November AMP_Digital

A D V A N C E D M A T E R I A L S & P R O C E S S E S | N O V E M B E R / D E C E M B E R 2 0 1 9 2 0 consideration by analyzing not only how many patents the key players in a field have, but also in what technology sub-areas they are focused and when they have been inventing in this area. While the patent analytics shown here do not by themselves single out the best strategy moving forward, they do use facts to narrow down the options to consider for further exploration. LEGAL/IP FUNCTION—PATENT PORTFOLIO OPTIMIZATION Driver: A global metals company has developed and patented technol- ogies related to additive manufactur- ing (AM) of parts used in the aerospace industry. Due to the company’s shift in strategic direction, as well as its desire to lower its patent portfolio mainte- nance costs, the company may consider “monetizing” (i.e., licensing or selling) AM IP assets that are no longer vital to its business operations and growth objectives. Questions: How valuable are the company’s patents related to AM for aerospace applications? Whomight rec- ognize that value and be interested in ENABLERS OF PATENT ANALYTICS The creation and use of patent analytics are enabled by three key factors: Structured Data: Patents are highly structured documents that can be organized into searchable databases. A patent’s format is defined by patent authorities such as the U.S. Patents & Trademarks Organization (USPTO). Each U.S. patent has a number of specifically required fields [1] that can be represented as patent “metadata” (e.g., a publication number, a publica- tion date, or patent assignee(s)), as well as a variety of text fields such as the patent title, abstract, and claims. In addition, a patent typically includes technology classifications related to its subject matter assigned by patent authorities such as the USPTO or the World Intellectual Property Organization (WIPO). Additional “value-added” classifications and summaries might be provided by companies such as the Derwent division of Clarivate Analytics. The advent of digitization andmachine translation enables patents to be organized into large databases that are relative- ly easily searchable online by any one or combination of several patent fields. Several databases with limited search capabili- ties are available free online, including websites of various country, regional, or global patent authorities [2,3,4] , as well as other open sources such as Google Patents [5] and PatentLens [6] . Other commercially available, fee-based databases such as Derwent Innovation [7] , Questel-Orbit [8] , or GridLogics-PatSeer [9] enable more complex searches, along with more powerful data down- load capabilities. Analytics Software: The fielded data for large numbers of patents, once downloaded, can subsequently be imported into other software packages for additional data mining (to reveal statistical patterns and relationships in the data) or text mining (to reveal concepts expressed in the text fields as well as similarity between text fields or documents). Such capabil- ities can be provided in “shrink wrapped” software applications, such as Microsoft Excel, or in more sophisticated software packages specifically designed for document analytics, such as Derwent Data Analyzer [10] . Visualization Tools: The third enabler of patent landscaping is the now nearly ubiquitous presence and availability of visualization tools [11] . Such visualization capabilities have been rapidly embraced by the patent analytics community. By con- verting vast quantities of patent data into easily understood graphs and charts, key trends and factors that impact those trends can be more easily grasped by the non-expert and communicated broadly across an organization. in-licensing or purchasing these patents? Patent Analytics: The pricing of individual patents for license or sale is typically based on relatively complex tools and methods for non-tangible goods valuation using market-, cost-, or income-based approaches. While such valuations are important for establish- ing patent licensing royalty rates or pat- ent sale pricing, these approaches are expensive and time consuming, involv- ing more than what would generally be required for assessing whether patents should receive further maintenance, to what extent they should be filed in mul- tiple jurisdictions, and if they should be exploited for further development. However, a relatively quick as- sessment of the perceived importance of specific patents in a portfolio can be achieved through metrics that reflect the collective views of entities connect- ed to the patent: those citing the patent, the examiners who ultimately allow or deny the granting of the patent, and the patenting entity itself. At Clarivate Ana- lytics, a Derwent patent strength index (DSI) has been developed that ranks the level of interest of these three groups for all patents in a collection and uses this parameter as a proxy for commer- cial value. While such assessments do not necessarily identify which patents will eventually yield the greatest finan- cial return, they do provide a consensus view (somewhat similar to crowd sourc- ing) regarding which patents are de- serving of priority attention. Figure 4 details the DSI ranking of the metals company’s AM-aerospace patents relative to a peer group of sev- eral hundred patents filed globally since 1996 in this field. As noted, about 70% of the metals company’s patents in this area can be considered either “very strong” or “somewhat strong,” with the remaining 30% of the patents rated as only “somewhat weak” or “very weak.” The higher strength patents (i.e., those of higher perceived value) appear to have higher rankings based on forward citations to those patents, patent grant rates, and global geographic coverage. Once the portfolio is segmented according to strength ranking, the next task would be to identify what entities might be interested in licensing or buy- ing these patents. One likely indicator