June 1, 1995
Stanford University's Computer Industry Project (SCIP) is an interdisciplinary study of the worldwide computer industry. Funded in its initial years by the Alfred P. Sloan Foundation, this project's goals are to characterize the nature of the emerging industry, analyze trends and identify and monitor important industrial, educational and public policy issues. We are reporting here on the efforts of one area of the SCIP's research, the Software Industry Study.
At first, one might think of the software industry in terms of packaged products from companies like Microsoft, Oracle, IBM and Novell -- the software products segment, as we call it. However all computing involves software. The software services segment (consulting, custom systems, systems integration, etc.) is a larger industrial activity than is software products. And much larger still is the effort put into software that is developed "in house," either as part of attempts to use computing to improve business operations or software to be embedded in the companies' products themselves to add flexibility and functionality. Software has an increasingly critical impact on almost every industrial and government organization.
In the fall of 1993, using the same methods we used in Japan, we turned our attention to the prepackaged products segment of the US software industry. It is in this software products business that we find the most vivid contrast between the US and Japan's (or any other nation's) software industry. In the context of three graduate research seminars, students and staff again conducted extensive library research and interviewed over 40 people associated with the US software industry: senior executives of software publishing companies, venture capitalists, software association directors, journalists, technology futurists, corporate IT managers, and government researchers.
Japanese Software Industry: The reasons for the troubled state of the Japanese software products industry go much deeper than the language barrier or the relative size and fragmentation of the Japanese PC market. Software, perhaps because it is a non-material product, has a lower status in Japan than other highly engineered goods. This perception, in turn, makes bundling (packaging "free" software with computer systems purchases) a common phenomenon. The lower status of software as a product also puts less of a stigma on software piracy, which, at least until recently, has been rampant in Japan as it is in many other nations. Both bundling and rampant piracy make it very difficult for an independent Japanese software publisher to find a sizable domestic market. The lower status of software also affects the career choices of bright young engineers and the investment choices of Japanese capitalists. Software in Japan is not as prestigious as it is in the US.
The future of the Japanese packaged software industry is also problematic. For reasons involving government educational and research-funding policies and the lack of prestige of software careers, the university system is not producing enough top-flight software professionals with up-to-date expertise. They will have a difficult time competing with the complex, cutting-edge software that is being developed elsewhere in the world. Furthermore, the relatively conservative stock market system, absence of a significant venture capital community, negative cultural attitudes toward individualism, entrepreneurism, and career risk, and the strictly seniority-based hierarchy of Japanese companies all contribute to an almost total absence of new software startups or spin-offs entering the market with innovative ideas.
Piracy. Despite public statements to the contrary by industry associations, software piracy in the US is not considered to be a top-level business problem among executives interviewed at US firms. Attitudes about software and intellectual property generally, along with the consciousness-raising and policing activities of software industry associations, have minimized the extent of software counterfeiting in the US. In many other countries, however, piracy is rampant because of different cultural perceptions of software as intellectual property, different histories with respect to bundling of software free with computer systems, and lack of enforcement of copyright laws. As mentioned above, a secondary impact of rampant piracy is debilitation of the domestic software industry, even in countries where there is a well-developed infrastructure that could support such an industry.
Off-shore Software Development: While offshore programming is still a very small part of the global software market (approx. $500M/year), software services exports from India, Taiwan and other exporters of contract programmers and custom-built software are growing at two to three times the annual 15% growth rate of the US software services industry, driven by the significantly lower wage levels for high-quality foreign software developers. In India, especially, there is an extensive supply of high-level programming talent who, in addition, are generally fluent in English. The sophisticated software export firms in India are continuing to invest in communications infrastructure, modern computer hardware and up-to-date software skills training for their programmers. Many projects that would be done by average programmers in the US can be done at lower cost and by better programmers abroad. Quality is as much a factor in decisions for offshore outsourcing as is price.
Software Education. In both Japan and the US, all interviewees were asked whether their respective education systems produced well-trained software professionals. Japanese interviewees in government, academic, and corporate settings were uniformly dissatisfied with the technical skills of newly graduated programmers and software engineers. In contrast, interviewees at software publishing firms in the US were satisfied with the graduates of US computer science programs, at both the undergraduate and graduate levels. (US respondents did express some reservations regarding training in project management and team development, however.) It is important to note that historically, the ability of the US education system to keep both faculty and graduates at the cutting edge has been due in large part to defense-funded university research, and may be threatened by post-cold-war policy changes.
Video game industry: An investigation has been made of the competitive dynamics that are widening the scope of the video game industry to encompass more general forms of interactive, multimedia entertainment. Japanese firms (Nintendo and Sega) have dominated this segment and have succeeded in establishing the architectural standards. The video game industry is already playing a significant role in accelerating the consumerization of computing worldwide. It also provides our clearest example of the growing convergence of the computing, consumer electronics, telecommunication, and entertainment/publishing industries. Through an increasing number of strategic alliances in interactive multimedia, game makers will continue to expand their markets and their influence on the direction of consumer software products in general.
Consolidation, Domination and the Sales Channel. In the PC software industry, recent years have seen a massive consolidation of the market resulting in its substantial domination by Microsoft. Furthermore, the "action" in this segment has moved from corporate desktop productivity products, a saturated market in the US, to the consumer and the 35 million PCs now in peoples' homes. The direct and boutique sales channel is dwarfed by the mass-retail outlets, mastered by companies like Intuit who view their products not as software but as a consumer products, like toothpaste. The effect on the industry of these changes, as well as new technology for software distribution and "per-use" charging, will certainly be profound, but is very hard to predict at this time.
Intermittent programmers. In a recent study, the class of programmers who work neither for software products and services firms nor in corporate IT departments were labeled "intermittent programmers." These are the people who write the code that goes into our automobiles, microwave ovens and telephones. Their numbers are doubling every year! They generally do not have any software training, and, in many cases, they don't feel they need it.
However, the software these intermittent programmers are developing is becoming more and more complex and finding its way into more and more parts of our lives. For example, the development of a modern computer chip is done using a hardware design language by electrical engineers who, for the most part, do not consider this coding to be software development and do not consider software development a respectable activity. The limitations of these engineers' ability to debug their code, e.g., in the case of the Pentium chip, must be viewed as an example of a software project failure, a phenomenon we will be hear much more of in the coming years (see below). Limited though modern software training may be, including courses on software structures and techniques may be called for in a wider range of professional curricula.
Software Invention and Intellectual Property Protection. Although there is universal agreement among publishers that the US patent system and case law regarding the protection of software inventions is terribly flawed and that the business consequences are serious, there is no agreement, among publishers or lawyers, about what to do about the situation. Many industry executives express concern over hidden patents and the cost to a small company of pursuing a patent infringement case or defending against one.
The Financial Aspects of the Software Industry. There are a few unusual characteristics of software companies. One important fact is that the bulk of product development expenditures are characterized as R&D and there is very little production tooling or manufacturing expense. As a result, current anti-dumping laws are inadequate because they are based on "cost of manufacture." (Also cover: Influx of money from Hollywood, etc. Furthermore, compared to most manufacturing industries, the initial capitalization of software companies can be very small. The extensive impact of a venture capital firm on the start-up.)
Industry Data. The data from industry analysts, although adequate to determine short-term market trends, is incomplete and inconsistent over a multi-year run. Thus, it is inadequate to truly analyze long-term trends in the industry. Government economic and labor data, whose collection cycle-time is longer, also have inadequate and out-of-date categorizations of software industry activities. We are working with both industry and government agencies to help extend and normalize their data gathering processes.
Software and computers cannot be separated. Software is defined as the instructions that a computer follows to perform a specified task. Software development starts with the specification of the task, whether it is the functionality of a microwave oven or the details of an airline's frequent flyer program, some person has to specify what he or she wants the computer to do, exactly. The process of translating that specification into instructions for the computer, testing to make sure the specification and its translation are correct, and documenting and maintaining this "program" as people use it and request modifications and new features is the work of the software professional.
We've divided the world of software into four major segments:
However, the true impact of software on the computer industry and on almost every other major industry goes well beyond product and services sales. We see the hundreds of millions of dollars that companies have been spending for thirty years on information technology, but industry estimates are that, over the years, 70-80% of the cost of business systems is in the development and maintenance of the software these companies write themselves. As corporations introduce automation into more "front-office," and therefore more rapidly changing, areas of their business (e.g., sales and service vs. accounting and manufacturing), even more intense development and maintenance efforts are required. And then there's the software in the products and services themselves: viewing the commercial impact of computing primarily in terms of corporate information systems is no longer valid.
And as the cost of computer chips drops, computers and the software that runs them are becoming important parts of many products. Companies in industries like telecommunications (cellular phones), airlines (frequent flyer programs), aircraft (on-board avionics), consumer electronics (programmable VCRs and microwave ovens), and automobiles (air bags and anti-lock brakes) now have products whose competitive features depend on the software they've embedded in them. And some industries, like the $10B cartridge game industry, are just selling software in disguise.
Some companies have indicated that 70% of their product development costs are now in software development. There are no estimates of the total amount of this "embedded" software in the products of so many industries, but this is a relatively new phenomenon and there is every expectation that it will continue to grow rapidly, both in cost and in the complexity of the software itself:
Our initial explorations of the ailing Japanese software industry and the world-dominant "products" segment of the US software industry have already offered some valuable insights. We are continuing these investigations to confirm our initial impressions and establish more quantitative results. It is important that we continue to take a serious look at the industrial impact of software. We expect to expand our study over the years to include all software activity: products, services, in-house and embedded. Software may be to the 21st century what chemistry was to the 19th and electricity to the 20th-- a driving technology that determines the fate of industries and nations.