Illuminating the dark side of patent data

Patent data includes a huge quality and quantity of information. Qualitatively, the information is a multi-faceted collection of varied technical disclosures, legal content and government records. It includes technical jargon, plain language, graphics and reference citations. Quantitatively, for example, World Intellectual Property Organisation records include approximately 65.1 million patent documents as of August 31 2017 (Table 1), while European Patent Office records include approximately 99.6 million patent entries up to August 31 2017 (Table 2).

Patent data has been used by a niche group of professionals as well as for specific – but also niche – purposes. Today, patent data should have wider availability and usage. However, there is no common, unified standard format for patent data, and each national patent office or software developer offers limited functionalities or user interfaces. Most users can access patent data from each national patent office only at a specific level (ie, the application or patent as published, without access to the associated prosecution history) and without many intelligent analytics. This is unacceptable considering the range of big data and analytical methods available.

There is an historical and widely accepted, although absolutely incorrect, belief that patents are not subject to change or challenge. In the United States, this has been exacerbated by the presumption of patent validity. This belief prevents less experienced users from appreciating the dark side of patent data. They are thus unmotivated to create innovative methods to use patent data or to explore the value of patent data outside the traditional areas where it may assist decision makers.

As an effort to make patent data more versatile, this article explores the various challenges caused by the so-called ‘dark side’ of patent data. Both national patent offices and the software industry are urged to overcome these challenges.

Improving national databases

There is a dark side to patent data, a wide range of issues that cast potential negative shadows. Patent professionals’ lack of care, knowledge, skill or experience often leads to errors, which in turn result in either valid applications being rejected, or issued patents being invalidated or limited in claim scope. Either way, the value of patent assets to their owners is reduced. Figures 2 and 3 provide the total number of patent applications filed, abandoned and granted in the United States and China for 2012 to 2016. This data suggests that a significant number of applications were rejected for lacking patentable subject matter, lacking novelty or non-obviousness. The ratio of allowed to rejected cases in each of these countries is considered too low.

The dark side of patent data persists after patent issuance, a fact illustrated by the data presented in Figures 4 and 5, which provide post-grant review proceeding statistics for the United States.

This data shows that despite the belief that patents are not subject to change or challenge, patents are under widespread challenge in the United States from a broad range of challengers.

Patent procurement is a lengthy, complex and specialised process. For example, the process includes invention disclosure, prior art search, draft and review of patent specifications, examination by patent offices and response thereto, and issuance (in the United States, the process may further include continuation, continuation in part or division). There are various stakeholders at each step and at each point. The dark side of this can be attributed to mistakes or a lack of information held by the various stakeholders. The problem will become more aggravated as the number of Chinese patents soars, and further because public access to patent data will continue to be deficient in many countries, including in the United States. While the US Patent and Trademark Office (USPTO) Image File Wrapper system provides excellent data, over the many years that the USPTO has provided patent search capabilities, the search interface has rarely been improved.

Figure 1. Patent documents by patent office (incorporated at the European Patent Office on or before August 31 2017)

The severity of the dark side problem calls for an immediate solution to prevent harm to the public as a result of questionable patent assets affecting wide-ranging business and government decisions. Such questionable assets come in the form of both rights wrongly given up by abandonment and rights that are detrimental to future developments due to wrongful granting. Solving these problems would especially benefit R&D programmes, which will no longer be blocked by poor-quality patents. Of course, technical disclosures in patent data impart reference value to the public regardless of any dark side issues.

Little help available for chief technology officers and CEOs

Patent assets affect business and government decisions. However, the dark side of patent data deployment manifests in decision makers’ lack of confidence in patent landscape reports from patent practitioners. For various reasons, many patent practitioners fail to acquire the requisite breadth and depth of business skills to analyse the data in front of them. This in turn leads to problems in determining patent values. In addition, many practitioners generate flawed patent reports due to misunderstanding or misinterpreting patent data. Sophisticated decision makers from the government, academia or industry sectors do not believe that such reports can be relied on to support their needs. These include setting government policies for the scientific community, strategies for specific industries, exploring new technology, planning for new products, advancing competitive positions in the marketplace, making M&A decisions, allocating R&D resources and deploying patents for monetisation.

For example, in Japan, Taiwan and China, these patent reports can be based on any of the nearly 70 patent landscape models used in those markets. These models are all based on the patent office’s classification system and generally poor-quality patent data. They do not rely on an information matrix (or dashboard) combining intelligence from industry, technology and products. Such information is what many users actually consult through different stages of R&D, and through different phases of patent asset lifecycle assessments. Consequently, these existing patent landscape models cannot generate interpretable information that would be more useful to decision makers. The patent landscape methods practised in the United States and Europe do not transcend a mere collection of data from databases into actionable intelligence information. In short, globally existing patent landscaping methods are ill equipped to support the needs of decision makers such as CEOs and chief technology officers.

Figure 2. US utility patent applications from 2012 to 2016

Creating value from patent data

The inadequacy of many existing approaches is unfortunate, because when properly made available and analysed, patent data has broad applications. This information can and should assist innovators, entrepreneurs, established businesses and patent asset management firms alike. Patent data also aids sensible decision making in industry, government and academia – look no further than Japan, China and Taiwan, where governments rely on patent information to craft science, technology and industrial policies. Below, we briefly address various contexts in which this data can support to decision makers.

Science and technology policies

Patent data can depict a technology’s global context and trends. It can also help industry and academia to properly allocate R&D budgets that are used to:

• conduct basic and applied research;
• build momentum for the next round of industrial transformation; and
• focus development in much-needed technical areas, such as material science, biology, wireless communications, artificial intelligence and the Internet of Things.

Industrial policies

Patent data can identify emerging technologies or organisations affecting or disrupting established industries or traditional investment and M&A networks or old supply chains. Such examples include the impact of light detection and radar and advanced driver-assistance systems on the self-driving car industry; artificial intelligence’s impact on the software and hardware industry; the impact of biotechnology in the pharmaceuticals and medical treatment field; the impact of microLED on the display and lighting industry; robot technology’s impact on manufacturing and agriculture businesses; and blockchain’s impact on the financial technology sector.

Figure 3. China utility patent applications from 2012 to 2016

Source: www.patentcloud.com, August 31 2017

Figure 4. Trial status of petitions from the US Patent and Trademark Office, Patent Trial and Appeal Board and America Invents Act

These figures reflect the latest status of each petition. For example, the outcomes of decisions on institution responsive to requests for rehearing are incorporated. Once joined to a base case, a petition remains in the Joined category regardless of subsequent outcomes.

Innovators and start-ups

Patent data has the potential to help entrepreneurs choose R&D subjects with precision, implement a R&D strategy effectively and connect entrepreneurs to various industry, M&A and supply chain networks. Patent data can also help to reduce start-up risks and failures. Examples include security and cryptography technology as applied to online payments; big data technology as applied across industries; sensor technology as applied to fields such as medical devices, robotics and the Internet of Things; and algorithms as applied to business modelling.

Mergers and acquisitions

Patent data is capable of depicting with specificity the global competition landscape in technology, products and patent assets. Patent data will facilitate mergers and acquisitions in the industry and reshape the markets and competition profiles. For example, Foxconn Group invested in Lytro’s light field imaging platform technology and, through its Sharp acquisition, also obtained access to 5G standard-essential patents and display technologies. Qualcomm’s NXP acquisition gave it ownership in products, customers and patents outside the communications field. Intel acquired Mobileye to own products, customers and patents in the self-driving car sector.

Technology trends

Patent data can detail global context and trends for technologies, products and patent assets in order to accurately predict future market needs in terms of product and technology features. This enables better planning and efficient resource management. The technology trending now includes artificial intelligence, 5G, IGZO, unmanned aerial vehicles, biopharmaceuticals and digital medical devices.

Talent deployment

Patent data contains a global picture of inventors and designers of technology products and patent filings. Patent data enables talent recruiting across borders.

Patent asset management

Patent data can be used to map the technical content and context of global patent assets as a whole, as well as the distribution and relatedness of these assets. It can also be used as a tool to deploy quality patent assets from various countries to form high-value patent portfolios. Products with big market shares and high gross margins can be developed with the support of high-quality and high-value patent portfolios. Patent data can also be used to run patent monetisation programmes in order to obtain royalty, sales revenue or cross-licence benefits. Further, patent data can help patent professionals to manage global patent risk and create solutions from technological, commercial and legal perspectives.

Figure 5. US inter partes review petitions terminated, to October 31 2016

A way forward

There are many important uses for properly prepared patent data, but many of the existing approaches to this discipline simply do not produce actionable intelligence. The advent of big data companies specialising in the patent field provides a potential solution to the dark side problems identified above. These companies seek to develop software and big data solutions that:

• normalise various patent databases;
• apply semantic, natural language processing, cross-language, image search and machine learning, and other big data techniques to analyse patent data from various countries; and
• develop algorithms to assess patent quality and patent value.

Moreover, we believe that applying artificial intelligence technology to patent analytics will help to identify and measure potential errors by an inventor, patent attorney or agent, or patent examiner. This will help to prevent the dark side of patent data from worsening. It will also help to significantly reduce the number of patents granted by focusing on quality and value, rather than quantity. This would go a long way towards restoring true respect for other people’s patents.