What makes a successful open source company? If there's one corporation worth examining to answer this question, it's Red Hat. Since awakening the financial community to the open source phenomenon with its remarkable IPO in August of 1999, Red Hat has steadily been writing the book on how to build a Linux business. Linux Magazine's Adam Goodman, Robert McMillan, and Chris Somerville recently caught up with Red Hat CEO Matthew Szulik to discuss the lessons learned during his company's first year and a half as a public company. Szulik, a mild-mannered Bostonian who had recently shaved his head for a charity, has seen a fair bit in his first year at the helm: A dozen acquisitions, executive flight (CFO and CTO), and a stock that soared to $300/ share and then plummeted to 1/30th of that value. But in the end, he says, it's the future, not the money, that motivates him each day.
|PHOTOS © GARY WAGNER|
What makes a successful open source company? If there’s one corporation worth examining to answer this question, it’s Red Hat. Since awakening the financial community to the open source phenomenon with its remarkable IPO in August of 1999, Red Hat has steadily been writing the book on how to build a Linux business. Linux Magazine’s Adam Goodman, Robert McMillan, and Chris Somerville recently caught up with Red Hat CEO Matthew Szulik to discuss the lessons learned during his company’s first year and a half as a public company. Szulik, a mild-mannered Bostonian who had recently shaved his head for a charity, has seen a fair bit in his first year at the helm: A dozen acquisitions, executive flight (CFO and CTO), and a stock that soared to $300/ share and then plummeted to 1/30th of that value. But in the end, he says, it’s the future, not the money, that motivates him each day.
Linux Magazine: The world has sobered up since we spoke with [Red Hat Chairman] Bob Young last year. How has it felt to have your stock drop from such a height?
Matthew Szulik: When our stock hit $300 and there was wild enthusiasm, I remember having this discussion with Bob. I said, “This is one of the worst days of my career.” He asked, “Why?!!” He was having a good time and was caught up in all of this.
I had watched a group of incredibly bright young people work their butts off, deal with just incredible changes and challenges, and I thought that a wild stock price — an $18 billion market cap on the strength of a boxed product — was going to take away from all of the good work that had been accomplished, and most importantly, from the bigger challenges that lay ahead of us.
All the Microsoft hoopla made good theater, but Microsoft is a company with $26 billion in cash in the bank. They’re an incredible American success story, and with all due…actually with a lack of respect for their practices, I knew that the hard work was before us. We had yet to be challenged.
With all of that in mind, I didn’t see the wild enthusiasm. I wasn’t one of the cheerleaders. I was saying, we’ve got to get the follow-on offering out; let’s move. We’ve got to get the Cygnus acquisition done; let’s go. We’ve just got to keep scaling and keep driving the company to perform as rapidly as possible.
LM: So if you could do it again, would you go public when you did?
MS: I wouldn’t change a thing.
LM: But you thought it was a bad thing when the stock hit 300?
MS: Because of what’s happening now. Today we’re sitting here and the stock’s got a $1.3 billion market cap and there are probably people writing stories that I should be slicing my wrists. When you take a look at the complete picture of who we’re choosing to compete against, when you look at our financial position — we’ve improved the gross margins of the business from 46 percent 16 months ago to 60 percent in this last quarter — we’ve got $320 million in the bank; we’re on the verge of being profitable.
We’ve got great relationships with IBM and Compaq and Dell and others.
Why do people want to make me feel lousy? We have introduced 16 new lines of revenue into our company. We’ve been able to hire great people. And I think one of the most important elements is that we have been true to our commitment to the open source model. The questions used to be, “When are they going to go to the dark side? When are they going to start adding proprietary layers? When are they going to flip the model and become a proprietary company?” Heck, we cancelled revenue streams that had a proprietary component in them so we could continue to keep this vision and keep this momentum alive.
LM: What revenue streams did you cancel?
MS: When we bought Cygnus, they had a 75-person development effort working on a proprietary IDE (integrated development environment) for the semiconductor simulation market. It was called the Churchill project. We cancelled it.
LM: How did your employees react?
MS: It made a lot of people very angry. People had devoted a year of their lives to developing this technology.
LM: What led you to acquire Cygnus?
MS: Let me describe some of the conventional thinking we were hearing at the time. We were hearing from enterprise customers that in order to be competitive with Microsoft and others, we needed to have a platform with a consistent API [Application Program Interface] from server to handheld device. If you’re really going to be embraced strategically by the Global 2,000 — especially as wireless and other technologies loom as part of their five-year IT plans — you need to have an answer to that, and we didn’t at the time. So that was issue number one.
Second of all, there was a very popular question a year ago. “What about the desktop? Why don’t you go out and buy Corel?” But Microsoft had 92 percent of that category. We didn’t think it was a very forward-looking category. A lot of smarter people than us fought that war and lost. As a Bostonian, I had watched Lotus’s franchise crumble under the weight of trying to compete with Microsoft. We made a strategic decision at that time to move forward and not compete on the existing playing field. We would get into the handheld and device marketplace. We thought that that was where the next frontier would be. And all of this gets tied together with the Red Hat Network. That is the cornerstone of the whole company.
LM: How is what you’re doing with Cygnus different from what you originally expected?
MS: The conventional rule of thumb is that 65 to 70 percent of acquisitions fail. Cygnus was 11 years old when we acquired them and had gone through multiple changes over its history. What we saw was great engineering talent; the core engineering leadership of the company had a strong commitment to open source. Most importantly, they had great customer references. They knew how to build sophisticated projects and get them delivered on time and on plan. That was a discipline that we thought was important for our company. When we acquired them, we were just completing the follow-on offering and we were also a very young company. So I think we’ve learned a lot about how to integrate companies. From that perspective it was a great learning experience for us. And we’ve learned to do better with our subsequent acquisitions.
LM: How has it played out then? Are there any customers that you can talk about that have been able to take advantage of your competencies in this wide range of offerings?
MS: Oh, definitely. I don’t think the people that we’ve partnered with would see us as a complete solution provider if we didn’t have an answer to the portable device story. If we didn’t have that, we would have given them a reason to move away from the Red Hat relationship, and we would have created a window of opportunity for Microsoft to drive their Windows CE technology.
LM: You see the Red Hat Network as the cornerstone of your model. How does everything fit together?
MS: As a privately held company, we made a decision not to recognize 100 percent of the revenue from the box sales we were recording. We recognized those sales as a subscription because we wanted to have a continuous relationship with the customer. We didn’t want to just sell them a box and say, “See you 12 months later,” because our design cycle was every three to four months. And at that time we said, “The relationship to have with the customer is one that is online, that is network-based, because that will, hopefully, enable the true direct relationship with the customer — provide continuous innovation, continuous updates, and do that on a monthly subscription basis.”
If you look at that, you’ll see that we chose to start a subscription relationship at the expense of perhaps some valuation upside. If you look at our balance sheet, you’ll see a line of deferred revenue there of $12-13 million, which unfortunately we don’t get a whole lot of credit for from [Wall Street]. This was the beginning of our strategy for getting into the subscription model — to build and monetize a relationship with the millions of people who have touched the Red Hat brand. The Red Hat brand could be expanded to include the distribution of not just our technologies, but other technologies as well.
LM: Bob Young has said that he sees Red Hat as the Wal-Mart of technologies — that you can buy all these open source technologies through Red Hat.
MS: He never said that to me, but I’m glad he thinks that way.
LM: He has a metaphor for almost every occasion.
MS: [laughing] He’s a piece of work.
LM: Do you still call him “Hollywood?”
MS: No, “The Great Communicator.”
Red Hat: ASP
LM: So do you guys see a way to derive a revenue stream from every person that has a cell phone that’s using Red Hat?
MS: I don’t think cell phones are the answer. Let’s use Ericsson and the Web Pad as an example — I think that’s a realistic one. My generation was raised with planned obsolescence: “We’re going to build something and it’s going to be obsolete in 18 months.” With consumable devices like a cell phone, that will continue to be the case.
You’ll throw it away, and new functionality will be added. But when we have spoken to executives at Nokia, Ericsson, Sony, and others, their cost of customer acquisition is so high that the more they can enable that device to stay in your hand and use the network to upgrade that, the more they’re able to continue and keep that brand relationship with you. It’s not the manufacturing cost that they’re worried about; it’s the cost of switching.
In speaking to Ericsson about the Web Pad, we would like very much to enable the enhancement of that device, or the extension of new services or new value-add to it. We would want to be involved whether it’s through Ericsson.com, which is being back-ended by Red Hat, or directly through Redhat.com, where we can introduce new services and new value-adds and participate in the revenue stream on the incremental dollar amount that Ericsson will be charging their customers.
LM: Why do you think the subscription model is the right one for Red Hat?
MS: I think two things make subscription work for Red Hat. One is our revenue-recognition practice, because for a company that’s been basing their business model and their financial reporting on the sale of software licenses — to go back and restate that as a subscription, especially as a public company, is not easy.
The second is that one of the great benefits of open source is the continuous improvement of the product. So when a customer buys a subscription — if the release is every two to three years like Solaris, or four years like Microsoft — where’s the value? In the case of Red Hat, we produce two to three versions every 14 months.
LM: So how fast is the Red Hat Network growing? Can you tell us how many subscribers you have?
MS: I believe when I left the office, we had in excess of two million updates of the product, and I think it’s an evolving activity for the company. We have stated publicly that this upcoming fiscal year we don’t expect it to have a major impact on our revenue model.
LM: So who is paying for the Red Hat Network now?
MS: It’s a free service right now.
LM: How will it make money?
MS: Shortly, through registration. Are we smart enough to create enough value that people would want to pay for that service? This is a strategic move for our business, because it’s not simply creating an online service. There will be many other strategic initiatives related to how we will reposition our company to support this. And it’s not trivial. When you start telling people that you’re going to manage their 5,000 or 10,000 server environment, and that they’re going to be able to have a complete snapshot of their entire architecture, and that you’re going to maintain it and manage everything from security to remote systems management…I think we’re learning a lot.
So we have a number of very large strategic customers that we’re working with right now. We’ve been able to learn a lot and we continue to learn from them how to do this responsibly, making sure we don’t shoot ourselves in the foot by over-hyping it.
LM: So you see yourself becoming an application service provider (ASP)?
More than Money-Motivated
LM: We’ve been talking a lot lately about the relationship between commercial and community interests. I have a hard time coming up with the names of any developers who have come into the Linux community — say from IBM — and become well-known leaders on the basis of their work. Why do you think that is?
MS: Well, I have a funny story. We hired a very, very bright developer — a very smart young man. And he came in after his 90th day at Red Hat and said, “I really believe in the technical direction of open source,” but it turned out that even though he had the technical capability to make contributions, he found that access to information — and to the leadership of the project that he was working on — was highly fragmented and it was difficult to get his ideas heard.
It’s no different from when you get married and have a child. That represents change. When you have another child, that represents more change. I think change is hard for some people. What was a somewhat quiet, very technically-driven community of people using technology for their own purposes is now mushrooming into a commercial industry. And I think what has to be taken into consideration is; A) What are the long-term benefits of expanding out the community? We need to recognize that the community is going to change. There are social and market forces which will cause that to happen, whether we like it or not. And B) How do you make sure that the technology advances along the principles and the values of the original open source community?
LM: So what comes first for you? What if your board of directors decided that Red Hat had to begin selling proprietary software?
MS: [laughing] What is the upside for me to answer that question? I think it’s a personal choice. For me it’s not something I would want to participate in. Selling proprietary software puts me in no different position than all kinds of companies selling widgets out there. It’s the commitment to open source that differentiates us.
My personal goal was never money. And I think when I came to Red Hat there was a lot of suspicion. I remember sitting down with Mark Ewing and him saying, “you are going to change this into a capitalist organization.”
I believe that what’s at stake right now is the whole notion of how information will be available for the next generation of citizens, whether they be in Germany or Japan or the United States. In the United States right now, a lot of that policy is being formulated and being discussed without both parties sitting at the table. With [Microsoft VP] Jim Allchin’s recent comments to US Senators, the Senate only got one viewpoint. There was no open source representative to challenge his statements and his opinion. There are policy decisions being made on the state and federal levels. If the open source community is not able to establish a broader platform, then we will be forced to live with decisions that are being made.
As an example, Red Hat spends a fairly large sum of money now to battle UCITA (Uniform Computer Information Transactions Act). But that’s only one small element. There are many more. And the proprietary vendors are having a relatively easy time with all of their lobbying efforts because the open source community has not been able to galvanize itself to insure that our views are properly represented in Congress.
LM: There is a sense in the open source community that the political process has been irrelevant to the development of important technologies like the Internet. Are you in dialogue with people in the open source community about this?
MS: I have tried to broach this topic with other CEOs in the open source industry and it didn’t go too far.
LM: Why not? What did they want to talk about?
MS: Red Hat’s dominance of the distribution market, which I think is completely irrelevant.
Robert McMillan is editor at large with Linux Magazine. He can be reached at firstname.lastname@example.org.