How to perform a patent landscape analysis in 5 key steps
Patent filings are occurring at an unprecedented scale. In 2016 approximately 3.1 million patent applications were filed worldwide (8% higher than 2015 statistics) — that’s over 8000 patents filed every single day. And because much of this information may not be published anywhere else, patents offer a one-of-a-kind window into the business, technical and product development activities of corporations, universities, governments and research institutions around the world.
One challenge for modern corporations developing innovative products, is finding the critical and relevant patent information in a given technology area and analyzing that information in a manner that empowers actionable decision making.
Part 1 Recap:
In Part 1 of the Patent Landscape Analysis — An Overview, we set out to explain what a patent landscape is, and why it is so critical to all competitive innovation work. i.e., reduce redundant search, accelerate time to commercialization and protect teams from defensive litigation.
Here, in Part 2, we’ll recap that information and provide an overview of the process and the five steps involved in conducting a patent landscape analysis.
Patent Landscape Analysis – What is it?
Patent landscape analysis, often referred to as patent mapping is a proven multi-step process, employing computer software and human intelligence to review, organize, and extract value from extensive patent search results in a specific technology area. A completed patent landscape analysis project consists of a set of references and accompanying analytics from which important technical, legal, and business information can be extracted.
Patent landscape analysis provides a basis for understanding innovation activity including:
- Which organizations (companies, research institutions, others) are working in the area
- What technologies and industries are being targeted
- How technical problems are being solved and which product features are disclosed
- How long it takes for innovations to reach the market
- Where patents are being filed
- Who are the key inventors.
The information derived from a patent landscape analysis is useful for an organization to:
- Generate novel technology
- Monitor competitor activity
- Design around others’ technology
- Identify licensing and M&A targets
- Reduce legal risks
- Avoid spending time and money on duplicate technologies
- Engage thought leaders from R&D, marketing, legal and business development groups
- Optimize internal R&D processes
- Establish a comprehensive and well-informed IP strategy
Companies and investors realize that intellectual property — patents in particular — can provide significant product differentiation and monetization opportunities beyond product sales. Patent landscaping is a critical process for corporations to derive this value. The key benefits include:
- Reducing costs (R&D – time to commercialization, defensive patent litigation, eliminating redundant research)
- Increasing revenue (offensive litigation/licensing)
To reap the full benefits of a patent landscape analysis, forward looking organizations will establish an Intellectual Asset Management (IAM) team, including representatives from legal, technical, business development, and marketing departments. The combined expertise is essential to derive actionable conclusions from the data.
Patent Landscape Analysis — The Process:
There are five key steps in the process combined with a variety of decisions that are made based on the purpose of the study and the preferred outcome.
- Search, review, and refine the subject matter
- Data cleanup and normalization
- Review data, create categories and populate
- Create charts/tables and visualizations
- Ongoing monitoring and analysis
Before starting any of these steps, it’s critical to identify the purpose of the study and develop a clear understanding of projected outcomes of the analysis. Furthermore, gaining buy-in from key stakeholders on the IAM team (representatives from legal, marketing, R&D, and corporate development), prior to initiating the analysis is essential to establish a project with actionable conclusions. Below are a few examples:
Market entry or new product development – To understand the overall field, the players, technologies, and level of activity
- Freedom to operate – infringement or licensing avoidance – The project aims to design around technologies or identify “white space” (areas that appear unpatented), for limiting infringement risks
- Exploring competitor activity – To enhance competitive and market intelligence. Where are competitors spending resources and what are the next products that they will likely release?
The five key parts involved with performing a patent landscape described below:
1. Search, review, and refine the subject matter
Once the key stakeholders from the IAM team have specified the technical area to be studied and agree on the purpose of the landscape analysis project, the first step is to review current intelligence from the IAM group. Next, an iterative process takes shape. Stakeholders search patents, technical literature and other market related information, review and refine the landscape to limit the study to the relevant results. This process of search, review, and refinement continues until a landscape targeted to the relevant subject matter is achieved. Search is one of the most critical components in developing actionable results. All of the remaining steps of the analysis are dependent on the quality and accuracy of the results found in the search. A successful search strategy will find the relevant results, not miss “counter-intuitive” results and remove irrelevant results.
Here’s an outline of how the search, review, refinement process might evolve:
- Kickoff meeting with key stakeholders i.e. legal, marketing, R&D, and business development
- Review the current state of technical, market, and corporate intelligence
- Search for exemplary patents, non-patent literature including academic articles and patent class codes that are generally related to the subject matter
- Search patent data related to technologies in your area of focus, and other industry specific language to identify relevant examples
- Search non-patent literature databases such as Google Scholar, Elsevier, Pub-Med, Science Direct etc.
- Review the International Patent Class (IPC) and the Cooperative Patent Class (CPC) code hierarchies
- Refine the list of technical/product features to those limitations of interest
- Eliminate redundant and/or irrelevant results
- Determine which international jurisdictions to include
- Where will the products be sold? Where are competitors filing patents?
- Determine a date range to search
- Look at patents filed during the past five years, past twenty years, etc.
- Based on the technical/product limitations, develop and run targeted patent searches across worldwide data
- Test a variety of search strategies
- Full specification searches
- Title abstract claim searches
- Patent class code searches
- Test a variety of search strategies
2. Data cleanup and normalization
Data clean-up and normalization are important tasks for the patent landscape analysis process. It’s not uncommon for searches to return thousands of potentially relevant patents and patent applications.
These results need to be refined based on the technical criteria relevant to the search, and further cleaned up to generate a more accurate output. Normalization provides a means for accurately understanding which organizations are working in a specific field, which inventors work for each organization, how large or small each organization’s portfolio is relative to others in the area, and which organizations are working together. The purpose of this step is to further cleanse the data, ensuring that consistent comparisons are being made during later stages of the analysis.
Communication between the patent analyst and the technical/business team is critical during this phase of the project to ensure that documents disclosing relevant technologies and products are added to the landscape, while those less relevant are set aside for future analysis. Reducing the number of irrelevant results in the landscape helps the enterprise stay focused on the project objectives and increases the likelihood of actionable conclusions. At a high-level the aim of cleaning up and normalizing data is to:
- Reduce irrelevant or incidentally related results
- Remove duplicate, later filed versions of patent documents
- Fix naming discrepancies and data errors
- Keep or remove (depending on the purpose of the study) multiple versions of patents representing the same family
During the data cleanup and normalization phase of the project, the patent analyst, with support from other members of the IAM team clean up the data and apply normalization techniques to ensure non-essential and/or duplicative data is removed and names of assignees and inventors are codified. This includes, for example, combining results from assignees that have been acquired by other assignees, fixing misspellings, and matching multiple derivatives of similar names. An exemplary process for cleaning up and normalizing data is outlined below:
- Remove unwanted data
- Documents with duplicate kind codes (A3, A4, A8, A9)
- Published patent applications that have been granted
- Corresponding foreign applications
- Irrelevant results
- Normalize inventor data
- By inventor first, middle, and last name
- Normalize assignee data
- For U.S. published patent applications without assignee data, by inventor/year
- By original assignee, current assignee, reassignment chain
- By organization, licensee, acquiror, business unit
- By organization type, large, small, university, monetizer etc.
During data clean up, additional decisions will be made such as whether to select a single representative family member for each innovation, or to select any patent family member that the search query returns. If choosing the former, the earliest patent family member would be selected. If choosing the latter, all results that match the search query would be analyzed.
Other decisions that need consideration are:
- How to proceed with U.S. published applications that have granted
- How to proceed with continuation, continuation-in-part or divisional patent applications.
The decision to keep or remove such results depends on multiple factors, and the outcomes of such decisions impact overall analysis for users to understand:
- Overall number of patents filed in each family
- Number of unique innovations vs. follow-ons and improvements
- Geographic filing strategies
- Patent filing velocity
- Time to first examiner office action or patent allowance
- Different versions of claims that exist between published applications and subsequent patent grants
3. Create categories and populate
The next step in the patent landscaping process involves a thorough review of the information by technical and patent experts on the IAM team to classify the results into more manageable and useful categories. The experts must create the category schema as well as populate the categories with relevant patents. Due to the overlapping nature of some categories, it is likely that patents will fall into multiple buckets. The patent categorization step is iterative by nature and results in a valuable hierarchy that supports actionable insights to analyze the data from multiple points of view.
Determining the appropriate category schema for a patent landscape analysis depends largely on the underlying goals of the exercise. Whether you’re trying to find the “white space,” (areas that appear unpatented), understand what areas or jurisdictions competitors are working in, or how active one technology area is relative to another, it is essential to clearly define the project goals and apply them to the categorization process.
During this step it is also possible to add custom comments, highlight text, or add other meta-data to patent and technical information, enabling the enterprise to track and monitor key criteria related to the landscape project.
Here is an exemplary outline of the steps involved with creating and populating the categories:
- Category creation
- Patent and technical experts that are ‘Skilled in the Art’ manually read patents (title, abstracts, claims, claims only, or full specifications depending on project scope) to identify category schema through an iterative process
- R&D, business, and/or IAM teams review internal products and technologies to identify categories
- Apply the patent classification codes from the Cooperative Patent Classifications hierarchy (CPC), or the International Patent Classifications hierarchy (IPC) to identify categoriesReview non-patent literature which typically publishes prior to related patents, then generate categories based on review of emerging technology
- Populate the categories with patents
- Patent analysts review the text of the patents to determine which category(ies) each patent should be classified. This is the most accurate yet time consuming method.
- Search strings are created using Boolean operators to sub-search patent data and populate the categories. This is a less accurate and less time-consuming method
- Automated machine learning, semantic, or artificial intelligence systems are used. This computer algorithm based approach has a highly variable accuracy rate, depending on the technical area and requires review by an analyst to confer results
- Categorization criteria
- End user benefits
Traditionally, organizations rely on spreadsheets to manage the categorization process. This is problematic for a number of reasons including difficulties associated with:
- Keeping the spreadsheets up to date
- Ensuring the appropriate technical or legal representatives have the most current version of the spreadsheet
- Lack of full document search capabilities
- High costs during litigation discovery
Ideally, a web-based patent management software is deployed to categorize, archive and monitor the competitive patent information identified in the patent landscape analysis.
4. Create charts/tables and visualizations
The next step in the process of a patent landscape analysis is to create charts, graphs, tables, and patent mapping visualizations. Assuming the search produced a relevant document set and the documents have been thoughtfully categorized, there are a number of different types of charts and visualizations that can be produced. These “graphics” are valuable as an accompaniment to a written or oral presentation and help to portray the “story” of the patent landscape to management. Multi-factorial charts and visualizations should also be presented for maximum impact.
Some of the basic criteria that can be captured during this step are listed below:
- Charts/Graphs and Tables
- Patent filing or publication velocity
- Company vs. product
- Patent family connections
- Time from filing to examination or grant
- Number of office actions
- Heat maps
- Topographic maps
- Citation trees
- Product/patent matrices
There is no one size fits all process for performing a patent landscape analysis. Variables including the purpose of the study, the technology area, what information to include, what geographies to search and how to categorize and display the data will impact the time and effort involved with performing a landscape analysis. Whether using a traditional spreadsheet or a customized patent management software solution, how the data is managed will have an impact on the outcome of the study, particularly with respect to monitoring and keeping data current. A well-executed and thoughtful patent landscape analysis that involves diverse teams with multi-disciplinary backgrounds will certainly result in actionable conclusions that keep an enterprise steps ahead of the competition.
5. Ongoing monitoring and analysis
Completing the initial landscape analysis is a time intensive process. Once the initial patent landscape analysis has been completed on the backlog of information (i.e. data published up to the date of performing the analysis), keeping the landscape up to date requires limited maintenance while offering significant benefits.
Most patent mapping studies can take anywhere from several weeks to many months to complete. These projects often contain thousands of patents published during the previous 5,10, or 20+ years. However, once the search strings and filtering criteria are established to create the initial landscape, keeping the data up to date requires a limited time commitment.
Anywhere from 10 to 30 patents publish each month, relevant to most defined technology areas. At this point, a member of the IAM team can quickly review the new results, categorize them based on the category schema created in the initial analysis (or create new categories if necessary), and add custom comments or meta-data based on the type of data captured in the initial landscape analysis. In addition, charts, graphs, and visualizations can be updated on a monthly or quarterly basis.
Gaining insight into competitor activity as soon as patents publish enables the enterprise to quickly make business decisions such as:
- How to proceed with its own R&D, product development and patent filing strategies
- Whether to file a pre-issuance, inter-partes or post-grant challenge
- To pursue a licensing negotiation, begin a design around project or initiate litigation
Maintaining a searchable, centralized database of updated competitive patent information relevant to the enterprises R&D and business assets enables the enterprise to:
- Stay on top of potential M&A and licensing opportunities
- Reduce redundant research and development efforts and patent filings
- Predict competitor behavior and future product releases
- Reduce costs associated with litigation discovery
Why is the patent landscape process so important? Patent information provides a goldmine of information available to everyone on the internet, including your competition. Patent landscape analysis provides a map for navigating the competitive landscape, and is essential in establishing strong innovation identification, capture, and management programs at any enterprise.
Implementing our 5-step process for executing a patent landscape analysis, outlined in the pages above, results in actionable conclusions for the enterprise to make more well-informed and strategic IP, R&D and business decisions.