Monte J. Shaffer is a fourth-year Ph.D. student and job market candidate (2011) in the Department of Marketing at Washington State University. Monte is currently working on his marketing dissertation in Entrepreneurial Innovations. Prior to joining Washington State University, Monte received a Bachelor in Mathematics / MBA in Marketing from Brigham Young University (BYU) in Provo, UT.
Patent data is publicly available, serves as a instrument for doing patent-level and firm-level analysis for both private and public firms, and amidst the modern information age, may be the only way to secure intellectual property. Patent counts or forward-citation counts have been traditionally used to measure the innovation portfolio of a firm. Using network analysis, a variation of Google's PageRank algorithm is introduced to the patent citation network to define an objective measure for radical innovation -- ``Patent Rank". Two model types are considered: simple structure and technology ``class-match" using two temporal forms: cumulative network and five-year moving window. All utility patents from 1976--2009 will be analyzed; over 5.6 million patents and 40 million citations are evaluated to produce 332 million Patent Rank scores. Useful distributional properties are considered and these objective scores are compared to a recent subjective survey performed by PBS to assess the question: What are the most radical innovations of the modern era?
There has been a call for 'new' patent data (Kortum - see Tellis et al. 2009). I believe that I can contribute to the field of marketing strategy by improving the data available, and describing its potential uses. The new data source allows for large and rich information regarding patents that can be used in many types of strategic analyses. The most recent run of these data consisted of 73 IT firms in the S&P 500. Collecting data from January 1996 to June 2009 provides over 192,000 patents with information about forward/backward citations, classification matches, and more. The programming process to run this list took nearly 36 hours as it had to analyze over 3 million patents to create the informative dataset. This is my definition of new data, and the process is continuous and ongoing: (1) All Patent Data has been harvest (8 million patents); (2) Parsed Data is currently being stored in database format; (3) Firm boundary issues [IBM, Internation Business Machines, mergers, misspellings, etc.]; (4) with an intent to do new modeling research on the patent data: (a) Diffusion of Radical Innovations (patents); (b) Patent Rank (e.g., Page Rank applied to patent network of citations) - structural and weighted ranks (e.g., classification matching); (c) EIQ; (d) Race to the Patent Office; (e) Patent Pending