Technology Potency: Patent Citation Refinements for Merger and Acquisition or Joint Venture Analysis

The real value of a possible acquisition or joint venture partner is often in the potential of their next invention. As advisers, patent managers, and intellectual asset strategists, it is critical that the methodology and tools used for assessing patent quality accurately reflect a would-be ally's technological potency.

Recent studies have drawn positive correlations between the increasing or decreasing quality of a company's patent estate and its stock market value. Lanjouw and Schankerman, Economic Journal, 114 (April), pp. 441-465. Patent quality, in the broadest sense, is most often tied to the frequency a patent is cited by other patents. Hall, Jaffe, and Trajtenberg, Royal Journal of Economics, Jan. 2004. One study correlating market value with forward citations found that patents valued at $1 million received an average of about eight citations, while patents valued at $100 million received an average of about 14. Harhoff, Narin, Scherer, and Vopel, Review of Economics & Statistics, Vol. 81, Issue 3, August 1999, p. 511.

The underlying assumption of citation analysis is that citations reflect the influence a base patent has had on later developments. Gay and Le Bas, Econ. Innov. New Techn., Vol. 14(5), July 2005, pp. 333-338. This would certainly be true in the case of patents describing breakthroughs or a stepchange innovation. However, it is important when identifying a technologically brilliant partner to discern whether its patents are highly cited due to stepchange technology innovations or merely for purposes of completeness in succeeding applications.

This review identifies qualitative corrections where breakthrough citation signatures may not equate to high patent value.

Within Technology Class Citations

The answers surrounding whether a cited patent influenced a later application can be unclear even when analyzing individual patents with identical class and subclass categorizations. To solve this, a situation is identified where an external force has driven vigorous competition into a market, and stepchange versus completeness issues are less of a question.

In 2000, the National Institute of Standards and Technology published preliminary work regarding new, open-flame fire resistance standards for mattresses. In 2001, California enacted Assembly Bill 603, requiring the California Bureau of Home Furnishings to develop an open-flame standard for all mattresses and foundations sold in California. Patent applications for flame-resistant barriers that may be able to pass these proposed tests surged thereafter, as shown in the chart below.

[IMGCAP(1)]

The chart's snapshot is of 100 patent applications in the fire resistant textile industry gleaned from a Patent Cafe, ICO concept search query. “Premier Latent Semantic Analysis Natural Language Claims Based Patent Search Engine,” PatentCafe', April 27, 2005. This tool ranks all published applications by relevance to a supplied topic of query, and allows those most semantically related to be graphed as shown below (because U.S. patent applications typically publish 18 months after application unless early publication is requested, all data shown here for the period after June 30, 2003 reflect only early publications). In mature markets, this graph will typically be well distributed as a function of time.

It is hypothesized that as the applications issue, the patents presently existing in the technology will obtain breakthrough citation signatures but only because of their being cited for completeness by the applications. Citation signatures obtained in situations where markets spur rapid innovation, likely embody what has been characterized as the “potentially dangerous” within-technology class citations and are independent of technology quality or potency. Maurseth, Econ. Innov. New Techn., Vol. 14(5), July 2005, pp. 351-374.

USPTO Examination Effects

Traditional patent quality indicators segregate inventions by field of innovation according to the USPTO classification system. Lanjouw and Schankerman, Economic Journal, 114 (April), pp. 441-465. However, when a field of invention traverses traditional categorization, the manner in which individual applications are examined has dramatic impact on both citation signature and the importance of a given citation.

Two inventions related to Mass-Spectrometry, US Patent No.'s 5,130,538 and 5,135,870 were issued within 1 month of each other in 1992. The '538 patent was granted by examiners in Technology Center (“TC”) 2800, won the 2002 Nobel Prize in chemistry and has been cited 23 times. The '870 patent was granted by examiners in TC 1700, was classified in four times the number of class/subclasses, and has been cited 58 times. In this instance, the '538 patent could reasonably have been placed in more classes and subclasses than it actually was.

The effect of this classification variance appears to be twofold. First, the '870 patent has an increased likelihood of being cited for completeness in an aggressive market as described above. Second, because examiners tend to most closely search the default set of classes/subclasses to which they are assigned, patents with more cross-references are more likely to be cited by USPTO examiners. Therefore, the '870 patent is relevant prior art to a wider group of examiners.

Expanding on this second point, examiner citation of art has been inferred to artificially inflate the effect of “knowledge spillover,” that is, the impact one reference had on later work. Duguet and MacGarvie, Econ. Innov. New Techn., Vol. 14(5), July 2005, pp. 375-393. For the purpose of identifying technological brilliance, when an examiner objectively states that a given reference is directly related to an application, presumably more so than all others within the same class and subclass of a reference invention, this should be intrinsic evidence that the cited patent is of technological importance.

It is hypothesized that the TC in which an application is examined affects how closely the references cited by the examiner, and potentially all references within the application, relate to a given patent.

This is supported by the below chart showing data from the USPTOBoard of Patent Appeals and Interferences (“BPAI”). “BPAI Receipts and Dispositions by Technology Centers,” www.uspto.gov/web/offices/dcom/bpai/docs/receipts/index.htm, April 27, 2005.

[IMGCAP(2)]

When an applicant believes that the examiner has erred in maintaining a rejection, the applicant has the right to appeal to the BPAI. Assuming across all TC's a uniform distribution of art related appeals, the below chart illustrates cases where the examiner was completely affirmed as a percentage of the total number of cases disposed by the Board within each TC.

In the incident case, examiners from TC 1700, chemical engineering, are affirmed nearly 20% more often than TC 2800, electrical engineering. It may be a reasonable extension therefore, that citations by the 1700 examiners are most likely to be of direct technological importance.

The overall effect of this analysis is mixed. When a patent has a breakthrough citation signature, it should be possible to weigh references cited by the examiner as likely implying the base patent was of technological importance. However, if a patent is broadly categorized, both the number of examiner citations and the number of completeness citations may be inflated.

Patent citation analysis is often the most analytic, quantifying means employed in patent evaluation. As a path forward, all strategists evaluating the technological potency of patents having a high number of patent citations, should be encouraged to consider breaking the citations into groups, and analyzing each group in light of forces in the market and within the examination process itself.

The real value of a possible acquisition or joint venture partner is often in the potential of their next invention. As advisers, patent managers, and intellectual asset strategists, it is critical that the methodology and tools used for assessing patent quality accurately reflect a would-be ally's technological potency.

Recent studies have drawn positive correlations between the increasing or decreasing quality of a company's patent estate and its stock market value. Lanjouw and Schankerman, Economic Journal, 114 (April), pp. 441-465. Patent quality, in the broadest sense, is most often tied to the frequency a patent is cited by other patents. Hall, Jaffe, and Trajtenberg, Royal Journal of Economics, Jan. 2004. One study correlating market value with forward citations found that patents valued at $1 million received an average of about eight citations, while patents valued at $100 million received an average of about 14. Harhoff, Narin, Scherer, and Vopel, Review of Economics & Statistics, Vol. 81, Issue 3, August 1999, p. 511.

The underlying assumption of citation analysis is that citations reflect the influence a base patent has had on later developments. Gay and Le Bas, Econ. Innov. New Techn., Vol. 14(5), July 2005, pp. 333-338. This would certainly be true in the case of patents describing breakthroughs or a stepchange innovation. However, it is important when identifying a technologically brilliant partner to discern whether its patents are highly cited due to stepchange technology innovations or merely for purposes of completeness in succeeding applications.

This review identifies qualitative corrections where breakthrough citation signatures may not equate to high patent value.

Within Technology Class Citations

The answers surrounding whether a cited patent influenced a later application can be unclear even when analyzing individual patents with identical class and subclass categorizations. To solve this, a situation is identified where an external force has driven vigorous competition into a market, and stepchange versus completeness issues are less of a question.

In 2000, the National Institute of Standards and Technology published preliminary work regarding new, open-flame fire resistance standards for mattresses. In 2001, California enacted Assembly Bill 603, requiring the California Bureau of Home Furnishings to develop an open-flame standard for all mattresses and foundations sold in California. Patent applications for flame-resistant barriers that may be able to pass these proposed tests surged thereafter, as shown in the chart below.

[IMGCAP(1)]

The chart's snapshot is of 100 patent applications in the fire resistant textile industry gleaned from a Patent Cafe, ICO concept search query. “Premier Latent Semantic Analysis Natural Language Claims Based Patent Search Engine,” PatentCafe', April 27, 2005. This tool ranks all published applications by relevance to a supplied topic of query, and allows those most semantically related to be graphed as shown below (because U.S. patent applications typically publish 18 months after application unless early publication is requested, all data shown here for the period after June 30, 2003 reflect only early publications). In mature markets, this graph will typically be well distributed as a function of time.

It is hypothesized that as the applications issue, the patents presently existing in the technology will obtain breakthrough citation signatures but only because of their being cited for completeness by the applications. Citation signatures obtained in situations where markets spur rapid innovation, likely embody what has been characterized as the “potentially dangerous” within-technology class citations and are independent of technology quality or potency. Maurseth, Econ. Innov. New Techn., Vol. 14(5), July 2005, pp. 351-374.

USPTO Examination Effects

Traditional patent quality indicators segregate inventions by field of innovation according to the USPTO classification system. Lanjouw and Schankerman, Economic Journal, 114 (April), pp. 441-465. However, when a field of invention traverses traditional categorization, the manner in which individual applications are examined has dramatic impact on both citation signature and the importance of a given citation.

Two inventions related to Mass-Spectrometry, US Patent No.'s 5,130,538 and 5,135,870 were issued within 1 month of each other in 1992. The '538 patent was granted by examiners in Technology Center (“TC”) 2800, won the 2002 Nobel Prize in chemistry and has been cited 23 times. The '870 patent was granted by examiners in TC 1700, was classified in four times the number of class/subclasses, and has been cited 58 times. In this instance, the '538 patent could reasonably have been placed in more classes and subclasses than it actually was.

The effect of this classification variance appears to be twofold. First, the '870 patent has an increased likelihood of being cited for completeness in an aggressive market as described above. Second, because examiners tend to most closely search the default set of classes/subclasses to which they are assigned, patents with more cross-references are more likely to be cited by USPTO examiners. Therefore, the '870 patent is relevant prior art to a wider group of examiners.

Expanding on this second point, examiner citation of art has been inferred to artificially inflate the effect of “knowledge spillover,” that is, the impact one reference had on later work. Duguet and MacGarvie, Econ. Innov. New Techn., Vol. 14(5), July 2005, pp. 375-393. For the purpose of identifying technological brilliance, when an examiner objectively states that a given reference is directly related to an application, presumably more so than all others within the same class and subclass of a reference invention, this should be intrinsic evidence that the cited patent is of technological importance.

It is hypothesized that the TC in which an application is examined affects how closely the references cited by the examiner, and potentially all references within the application, relate to a given patent.

This is supported by the below chart showing data from the USPTOBoard of Patent Appeals and Interferences (“BPAI”). “BPAI Receipts and Dispositions by Technology Centers,” www.uspto.gov/web/offices/dcom/bpai/docs/receipts/index.htm, April 27, 2005.

[IMGCAP(2)]

When an applicant believes that the examiner has erred in maintaining a rejection, the applicant has the right to appeal to the BPAI. Assuming across all TC's a uniform distribution of art related appeals, the below chart illustrates cases where the examiner was completely affirmed as a percentage of the total number of cases disposed by the Board within each TC.

In the incident case, examiners from TC 1700, chemical engineering, are affirmed nearly 20% more often than TC 2800, electrical engineering. It may be a reasonable extension therefore, that citations by the 1700 examiners are most likely to be of direct technological importance.

The overall effect of this analysis is mixed. When a patent has a breakthrough citation signature, it should be possible to weigh references cited by the examiner as likely implying the base patent was of technological importance. However, if a patent is broadly categorized, both the number of examiner citations and the number of completeness citations may be inflated.

Patent citation analysis is often the most analytic, quantifying means employed in patent evaluation. As a path forward, all strategists evaluating the technological potency of patents having a high number of patent citations, should be encouraged to consider breaking the citations into groups, and analyzing each group in light of forces in the market and within the examination process itself.