Open Source--The Unauthorized White Papers Chapter 12, The Proprietary Software Business: New Opportunities . . . and Pressures

Open Source: The Unauthorized White Papers

Chapter 12

The Proprietary Software Business: New Opportunities . . . and Pressures

No one should be surprised that the shortest chapter in this book deals with proprietary software. The Free Software movement abominates it and is working and waiting for its disappearance from the earth. The Open Source movement, on the other hand, believes that proprietary software has its place in the world, and that there ought to be licensing arrangements that permit Open Source and proprietary software to work together.

In its pragmatic fashion, the Open Source movement believes that while some developers may develop and give away their code for purely unselfish reasons (as the Free Software movement teaches), many developers have done their work because they needed to do it, and they have made it freely available because they hope to have the bug-fixes and maintenance contributed by others. Proprietary software is needed in an Open Source world when there is a need for software that no one has the resources or the motivation to tackle. This chapter will look at the commercial distribution model we are used to today, and how that model might change with the arrival of more Open Source software.

Linux Distributions

The Linux distributions (distros) have shown that it is possible to base a business on the distribution of free software, including the free distribution of in-house improvements to a distro. This is possible because they are in a perfectly horizontal market ó every user needs an operating system.

Looked at as a shrinkwrap product, the distros have the same economic and marketing model as proprietary software. Although development costs are theoretically nil because the raw Linux product is free on the Internet, the distro houses spend a good deal of effort improving their methods of package loading and installation, while testing and tuning the entire ensemble. Even if we were to calculate the research and development costs for an Open Source software product to be less than those of proprietary products (even though UNIX programmers are more costly), we must remember that in any software company only a small fraction of the retail price of a product goes to R&D ó the majority of each dollar goes to marketing and distribution costs. There are ways, however, to make money from the intellectual property (IP) that each distro house controls.

The chief IP of each distro house is its brand, which may be licensed; in addition, the sales of advertising on the distro Web site or the inclusion of advertising or software in the distro boxes themselves are revenue opportunities. These opportunities grow in value in proportion to the power of the brand. While consulting and training can be handled effectively by third parties (although the distro house may claim superior experience with its own distro), certification is intimately tied to branding. Third parties or even community projects may certify that a person is knowledgeable about Linux, or even about a certain distribution of Linux, but only Red Hat can create a (trademarked) Certified Red Hat Engineer. The process could be handled in house or franchised.

From Red Hatís brand-centric point of view, anything that causes the market to identify Linux with Red Hat is likely to expand further Red Hatís major share of the Linux distro market. As market leader, Red Hat has a further advantage: anything it does to expand the entire Linux market helps Red Hat more than it does competitors. Red Hatís lead is so great that the only other distro to promote a certification program, Caldera Systems, emphasizes that the people trained and certified in its program have had a "distribution-neutral" training and examination. Caldera branding is achieved by emphasizing that the training and certification program is nevertheless sponsored by Caldera Systems.

Red Hat began its training and certification programs as part of its efforts to qualify Linux for the business user: the standard complaint was that there were no such programs for Linux. Now that the Linux Professional Institute (LPI) has begun to test and certify, we have an overlap of a proprietary program and a community-based program. We are going to see a lot of this in the Open Source world, beginning with the horizontal applications described later in this chapter.

Branding and user loyalty to the brand are important to the distros because these assets enable the distro vendors to sell through the large distributors, such as Ingram Micro (http://www.ingrammicro.com) and Tech Data Corporation (http://www.techdata.com). These in turn sell down to the retail chains, VARs, and integrators. Only products with substantial established sales and large marketing budgets can provide the customer pull to persuade the distributors to carry the product. Branding is likewise important in signing on resellers to carry the distro and promote it to customers. Red Hat, for instance, has the volumes to be picked up by the distributors, and Corel, another strong brand, has preexisting channels into which it can feed its new Linux products. Caldera, on the other hand, has been aggressively signing up resellers and placing a heavy emphasis on training and certification to address the lack of UNIX experience among them. Caldera has shown the patience, attention, and hand-holding necessary to develop a reseller channel; these skills come much harder to Red Hat, whose resources are strained by the tornado of skyrocketing sales and an IPO.

In the push to extend their markets, Red Hat (and TurboLinux) have become OEM vendors. By purchasing Cygnus, Red Hat is aiming itself at the embedded market, as well. Caldera was already there with the formation of a sister company, Lineo (http://www.lineo.com), to deal exclusively in embedded Linux.

Is it possible to make a proprietary distro? Only in the sense that some proprietary applications or utilities might be put in the box and not be made available to the other distros. Because the Linux kernel code and the packages around it that make up the Linux platform are under the GNU GPL and other Open Source licenses, it is possible for any other distro to take advantage of anotherís improvements. But it might be possible to make a distro that was specially tweaked to serve a market niche and that introduced changes that were unattractive to the other distros.

An example might be a distro for newbies, a Linux "for the rest of us," as Apple used to say. The majority of potential users will never want to see the source code that comes in the box (nor would their bosses want them experimenting with it), and will cheerfully use the binaries they receive. If an easy-to-use totally enclosed distro won the loyalty of its users, they would not notice or care that upgrades to its hundreds of packages were always handled from the distroís Web site, or that, in fact, they could load only applications that came from the distroís vendor. They would be, in essence, Macintosh users, and regard it as a benefit, not a limitation. The other distros would have no reason to take advantage of code that was focused on a competitorís Web site. Third-party software vendors would have to come to an accommodation with the distro. Open Source purists might scoff at the results, but they could not deny that the source code was available.

Office Suites and Horizontal Applications

As the first popular applications sold for Linux systems, office suite packages give us a look at the direction any set of horizontal (widely-used, or mass) products will take in the future.

In the first place, prices vary widely, from free (StarOffice in the desktop version) to relatively expensive (Applixware and eventually WordPerfect). In any case, the money paid for any of these is less than for the leading brand for Windows. Needless to say, no paid-for office suite can be freely installed on multiple machines. Although the UNIX market has always been envied by the Windows market for its relatively higher pricing structure, it is clear that mass products for Linux will tend to be cheaper than those for Windows. In addition, as completely free Open Source products like AbiOffice mature, they will add to the downward pricing pressure already exerted by three factors: 1) the original base of Linux among impecunious geeks, 2) the need for products in new markets to take losses to establish market share, and 3) buyers spoiled by a cheaper operating system capable of running on less expensive machines. In the Windows market the increasing price of software relative to hardware causes frequent complaints, and in the case of the operating system it invited government attention.

If you go into the large office-supply chains nowadays, or visit the mass-merchandisers, you will notice racks of software suitable for office work, priced very low relative to products from Microsoft or Symantec. In a couple of years (or even less) there is no reason we will not be seeing office suites in this price category ó for Open Source systems. This boon for the consumer will not necessarily be good for the independent software vendors (ISVs), although it will result from the answer to their frequent prayers: the arrival of the "level playing field."

For years ISVs have looked on as Microsoft relentlessly winnowed their ranks. Some of this attrition has been a normal function of the proprietary software market. In the days of Lotus 1-2-3 and dBase, each was surrounded by a school of small fish who offered little software products that added features to the hugely successful spreadsheet and database. Those add-ons that the market bought in large numbers because they worked and everyone had a use for them (the ability to print spreadsheets sideways was one such winner) found themselves doomed either to acquisition, or worse, to finding the same functions added to the main product they were supporting.

Microsoft intensified this process by several methods. One was to pull into the operating system itself functions formerly offered by third parties, sharply reducing the market for these ISVs. The Windows applets, for instance, forestall many third-party product sales, and the browser is the most famous example of all. Next, Microsoft aggressively promoted its Office applications. Users could still theoretically exercise their choice of best-of-breed among spreadsheets, word processors, desktop databases, presentation packages, and so forth, but Microsoft offered all of these in a suite that cost about what any two or three applications would cost. Furthermore, the applications were tied increasingly together: Word documents could be updated from changes made to Excel spreadsheets. It was a value proposition that users found irresistible, and as more users switched to Microsoft Office, other users found it harder to justify other choices. This is the network effect that establishes a de facto standard among users: if other people send you Word documents and Excel spreadsheets, the best bet for being able to read them is to own Microsoft Office.

Increasingly, ISVs who introduced a horizontal product found themselves competing directly with an aggressive giant who could either acquire them or acquire their market by the methods mentioned previously. A number of smaller companies ended up being acquired by Symantec, one of the few horizontal ISVs remaining. None of these competitive methods, however, irritated the ISVs as much as did the belief that by knowing the innards of the operating system as no one else could, Microsoft had an enormous advantage in building applications for it. ISVs longed for a level playing field.

If Linux fulfills that hope, it will not bring a paradise for the ISVs in Open Source horizontal markets. The biggest barrier to market entry will fall because no one will have privileged knowledge of the innards of the operating system, but the lowered barrier will draw forth many new ISVs, and the increased competition will result in lower prices. Linux will have the same price-depressing effect as Philippe Kahnís offering at low prices of the excellent Borland tools and applications back when Windows was young. Factors beyond the number of competitors will also push prices down. Established products in the UNIX market (like Applixware) have already absorbed the bulk of their development costs. Even proprietary products starting from scratch will be able to lower development costs by taking advantage of Open Source work already done (such as the GNOME graphical interface). And any horizontal proprietary products will likely face both proprietary competitors freely given away to support a marketing purpose (as the desktop StarOffice is given away to promote the Web-based product) and GPLíd competitors like AbiOffice, free for the downloading. Horizontal products will be doomed to low margins.

Niche Products

The only refuge from this maelstrom will be in specialized products sold to niche markets. Here the barriers to entry are knowledge of industries and business practices in the chosen niche ("domain" knowledge) and the necessary investment in developing the software. The relative lack of developer power in any one customerís IT shop further shields the proprietary ISV from competition. All of these barriers promote higher margins in niche markets than can be had in the mass market.

Open Source software practices will nevertheless have some effect on niche markets as well. The ISV may be able to appropriate Open Source code to put in a product (as Apple built chunks of BSD into its newest operating system). Open Source advocates believe that eventually developers within an industry will learn to cooperate across companies, working together to develop Open Source products that will serve the industry generally, and leaving to individual companies the addition of private (that is, not distributed) features that will give competitive advantages. In the meantime there is an opportunity for ISVs to create "less-than-finished" products to be sold to VARs and integrators who will customize them for end-users: either APIs ó as is generally done nowadays ó or, as Open Source becomes more accepted portions of proprietary source code could be exposed for customizing. Although niche markets are smaller than mass markets, and the revenues are correspondingly smaller, the marketing costs are considerably less, a factor that also contributes to higher margins for the ISV.

The more things change, the more they stay the same. In short, the world after the arrival of the level playing field is going to look almost like the software market we know today. Presently, any ISV with a horizontal product knows that Microsoft is the chief competitor, and will fight fiercely if the new product is popular and profitable. Todayís ISV also knows that Microsoft is not interested in niche markets, and that no one in those markets has a privileged view of the operating system, making the niche a better market for ISVs today. In tomorrowís Open Source software market, the mass market will be disputed by a host of horizontal vendors, rather than by one large one, and will continue to offer low profits for that reason. Even if an ISV offered a killer app to this market, it would have to carry a low price to forestall the communityís loving it so much that they clone it. This is what AbiWord is doing (and on the Windows platform as well as for Linux). The niche markets will offer more chance for profitability, and there are plenty of expensive Linux applications available today that cater to scientific and engineering specialties. And just as today, the higher-margin opportunities for ISVs will be in training, support, and product updates, especially when these are offered as subscriptions.

The cloning threat is the most interesting innovation that Open Source brings to the proprietary software market. Sun Microsystemsís failure to make Java an open standard, its use of restrictive licensing, and its favoritism in tilting the software towards Sun-manufactured hardware caused a number of Open Source projects to emerge, with the result that the fastest Java Virtual Machine and a number of Java development tools are now Open Source products.

You will have a chance to see how much of this turns out in the next year or two. Part of the speed of the changes will depend on how much progress Linux makes among general users. In the meantime, the next chapter will show you additional developments to watch for.

Next chapter: Companies and Projects

Table of Contents

Copyright © 2000 by IDG Books Worldwide, Inc. Published under IDG Books Worldwide, Inc. Open Content License.