Software Patents, Incumbents, and Entry

Abstract

Software patents have been controversial since the days when "software" referred to the crude programs that came free with an IBM mainframe. Different perspectives have been presented in judicial, legislative, and administrative fora over the years, and the press has paid as much attention to this issue as it has to any other intellectual property topic during this time. Meanwhile, a software industry developed and has grown to a remarkable size, whether measured by revenues or profitability, number of firms or employees, or research expenditures. The scope of software innovation has become even broader, as an increasing number of devices incorporate information technology, requiring modem manufacturing firms outside the software industry to employ developers and programmers to ensure that increasingly diverse functions are performed more efficiently.\nAlthough inventors have consistently asserted their need for patents in order to compete with industry incumbents, patent protection has not been easily or consistently available for much of this period. Rather, the legal system has responded gradually to the burgeoning software industry by broadening the scope and strength of protection for software-related inventions in fits and starts. The explosive growth of the industry is largely attributable to demand generated by the efficiency of software solutions; the expansion of the venture capital industry over the same period largely explains the lack of industry concentration. The "garage" mentality can be explained by the fact that even some of the largest industry incumbents began with one or two (largely unfunded) inventors. Also, there is every reason to believe that increased patent protection has contributed to the ability of independent inventors and smaller firms to compete.\nMoreover, the ability to obtain patents on software always has been important to some of the industry incumbents, while others have exhibited little need for patents and displayed, in some cases, strenuous opposition to the patentability of software. The incumbents are a diverse group. Some produce only software; others have substantial hardware product lines. Some sell to other technology firms and others sell applications to end users in a broad range of markets. And some sell prepackaged software products, while others focus on services-custom programming, installation, or maintenance. Regardless of the sector in which they participate, the incumbents spend massive amounts on research and development (R&D) – about 14% of their annual revenues, more than $60,000 per employee. However, there are important patterns in patenting practices that raw data on R&D investments cannot explain.\nThis Article examines the relation between patents and the different business models used by firms in the software industry. The analysis has four parts. Part II provides a brief retrospective on software patents, emphasizing the shifting role of patents as the industry grew into its modem form. Part III uses quantitative data about patent portfolios to discuss the role that patents play for incumbent firms in the modem era. We highlight the fact that business models explain much of the pattern of patenting practices. Part IV describes the use of patents in the three channels through which technology flows into incumbent firms – venture-backed firms, opensource developers, and independent inventors – all of which contribute to the development of technologies that might supplant or improve the products and services currently delivered by incumbent firms. Finally, Part V concludes with a brief discussion of present-day industry perspectives on software patenting. As incumbents are now leading the charge on patent reform on all fronts, we can expect that some change will occur. An understanding of the varying uses to which software firms put patents in their businesses provides a useful perspective on the types of reforms they advocate.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 53 citations exist upstream.