expertise from VA. Perhaps, but the trend has been that any time a company begins to embrace a core technology it will invest in human capital to support it. That means developing internal expertise rather than outsourcing it. The expertise is coming online. Nearly every major university teaches OS development using Linux. Most new computer science majors are graduating with Linux and Java development skills.
Successful IT services companies have learned that the key to staying in business is developing a practice that focused on training and business process management rather than just developing widgets. Both IBM's consulting business and Microsoft Consulting Services (MCS) are good examples of this business model. While no competitor seriously questions VA's technical chops, IBM's and Microsoft's services learned that they needed to talk management strategy as well.
VA's problems were far more straightforward, and they were tied directly to the fate of the dot-com revolution. It was obvious from the beginning that VA would never be able to compete in established markets against the IBMs and Dells of the world. But emerging ISPs and dot-com entrepreneurs were fair game. Most traditional vendors had no
channels targeting them. In addition, many of these smaller companies were headed by teams that were looking for bargains, and who hated Microsoft products for reasons that were equal parts ideological and performance related. It was a tempting target.
Unfortunately, the dot-com boom was well on its way to becoming a bust even before VA floated the idea of going public. Whatever the deficits of VA's business plan, it was brilliant compared to businesses that decided that the Internet was the ideal way to sell broccoli or Kibbles n Bits. Unfortunately for VA, by the time the company had fully
ramped up to sell to this market it had disappeared.
In any case, VA has moved on. According to CEO Larry Augustin, VA's strategic focus will center on enterprise versions of the company's SourceForge software, which lets businesses set up collaborative programming sites.
This makes a fair amount of sense.
Today, large enterprises develop new code on a 24x7 basis. Programmers from around the world may collaborate on software design and development without ever meeting each other. There's a real need for process management systems that help track and organize that collaborative process. In many ways enterprise code development parallels the process that created the Linux code base. Projects get started, then stopped, then re-started months later. IT personnel turn over. Frequently program originators are long gone by the time a project is completed. Requirements often change drastically during the development cycle.
One of the biggest attractions Open Source has for many IT managers is that open, documented code allows for broad participation in code development and supports iterative
changes. As everything is open, everything is viewable and testable. There are no hidden "gotchas" squirreled away in compiled code. Individuals working outside a "core" development team can develop and test a change on their own and post it back to that development team.
There is little question that collaborative development of this type needs an organizational framework to identify projects and their owners and to track relevant code libraries and remaining
But collaborative development requires training in a process as well as access to a software program. Where is the value? Should VA focus on morphing SourceForge into a collaborative development tool, and take on CollabNet -- and eventually Microsoft?
Or is VA missing yet another opportunity? Is there a better business to be had teaching clients how to develop their collaborative programming methods by using the "community" approach that is responsible for Linux's success in the marketplace?
It's a tough call. I'm glad I don't have to make it. I don't even like discussing it.