In February, 2000, the Internet bubble was just about to burst and Paris-native Marc Fleury was in Silicon Valley hitting the venture-capital circuit, trying to get funding for his fledgling open-source startup. When one VC after another told him it was a horrible business plan, he packed up his PowerPoint slides -- along with his bruised ego -- and moved into his in-law's house in Atlanta, setting out to build his startup the old fashioned way -- with his own money.
Fast-forward to 2004, and Fleury is having the last laugh. His company, JBoss, has decided to go for its first outside investors, sewing up a tidy $10 million in funding from VCs clamoring to get in on one of the hottest open-source companies in high tech. JBoss, along with a handful of others, such as MySQL, have proven that big corporate customers have an interest in free software that extends beyond Linux (see BW Online, 10/19/04, "Redemption for JBoss's Boss").
JBoss makes application-server software -- the kind of that's responsible for running mission-critical Web programs, such as JBoss client Travelocity's reservation engine. It's a market that has been dominated by BEA Systems (BEAS) and IBM (IBM), but analysts say JBoss is posing a serious threat by offering free software licenses while charging for support, maintenance, and training.
In 2004, Forrester Research surveyed 95 open-source-friendly tech companies and found 15% were already using JBoss software and 8% more were planning on it. While just 3% or so of JBoss users are paying customers, almost 6 million companies have downloaded its software.
Despite his success, Fleury is skeptical of the new generation of open-source startups now being funded by VCs. BusinessWeek Online Silicon Valley reporter Sarah Lacy caught up with Fleury on a recent trip to San Francisco to talk about making money in the world of open-source software -- and why it may not be as easy as JBoss and others have made it seem. Following are edited excerpts of their conversation:
Q: What's the state of the open-source market today?
A: Linux is the big brother. Linux has fought all the hard battles to get the keys to the car and stay up late. There's this funny [ramp] to adoption in open-source where you have a very high barrier. The perception to open-source [used to be that] it's dirty -- nice IT managers don't do that.
But once the developers started playing with it, [there] was definitely a bottom-up movement. What's going on now is the reverse. It's a top-down movement. It's really the [top executives] now that are saying, "Hey, wait a second. We've tried this open-source thing in production, and it works. The support is supposedly there. Let's make this the corporate standard."
Q: When did you see that change?
A: About a year ago. We called it the "roll it out" phase as opposed to "trying it out." Web servers or file servers [where open-source software was already being used] aren't mission-critical. Application servers [are].
We power the reservation engine for Travelocity. We monitor the power grid here at California ISO [a nonprofit organization that regulates the flow of electricity in California]. So these are completely mission-critical environments. We make our money when the large [corporate customers] want to standardize and they need a relationship with a vendor.
Q: Most open-source companies are following a similar business model where they give the software away for free and charge for service and support. Are there limits to that business model?
A: I'm not sure the model applies to other regions of software. The origin of open-source was definitely nonprofit, right? It was very high on passion and church, but not at all with a business model behind it.
Then, in the 1990s, you had the first-generation business models, like Red Hat (RHAT). The difference between them and us is that Linux is developed independently, and then [there are companies such as Red Hat] who package it. So they all compete for the same business, and they all compete on services and have to compete on price.
It's a difficult model to differentiate.
Our generation's business model simply says: We're software vendors. We write the software and offer the service. And I know my business model will fail for other parts of software, so don't try this at home. You may get badly burned.
Q: What kinds of companies could get burned? Those offering applications, such as customer-relationship management?
A: I don't even want to talk about it, because I really haven't thought about it. But let me give you [a few] arguments why it's harder. I'm a developer. I work during the day and at night. If I'm going to pass on seeing my girlfriend or my kids, it better be some sexy software that I'm working on. Writing a financial graphics user interface on an application? That's what I do for living. The idea that I'm going to sit up at night doing it is ridiculous.
No one is going to work for free. That's the myth of open-source. [New open-source startups are] trying to buy distribution. And maybe they'll succeed, because if you give something away for free, people will come. That's the only part of the model they're replicating: Give it for free and charge for the service. The point is, do you have enough volume? If that's your model, you monetize 1% to 2% of the user base, ergo you need a pretty large user base to have scalability.
Q: How does the open-source community view VC efforts? Is there resentment to those using open source to build a for-profit company?
A: It's funny, I was at this conference the other day. And I was traveling back from London, a little bit tired, so I was going to blow the conference off. I looked at the list: executives from Microsoft (MSFT), executives from IBM, and the CEO of Novell (NOVL). It's a real audience, right?
O.K., so I go in [and start to do my] total business presentation. This guy in the front row says "You've got to stop banging on people whose motivation is something other than money." There's always a Hari Krishna in the audience: "It's illegal to make money at this. We're all garage bands, and you sold your soul to the devil for a handful of dollars." So I go, "Have you contributed anything?" and usually they say no and I stop it there.
Turns out the guy is the founder of a pretty significant chunk of Linux, so Point A goes out the door. So I say, "You are what I call amateur open-source or hobbyist open source, which is you have a job and then you do this because that's your passion." And then somebody in the audience yells "You mean amateur open source as opposed to asshole open source?"
So there's always that. It's normal. There are always a bunch of amateurs because they've never made money at it, and it kind of pisses them off that there was a way to do it.
Q: Why is it a myth that a startup will get developers to hone the product for free?
A: Think for a second, who works for free? I think it gets perpetrated because it's such a nice myth -- you would get love and peace, the old hippie dream you know? And it's mostly true, but across all of software, not just open-source, you have a pyramid of productivity. It's an art still -- a black art of creating great software.
At top of the pyramid, you have these top 2% of developers that are 10 times -- in some cases 100 times -- more productive than the rest. It's true in proprietary developments like Microsoft and true of open-source too. The value is the QA [quality-assurance testing to make sure the software works and finding and fixing bugs]. They cover more ground than we could ever test.
Putting aside the QA, there are 20 people who write the kernel, and guess what? These guys are all professionals. If you get free, you want a lot of it. If you give free, you're going to give until you're tired of giving, and that's exactly what happens in the open-source community.