Originally Published: Saturday, 28 August 1999 Author: John Zedlewski
Published to: featured_articles/Featured Articles Page: 1/1 - [Printable]

An Open Business Plan for Red Hat, Inc.

This article originated in osOpinion and is provided under the OpenContent License.

Dear Mr. Young and the Red Hat, Inc. Board of Directors,

Your recent, wildly successful, IPO has shown, to the joy of Linux fans everywhere, that Wall Street does have faith in your business model. But now it's time to worry about the specifics of what you'll do with your fabulous market cap and how you'll really move towards long-term profitability to satisfy the suits. As you've built the foundation of your success on open software, I'd like to offer an open business plan. Of course, comments, additions, and changes to this document are welcome from everyone.
...

   Page 1 of 1  

This article originated in osOpinion and is provided under the OpenContent License.

Dear Mr. Young and the Red Hat, Inc. Board of Directors,

Your recent, wildly successful, IPO has shown, to the joy of Linux fans everywhere, that Wall Street does have faith in your business model. But now it's time to worry about the specifics of what you'll do with your fabulous market cap and how you'll really move towards long-term profitability to satisfy the suits. As you've built the foundation of your success on open software, I'd like to offer an open business plan. Of course, comments, additions, and changes to this document are welcome from everyone.

Acquisitions: There are occasional rumors that Red Hat would consider buying one of the other Linux distributors with its newfound wealth. Bad idea. What do you gain? A little bit of resentment and a few million in boxed Linux software sales. But you don't want boxed Linux software sales, that's not your long-term model.

Instead, and this is key: buy Santa Cruz Operation. No, I'm not kidding. Their market cap stands at about $300 million, even after gains related to their Linux announcements. What does this bring you?

  1. Revenue. IPO money is great, but it's wealth, not income. SCO will give you the funds to mask your losses ( ANY loss seems huge when you only have $10 mil in revenue ) and fund new ventures.
  2. Technology. Continue to sell/support the current SCO line, but, for the future, merge worthwhile SCO technology into Linux. Use your control of SCO technology to create an instant migration path from SCO to Red Hat Linux.
  3. Market share. With this acquisition, you can rightly state that, in terms of volume, Red Hat, Inc. is the world's largest distributor of both Unix and Linux systems. That carries a lot of leverage in the long run.
  4. Human capital. SCO is already gearing up for Linux support, and retraining their current staff to also support Linux would not be a challenge. In the current labor market, it will be painful to gear up a large, national consulting force. With SCO, you can just buy it. In the longer run, you'll be able to shift the company's engineers over to Linux and Linux-application development.
  5. Relationships. This is your immediate entry into long-term relationships with ISVs, corporations, and resellers worldwide. On top of that, you obtain a close relationship with IBM in the form of:
  6. Monterrey. The wildcard OS that might be a flop or might be the greatest Unix ever. In either case, you're covered. Unlike SCO by itself, which would die a flaming death on the failure of Monterrey, you always retain the option to cut your losses and bail (leaving IBM with another OS/2). But if it succeeds and doubts persist about Linux's scalability beyond 4-8 processor systems, you can position Monterrey as the high-end partner to Red Hat Linux. Additionally, you gain a base of developers skilled in working with Merced and the opportunity to smooth out APIs between Linux and Monterrey so that porting between the two will be as easy as possible.
  7. Brand name. In some circles, the SCO brand name is strong, in others, the Red Hat brand name is huge. By controlling both brands, you can cater to both markets, even providing a SCO (Stodgy Corporate Option?) Linux-solution at some point in the future. It also allows you to market some non open-source products, such as Tarentella, under a label other than Red Hat, allowing you to maintain the close association between that brand name and the free software principle.

Developer relations- Good developer relations lead to good application availability, which leads to commercial success. Remember: People don't run an OS, they run applications, and Linux needs to not just catch up to commercial Unices, but surpass them in terms of application availability. But you knew that already. The question is, where to go from here? I think it's a multi-pronged approach: online developer information (redhat.com), printed media (the Red Hat press), tools, and technologies.

Online- You'll never make a real profit on the portal. Slashdot, freshmeat, and Linux.com have the community support that are necessary for success, and they don't have the stigma of being potentially distribution-biased. But that's ok. You can break even on the portal and accomplish something more important: create a more in-depth linux.com that contributes to the Red Hat brand name and strengthens developer relations. Sure, dozens of sites exist that are supposedly Linux development-related. But none really provides all the information an aspiring Linux developer would need. Red Hat can create the DEFINITIVE repository of library documentation, open books (like the upcoming open book on GTK+-programming), whitepapers, language tutorials, and so on. Remember one thing: developer relations are the key to OS success. If you have to take a small loss in order to get 1000 pages of great info to Linux developers, do it. You'll make the pie bigger in the long run. The portal also gives you a chance to promote sales of Linux applications. Customers need to know that they can always go to a single, central location to buy any sort of Linux software, from Mathematica to Oracle 8i.

Print- Definitely gear up the Red Hat Press, but partner with an existing, experienced publisher. You can differentiate yourself from the herds of virtually-identical Linux publishers with your brand name, but, again, revenue isn't really the key here. Wide availability of educational materials will vastly expand the acceptance of Linux. Right now, the market still largely consists of a thousand general Linux books. You can publish in-depth guides to the key Red Hat technologies, and each one will effectively be an ad for your brand, because, even if you cover other distros, it will still bear the Red Hat name.

Tools- Oh boy, here it comes. My one reference to you-know-who: Microsoft makes development tools and technologies the backbone of their platform, and that strategy has been crucial to the acceptance of Windows and MS Office. But do you need to run out and buy Cygnus? Not really. In fact, open source development tools are the ultimate platform booster, because they lower the cost of entry to the Linux software market to $0. Work very closely with Borland/Inprise and Cygnus; they have a lot to offer, but Borland needs some Linux expertise that you can provide. Create a separate development tools division which pushes open source projects such as KDevelop (the most complete free IDE around), gIDE/Glade, and, perhaps more importantly, development tools in the spectra that haven't been widely covered: database/distributed/enterprise technologies and RAD/component/desktop tools.

You can market enterprise-level add-ons to these products as they progress. And, by the way, web development tools will be key as well, both at the programming level, like Cold Fusion or Allaire's Spectra, and at least one visual design/management tool, like NetObjects Fusion. Until that comes around, web houses have no choice but to keep a spare Mac for WYSIWIG work.

Technologies- Components, components, components are where it's at. EJB support in free application/web servers, if combined with a Java2 JVM from IBM that lives up to the speed of their 1.1.x versions, will be critical to speeding migration towards Linux. Components on the desktop and distributed levels will also be key, however. Bonobo and KOM both show promise, but until they can achieve some level of interoperability, neither will take off at the commericial or enterprise levels. Similarly, both of the free ORBs need development help. ORBit isn't really designed to be enterprise-ready (that was, of course, a conscious choice, not a flaw!), which MICO needs some speed/robustness help. Turning MICO (or another alternative) into a credible, low-end ORB for enterprise applications would be a tremendously important move in support of CORBA as opposed to its proprietary competitors.

Red Hat Labs- Obviously, your new capital will allow you to expand both the size and the scope of RHADL. But in what direction? For starters, you need to clarify things. You need to separate Red Hat "fellows" from Red Hat "programmers." The former are the type who have made substantial commitments to free software in the past and are largely working on their own pet projects, but with nudges and support (financial and otherwise) from Red Hat. The latter are regular employees who receive much more direction from above, even in terms of the free software they work on. Your past focus of RHADL is pretty much on-target for the fellows, but should also include the technologies I've mentioned earlier: GNOME, with an emphasis on usability (integrating a window manager!), Bonobo / KOM component models, CORBA, enterprise-strength kernel technologies, and development tools.

Red Hat Consulting Services- Here's where a lot of the non-fellowship programmers will also end up. You can offer those employees a great deal: when you're not on a project, we pay you to write free software that you love, but when you are on a project, you use that same expertise to help ensure the future strength of Linux (and your stock options) by helping major corporations and ISVs support the platform. In the mid-term range (2-3 years), there's an incredible market opportunity for name-branded Linux programming expertise. A development company wants to port compilers and libraries? No problem, they can hire Red Hat engineers under NDA who have actually contributed to gcc and glibc. By providing a central source for contract programming talent, you both speed the transition to the platform and create a real revenue source. The relationships you form with these projects will also be your entryway into long-term support contracts with big-name firms.

Red Hat Software- Last, but not least, what will you be selling two years from now? The answer, I think is clear: integrated solutions (buzzword buzzword). Your new Red Hat e-commerce server is a nice start in this direction, but it still falls into the Linux trap: assuming the customer can do it for him/herself. Sorry, guys, the small businesses you're targetting with this product just don't have a lot of PHP expertise. In the future, Red Hat will offer both a low-end e-commerce server, which includes database-driven templates and lots of hand-holding, and a mid-range server that truly integrates with a free application server, such as Midgard or Zope, and provides real examples, documentation, and templates for a business to get a web site up overnight. This is an important step towards getting e-businesses to start using Red Hat support for real, because it takes your products to a new level: the development and applications levels. Of course, all this also takes support from the Red Hat press ("Building E-Businesses with Red Hat ZopeServer") and redhat.com.

John Zedlewski is a student of economics and computer science at Princeton. He can be reached for response to this column at: zedlwski@princeton.edu.





   Page 1 of 1