How Does Red Hat Make Billions of Dollars Selling Free Software?
People frequently ask me, “How does Red Hat make money selling free software?” This question comes up so often that I can’t imagine not addressing it in this book, even if it’s in an appendix. Let’s start with an analogy: why do people choose to buy a bottle of water, when the same substance is freely available in most of the world? All you need to do is open a tap to get it, just as you can go online and download, say, open source software code. Yet many people find good reasons to pay several dollars for that bottle of water or sign up for a delivery service: the convenience and portability of the bottle, the perceived safety of the water’s source, or even because people prefer the taste compared to tap water. The buying decision comes down to choice and whether the vendor is providing something of value to you.
Similarly, although it might appear that Red Hat’s business model is to sell something that is supposed to be free, our company actually is in the business of adding value to free code. Red Hat is an ever-growing collaboration of both internal and external communities of contributors who update and improve software by working together. While Red Hat first made its name in the software world by working with the Linux operating system, which may be the most widely used software technology in the entire computer industry, we now collaborate with a multitude of other open source communities such as OpenStack, JBoss, GlusterFS, and Apache Camel in ways that provide value for our customers.
For our flagship product, Red Hat Enterprise Linux, for instance, we take the latest version of Linux and dissect it so that we can test and document every modification or bug-fix that has occurred since the last release. We certify it against the thousands of hardware and software programs that enterprises typically use, so they can confidently implement Red Hat Enterprise Linux without worrying about if or how it will work with the myriad other components of their data center. In addition, we have a dedicated security response team to identify, patch, and distribute fixes for malware. Unfortunately, the internet is making the “pipes and sewers” of computing as risky as drinking from a pool of water in the middle of a swamp. Our “bottled open source” solutions provide security that simply cannot be achieved otherwise. In short, we sell our customers peace of mind that their entire system is the most stable and secure system possible on the planet. That’s something they’re willing to pay us for.
Learning from Linux
We alone don’t do all the work that our customers benefit from. We work as part of an open source community, which is much more powerful at delivering results than any one single siloed organization can.
Consider how the dynamics of the Linux community, of which Red Hat is part, operates. Created by Linus Torvalds in 1991 (he famously kicked off with a listserv note that read: “Hello everybody out there . . . I’m doing a [free] operating system . . . just a hobby, nothing professional . . .”), Linux now runs everything from the most extreme scientific computers to web servers, network devices, point of sale systems, military systems, cameras, automobiles, and mobile devices. If you’re using an Android phone, you’re using Linux.
The Linux story is fascinating from an organizational management perspective precisely because it involves a system in which literally thousands of people, including many Red Hatters, self-organize and take direction from informal leaders who emerge based on the quality of contributions they have made over time—which some community members have come to refer to as a “mega-self-organizing collective.” The Linux source code itself, some 30 million lines of programming instructions, is accessible to just about everyone out there; hence, the term “open source.” To create this code, community members came together in a decentralized manner, rather than working through a conventional top-down hierarchy, where higher-ups issued orders that were fulfilled by programmers at the lower rungs.
I am constantly amazed at the effectiveness of this self-organizing, opt-in, global community that we contribute to. This virtual army fixes bugs, heads off malevolent hackers, and collaborates on making Linux a stronger product than it would be if just a single entity tried to control it, which many people once thought was the only way to build software. Linux is modular enough to run things as diverse as refrigerators to smart phones to mainframe computers. It’s precise enough to run some of the most mission-critical systems in the world, including stock exchanges and nuclear submarines. And it’s scalable enough to run Google’s infrastructure. The Linux community also reacts at amazing speed. Everything from security issues, when they arise, to breakthrough enhancements are typically addressed within hours or minutes.
For example, the US Navy, one of our customers, came to us with a major problem. Events in recent conflicts had proven that smaller boats armed with missiles posed a serious threat to the larger ships like aircraft carriers. The navy wanted to develop antimissile systems that could react in real time to those kinds of threats. The problem is that most operating systems don’t work in real time. Think about the last time you were typing a document or working on a spreadsheet and your computer paused or experienced some kind of delay. That was because your operating system was working on other things in the background. While that’s fine for most of us, the navy clearly needed something better.
But rather than trying to come up with its own solution, the navy came to Red Hat and the Linux community for help in building a “real-time kernel” that wouldn’t experience those kinds of delays. We then helped catalyze the various Linux communities to tackle the project as something that everyone could benefit from. What happened over the next eighteen months was that the community worked together to create what the navy needed. But that same technology now used to help shoot down missiles is also used by stock exchanges to help ensure there is never any lag during the trading day. That’s because our community includes members from organizations like JPMorgan Chase and Goldman Sachs, which saw the value in this effort, and led to a way to solve their problems. In other words, an open source community like Linux works because its members both directly and indirectly benefit from the work that they contribute.
People invest their time in making Linux a better operating system because, by virtue of making computing safer, faster, and more open, they feel as if they make the world a better place. It’s that compelling purpose that brings so many people together to work on Linux and other open source projects. Not just programmers, but also writers, testers, marketers, designers, and administrators contribute their combined efforts to create something no one traditional organization could do on its own. One study estimated, for instance, that if you started from scratch, it would take about eight thousand years of development time to recreate Linux, at a reported cost of more than $10 billion.1 That’s powerful stuff.
Harnessing the Power of Community
The Linux story helps explain why we at Red Hat have come to appreciate that the open source development model is incredibly powerful at delivering innovation at a faster and less costly rate than the traditional software development model. And as Linux has evolved, so, too, has Red Hat.
In 1998, Red Hat had seventy-five associates and was earning $10 million in yearly sales, mostly from shipping Linux on CD-ROMs and selling T-shirts. People paid Red Hat for a version of Linux so they could get access to documentation and support. While Red Hat experienced a highly successful initial public offering in August 1999 using that business model, the dot-com crash almost brought the company down with it. The company’s stock, which traded as high as $286 a share in December 1999 (leading to an eye-popping valuation of about $19.7 billion), had cratered to about $3.50 by 2001 as the company continued to hemorrhage cash and struggled to grow revenue. That meant that if Red Hat was going to survive, it needed to evolve its business model to become profitable—fast. That was especially true because Red Hat competed against some of the mightiest tech companies around, a few of whom were extremely vocal in their derision of a business model they called “communist.” As Bill Gates, former CEO of Microsoft, told the press back in 2001, “We think of Linux as a competitor in the student and hobbyist market but I really don’t think in the commercial market we’ll see it in any significant way.”2
Unlike other business stories, which involve heroic CEOs riding to the rescue, Red Hat’s turnaround came from within. A collection of engineers, businesspeople, and lawyers identified as “the smartest people in the company,” according to Mike Evans, Red Hat’s vice president of business development, helped the fledgling company pivot its early business model. After debating several options, including some that would make Red Hat’s software proprietary, the team came up with the idea to completely shift focus from the desktop to the enterprise server—a key turning point in Red Hat history, though not a popular decision at the time. In fact, Paul Cormier (now Red Hat’s president of products and technologies), a key architect of the now powerful “enterprise open source” business model, received a “does not meet expectations” for his performance review in the year that he developed the business model. (He’s obviously more than met expectations in the subsequent decade.)
But when Red Hat shifted its focus to the enterprise IT market, it dawned on folks like Cormier that as powerful as open source is at solving problems, it can also create quite a few issues for companies as well. For example, open source development breaks down problems into smaller chunks that groups work on, which results in the frequent and early release of software that can be problematic for companies needing stability perhaps more than rapid change. And unlike traditional software with long release cycles, developers don’t fix bugs in older versions of code. They simply fix those problems in their next release, which again creates issues that affect applications for many years. The fast-moving, modular approach to writing software has many advantages over the traditional software development model, but it does not make it easy to use for running large, mission-critical applications liking billing systems or trading platforms that require absolute stability and as little change as possible.
If you’re running, say, a stock exchange, you need to know that any updates to your software won’t crash the entire system. Or, if you run several kinds of hardware in your business, you need to know for certain that the latest software update will run equally well on all of your machines or it simply won’t be valuable to your business. Red Hat steps in to stabilize the software while providing support and documentation. When we “freeze the spec,” as it’s called, Red Hat is actually making a significant investment in that we dedicate several hundred engineers to supporting each version of Linux for the next ten years. As changes occur upstream in the Linux community, our team will be working to create patches to the older versions of the code that our customers are running in a way that keeps their business running smoothly. Put another way, we generate our revenues by making open source software consumable, and safe, for enterprises to adopt.
By investing in tech labs of our own, not to mention the kind of top talent needed to run that kind of support backbone, Red Hat is able to provide value to our customers through services they pay for on a subscription basis. One of the real values that Red Hat delivers through its subscription model is that we don’t have any built-in incentives to deliver software to our customers that they don’t want or need. When you buy a Red Hat product, you are buying the support and value-added infrastructure for whatever product you buy without any hidden incentives to add new code and features—known as software bloat—that you didn’t ask for.
In short, by providing these services to our customers, and actively contributing and giving back to the open source communities that make everything possible, we’re able to make billions of dollars selling free software. We’re proud of that because we feel that our customers are receiving an incredible value compared to proprietary alternatives. Since the code itself is free, our customers pay only for the added value we provide. That means that customers not only reap the benefits of the innovative capabilities of the code created by open source communities, but also get the stability and security they need, all for about one-tenth the price of what our competitors offer. That’s why it’s easy to see that the future of software is all about the open source way.