Using patent information for organisational decision making and forecasting
Patent landscape reports support informed decision making and are designed to efficiently address the concerns associated with making high-stakes decisions in technologically advanced areas with maximum confidence. For many years decision makers operated based on personal networks and intuition. With the introduction of patent analytics and patent landscape reports, it is now possible for these critical decisions to be made with data-driven approaches that deliver informed choices and lower risk profiles. The insight gained from the preparation of a patent landscape report can be applied to almost any organisation engaged in the evaluation of technology and its impact on society.
While patent landscape reports are useful instruments for informed decision making, producing one can be time intensive and expensive. An organisation willing to devote the necessary resources to generate a patent landscape report often does so when it is preparing to make a significant monetary or headcount investment in developing or moving into a technology area. It is critically important to ensure that a patent landscape report is prepared properly in order to ensure that the insight it provides is accurate and directed towards the key issues associated with technological implementation.
This chapter provides practical instructions on the use of patent families, how the decision to use each type will impact on the statistical analyses conducted and how forward citations can be used to identify influential portfolios. These elements, when properly conducted, will result in the generation of a patent landscape report that will deliver real value to the corresponding decision maker. An example will be provided as a means for instructing practitioners on some of the steps required in generating patent landscape reports, which will also be useful for individuals who have patent landscape reports produced for them by third-party consultants and are looking for a means to evaluate them.
Since patents provide a right to exclude others from operating in a technological area, they have business and legal implications. Receiving a patent can be a reasonably expensive endeavour, costing from $10,000 at the low end of the scale to between five and 10 times that for more complicated applications. The substantial costs involved when pursuing a patent indicate high interest and investment in the area.
Patents are also critical sources of information that may not be found elsewhere. A 1986 paper citing a report from 1977 claimed that 80% of the information found in patents is not found elsewhere. A more recent study suggested that the percentage in some technical areas may even exceed 95%. While patent data can be difficult to work with, and misleading if not handled correctly, it is critical to a thorough understanding of most technological areas. Jacob Schlumberger best encapsulated these feelings more than 50 years ago when he wrote: “We have the choice of using patent statistics cautiously and learning what we can from them, or not using them and learning nothing about what they alone can teach us.”
Patent statistics and analytics are the tools used to study patent data. One of the most frequently used means of delivering insights, and the path for organisations associated with this sort of research for clients and decision makers, is the patent landscape report. The World Intellectual Property Organisation (WIPO) defines a ‘patent landscape’ as: “An overview of patenting activity in a field of technology. A landscape normally seeks to answer specific policy or practical questions and to present complex information about this activity in a clear and accessible manner. Industry has long used patent landscapes to make strategic decisions on investments, research and development (R&D) directions, and competitors’ activity as well as on freedom to operate in introducing new products. Now, public policymakers are increasingly turning to landscaping to build a factual foundation before considering high-level policy matters, especially in fields such as health, agriculture and the environment.”
In this definition, there is a clear emphasis on creating patent landscape reports that provide insights for supporting informed decision making and lowering the risks associated with making high-stakes decisions in technological areas.
Patent families – impact on statistical measures in patent landscape reports
Due to the territorial character of the patent system worldwide, patents must be applied for in individual jurisdictions. These priority claims lead to relations between different national patent applications (so-called ‘patent family relations’). Since international patent treaties expressly permit the claiming of more than one priority, complex family relations may exist depending on whether two applications share priorities in full, partially or only indirectly (ie, through other related applications).
This creates a situation where a single invention might have many individual patent documents associated with it, depending on the number of countries in which the applicant sought protection and how many stages of publication the case has gone through.
In order to identify the correlation between inventions and the numerous patent documents that can be associated with them, the concept of a patent family was created. There are a variety of different definitions provided depending on how tightly linked the documents are based on priority filings. According to the WIPO Handbook, these are defined as:
the domestic patent family – consisting solely of a single office’s different procedural publications for the same originating application;
the simple patent family – relating to the same invention, each member of which has for the basis of its priority right exactly the same originating application or applications; and
the extended patent family – relating to one or more inventions, each member of which has for the basis of its priority right at least one originating application in common with at least one other member of the family.
Organising patent collections into some form of family is an essential activity, which will have a major impact on how statistics are generated for a patent landscape report. Determining which method will be used and consistently applying it across an entire project will ensure that accurate comparisons can be made between the different attributes being studied. For instance, analysts must determine whether the use of an extended family will severely underrepresent the amount of investment made by an organisation when deciding to use that method. If one of the objectives of the patent landscape report is to identify which organisations have invested the most resources into a technological approach, then a simple family or alternate method would be better suited than the use of an extended family.
In April 2013, just before personal fitness monitors became popular, Jawbone – an early pioneer in the area – had 80 patent applications associated with the Up™ product filed around the world. After removing redundant applications filed in multiple countries, a collection of 29 unique application numbers were discovered. Using an extended family, all 80 documents were collected together as a single family. In an analysis where the documents had been reduced using extended families, Jawbone appeared to be a very minor player with only a single family, but in reality it had made an enormous investment in this space.
Many of the 29 applications had identical original titles and specifications, but differed significantly when looking at their claims. To determine how many distinct technology concepts were covered by this collection the claims of each document were read and categorised. Figure 1 shows that 12 individual concepts were discovered within the first claim of each application.
Figure 1. Breakdown of Jawbone Up™ portfolio by application number and technology concepts from the claims
The collection represents a particularly diverse set of individual inventions related to the development of a single product – a flexible fitness monitoring device.
While reduction of patent collections to a single extended family member is a popular method for eliminating multiple country filing biases, it can often lead to situations where distinct inventions are removed. When this occurs, the breadth and depth of inventive output from a single organisation can be understated. For this reason, alternate methods for reducing patent collections should be used. In the case of the Jawbone example, more than 80 individual documents were reduced to a single extended family. A better approach involves the use of a simple patent family or one of its close alternatives that more closely represents the true level of invention within the portfolio.
Forward citations – identifying influential patent portfolios
During the prosecution of a patent application, an examiner will look for prior art that may speak to the novelty, obviousness or inventive step associated with an invention. When references of this nature are discovered they are cited within the document during different publication stages.
In the United States there is also a duty of candour that requires applicants to share prior art with the US Patent and Trademark Office (USPTO) during the examination of an application. These documents are also citations and appear on the front page of granted US patents along with the prior art identified by the examiner.
Any discussion of patent citations begins with a root document. The references that the root document cites or references itself are referred to as ‘backward citations’. Conversely, going forward from the root document, any more recent documentation which references the original root document will be referred to as a ‘forward citation’ for the root document.
Citations represent a relationship between two inventions. Studying them provides a means of identifying seminal documents or families that could have had a significant impact on the development of a particular technology. Looking at these in aggregate allows analysts to identify influential portfolios.
The impact of redundant applications and patent families on citation counts should not be underestimated. Citations are based on the referencing of discrete documents, so a recently granted patent may not have any forward citations associated with it, but the corresponding, previously published pre-grant application with the same application number may well have several. An analyst can also look at the entire simple or extended patent family and find additional unique forward citations for that family. These citations must be aggregated in some manner, so that the family being discussed is reconciled to include all of the individual publications that it represents.
Once citations have been reconciled an analyst can start to compare portfolios within a technological space to identify which organisations have the most influential patents in the field. There are a number of ways of doing this, but the simplest involves adding up the number of reconciled forward citations for each family belonging to each company. An illustration of this can be seen in Figure 2.
Figure 2. Count of forward citations to personal fitness monitor portfolios by top patenting companies
In this example, the listing of organisations was ordered by portfolio size from largest to smallest, as opposed to ordering it by number of forward citations. While Philips had the largest number of patent families in the area of personal fitness monitors, it did not have the highest number of forward citations when this chart was created. Theranos, BodyMedia, Valencell and Healthtech had a disproportionately large number of forward citations for the relative size of their portfolios by count of patent families. This ratio of the number of forward citations to portfolio size suggests influential portfolios.
Another way to represent this concept is by using a network diagram to visualise how companies are related to one another in a network based on shared citations. Figure 3 provides a forward citation network diagram for a collection of more than 2,000 cryptocurrency patent families. In this example, the largest portfolio belongs to Mastercard, but within the citation network Visa clearly has the most influence with the largest number of connections to the most additional organisations based on the size of their node and the number of edges to other companies. Other key influencers in the network include Bank of America, eBay, Google, Square and Apple. Some of these are financial institutions, while others are surprising additions in financial technology.
Figure 3. Forward citation network diagram of cryptocurrency patent families by company
Patent analytics driving corporate strategy
In April 2013 Jawbone had a collection of 29 unique applications associated with its Up™ product. All of these documents were based on an initial application filed in 2010 before the product went on sale, and while there were a reasonable number of applications, Jawbone should have been concerned about whether a relatively late filing date would allow it to protect its market share.
The situation with Fitbit was even more dire. During the same April 2013 period, Fitbit had only 13 unique US patent applications, despite being the market leader in the space. These applications did not cover any of the device components or technologies used to assemble it. Instead, they covered only a generic monitoring device with and without a monitor. The pending applications were associated with activity algorithms – particularly stair climbing applications – and were being challenged at the USPTO.
In early 2013 it could be reasonably concluded that Jawbone and Fitbit were exposed in terms of their existing patent portfolios. Both companies were likely to be considering patent-buying programmes, strategic acquisitions or partnerships with other companies that owned valuable intellectual property in the space to address the shortcomings in their own coverage.
It was therefore of little surprise when on 30 April 2013 Jawbone announced that it had acquired BodyMedia – an early pioneer in the fitness monitoring field – for $110 million. It was reported that the acquisition was primarily driven by BodyMedia’s patent portfolio, as opposed to their marketed products. The BodyMedia portfolio was a good fit for Jawbone since it was one of the best in the area with technology depth and breadth, as well as early priority dates. It was a core asset that Jawbone could use to continue building its patent portfolio and protect its products.
Figure 2 clearly shows that BodyMedia had the most influential patent portfolio of any of the companies in the field at the time based on an analysis of their forward citations. The BodyMedia acquisition compounded the value of Jawbone’s portfolio and enhanced its ability to protect its share of the personal-fitness monitors market. In this case, patent analytics suggested that BodyMedia would be the best choice for Jawbone or Fitbit to acquire, and a few weeks later that transaction came to pass.
The choice of patent family reduction is the single biggest factor in determining the inventive output of an entity in these reports. Forward citations are universally recognised as a key value indicator but must be reconciled to be used effectively.