|
The question of how patents affect innovation in the software industry draws interesting debate. A recent paper from John Allison, Abe Dunn and Ronald Mann of UT Austin fills in a gap that research in the area has not addressed: how patents relate to business models within the software industry. The paper, Patents and Business Models for Software Firms, U of Texas Law, Law and Econ Research Paper No. 77 (March 24, 2006), produces important findings: ° Patenting is a “regularized and important part of a well-organized operation, rather than a random or happenstance occurrence.”
° Firms patent more, or less, depending on their reliance for revenue on selling hardware, software and services, respectively.
° R&D intensity positively correlates patenting activity more in the software industry than other industries. Especially noteworthy is the implication that patenting in the software industry results from profit motivated decisions companies make based on their business models. This is important, considering the deterring effects on discourse caused by the principals-based arguments of open source advocates (see Eben Moglen in Open Source politics are American as Apple Pie, and Lawrence Lessig in A War Against the Freedom to Innovate). While the open source movement argues that open source development is more innovative than software protected by intellectual property and patents, what they really advocate is their business model. However, whether business models innovate depends on their performance in the marketplace, not how much they reflect “freedom” or the “spirit of open source.”
The purpose of the paper is to “understand the relation between the business model of a particular firm and the patent practices…” 17. The authors acknowledge, “what little we know about patenting in the software industry suggests every reason to believe that the role of patents differs substantially within the industry itself.” 4. The context of the study is that while patents in the software industry are more controversial than in other industries, there remains little empirical research on the role of patents on a firm-by-firm basis. 3. The authors state that a study of the software industry must accommodate its heterogeneous characteristics, where the role of patents may differ substantially in sectors within the industry. 4. Thus, to answer the question of the relation of patents to business models in the software industry, the authors focus on how patents correlate with distinctive business models in the industry. 4.
The paper is important because it shows how patenting differs according to business models in the software industry. Patenting is found as a “regularized and important part of a well-organized operation, rather than a random or happenstance occurrence.” 38.
The authors assume that “the number of patents applied for in a year is a function of a firm’s R&D spending and other characteristics of the firm.” 18. Other factors are considered as the paper notes, “there are important influences on patenting practices beyond the raw amount of investment in R&D.” 14.
The study takes as its framework the different types of software firms outlined by Harvard Business School Professor Michael Cusumano in his work, The Business of Software (2004). Cusumano’s framework is combined with data from five years of Software Magazine’s Software 500. 5. The basic framework for the study outlines firms by: ° Product firms with higher operating margins, higher growth rates and less stable market shares. These firms often sell non-customized off the shelf products.
° Service firms with lower operating margins, lower growth rates and more readily establish stable market positions. These firms generate revenue by helping firms install, design and maintain software.
° Hybrid firms. While these firms resemble product firms, they also focus on customization and services. Cusumano’s outline of firm models is adopted because the “value of the products-services distinction in explaining other aspects of software business suggests that it might provide a useful lens for exploring the reasons for the apparent disparity of patenting practices in the industry.” 6. Further, “a number of practical aspects of the industry suggested to us the likelihood that patents would be more useful for products firms than for services firms.” 6.
The authors immediately draw implications from Cusumano’s framework: the products model is “relatively more effective for venture-backed startups than the services model.” It is much easier to duplicate “a product 10,000 times than the employees that provide services.” 6. “Products firms, because their technology is more difficult to protect than the technology of services firms, will produce more patents than services firms, all other things being equal.” 7. “Successful product firms “are more likely to produce the high returns venture capital investors seek”. 6. Further, the study finds many practical aspects in why patents would be a “more effective tool for protecting innovation in products than in services. To the extent a firm provide a unique level of skilled services, it may be feasible to maintain much of the differentiating knowledge in a tacit form, bound up with the skills of the individual employees.” 6. Products firms, by contrast, send products to market and risks mis-appropriation by competitors unless they pursue patent protection. 7.
After comparing the definition of software patents used by other researchers such as David Mowery and Stephen Hunt, the study defines a software patents as that “in which at least one claim element consists of data processing, regardless of whether the code carrying out that data processing is on a magnetic storage medium or embedded in a chip.” 11. The authors state that their approach will overcome the under inclusiveness of Mowery and the inconsistency of Hunt.
Sectors within the software industry found with the highest propensity to patent include: vertical industrial applications, application service providers, geographic information systems, operating systems, publishing and graphics, security, database and wireless and mobile. Sectors that patent less include financial applications, enterprise resource application, e-Business applications, enterprise application integration, application development, content and document management, business process management and data warehousing. Among the major findings, the study suggests: ° “An increase in the fraction of revenues from software services implies fewer patents are produced... a 1% increase in the percent of software sales coming from services e.g. percent of sales increasing from 50% to 51%, implies a 2.3% decrease in the number of patents produced. “ 22.
° “A firm that derives all its revenues from products is expected to produce 230% more patents than a firm entirely devoted to providing services. “ 22.
° “Returns to scale in number of employees is approximately constant in the software industry. In other words, the firms patent in proportion to their size, so that a doubling in the size of a firm is predicted to cause a doubling in the number of patents produced. “ 23.
° “The effects of size in the software industry are about the same as those in the semi-conductor industry, but that the effect of R&D intensity on software patenting is quite a bit greater than its effect on semi-conductor patenting. “ 23.
° “Patenting practices will vary depending on how much the firm is devoted to selling software products. 24.
° The devotion of a firm to a products or services model is important, even within a particular sector... there are important differences along the products/services continuum, even within particular sectors, and that differences even at that specific level relate to differences in patenting activity. “ 27.
posted by Noel Le @ 1:47 PM | Patents
Link to this Entry |
Printer-Friendly |
Email a Comment | Post a Comment(0)
|
