Patent research guide

How to do a patent landscape search

A patent landscape organizes public patent information around a technology so a team can see who is filing, where activity concentrates, and what the claims cover. This guide shows how one is put together and where its limits are.

Define the technology scope first

A patent landscape search begins with a scope statement. Describe the technology in plain terms, list the features that matter, and say what is out of bounds. A scope such as "wearable sensors" will return thousands of documents. A scope such as "wrist-worn devices that detect falls using motion sensing" is narrower and more useful.

Decide at the start what the landscape is for. A team choosing a development direction needs a broad view of who is active. A team preparing for investor questions may want a close look at the few patents that seem nearest to its design. The purpose shapes how wide to search and how many documents to read closely.

Set a time window and the jurisdictions you care about. Patents are filed country by country, so a family of related filings can appear in several places.

Pick the databases

Several free public databases cover most of what a business-level landscape needs. Google Patents offers full-text search across many countries and is easy to use. The United States Patent and Trademark Office publishes granted patents and published applications and shows legal status and prosecution documents for US filings. Espacenet, run by the European Patent Office, covers a very wide set of countries and groups related filings into families.

No single source is complete or always current, so it is common to use two and compare. Record which databases you used and on what date, because patent data changes as applications are published, granted, abandoned, or assigned to new owners.

Paid commercial databases exist and add analytics and cleaner ownership data. They are not required to build a useful first landscape.

Keyword and CPC classification searching

Searching works best with two methods used together.

Keyword searching uses the words inventors and applicants write. Because different companies describe the same idea differently, include synonyms and related terms, for example "fall," "drop," and "impact," and use proximity operators if the database supports them. Keep a log of each query and how many results it returned.

Classification searching uses the Cooperative Patent Classification, or CPC, a hierarchical system that assigns patents to technical categories. Once you find a few relevant patents, look at the CPC codes on them and search by those codes. This finds documents that describe the technology in unusual words. Combining a CPC code with a few keywords is a common way to narrow results.

It also helps to follow citations. A relevant patent lists earlier documents it cites and later ones that cite it, and both directions often surface close neighbors.

Filter to a readable set

The first pass usually returns far more than anyone can read. Reduce it in stages.

  1. Remove duplicates and group family members so each invention is counted once.
  2. Read titles and abstracts, and drop clearly unrelated documents.
  3. Skim the claims and figures of what remains, and sort documents into close, related, and background.
  4. Choose the close group for detailed review.

Write down why documents were excluded. This makes the process repeatable and shows a reader that the set was not chosen arbitrarily. For a standard business-level review, the close group is often a few dozen documents at most.

Build a comparison matrix and read claims at a business level

A comparison matrix puts the selected patents in rows and the features that matter to your product in columns. Typical columns include the owner, filing and priority dates, legal status, key technical features, and which markets the family covers. Reading across a row shows what a patent emphasizes. Reading down a column shows which features are common and which few documents mention them.

Claims define what a patent protects, so they deserve the most attention. At a business level, read the independent claims first and note the main elements. Compare those elements with your own design to see where there appear to be overlaps and where there appear to be distinctions. This tells you which documents deserve a closer look from an attorney. It does not say anything about whether a product infringes.

The matrix can also show patterns, such as which companies file most, in which years, and in which technical areas. These patterns give business context to the technical reading.

Limits of a landscape, and when to bring in counsel

A landscape is a research summary of public information. Keep these limits in mind.

  • It is not legal advice, and it does not determine patentability, freedom to operate, infringement, or validity.
  • It reflects only published information. Applications are generally not public for a period after filing, so very recent activity may be missing.
  • Searches can miss relevant documents. A thorough process lowers that risk, and it cannot remove it.
  • Legal status and ownership can change after the search date.

Bring in qualified patent counsel before filing your own application, before relying on any view about freedom to operate, when a close patent appears to overlap with your design, or when you receive a notice from another party. Counsel can analyze claims against your specific product, which a landscape does not do. The relationship between the two is covered in our guide on landscape versus freedom to operate.

ZKK offers Patent Landscape Research as business and market research, which you can pair with market work from our services. To discuss your technology, book a discovery call.

Common questions

Is a patent landscape the same as a freedom to operate opinion?
No. A landscape organizes public patent information for business decisions. A freedom to operate opinion is a legal analysis by a qualified patent attorney. A landscape can help prepare for that conversation, but it does not replace it.
Which free databases are most useful?
Google Patents, the USPTO's public search tools, and Espacenet are commonly used. Each has different coverage and update timing, so many searchers use at least two and record the search date.
How many patents should a landscape include?
It depends on the purpose. A broad landscape may screen hundreds of documents, while the close-review set is usually much smaller. ZKK's standalone service covers up to 15 identified patents, with a custom quote for larger scopes.
Can I do the search myself?
Yes, public databases are open to anyone, and a first pass is a reasonable way to learn the field. A structured process, a documented search log, and careful claim reading take time, and that is where help is often used.

Related

Want a patent landscape built for your technology?

Book a discovery call to define the scope. Patent Landscape Research is a fixed $1,349 for up to 15 patents, and this is business research, not legal advice.

Reports, models, and other deliverables provided by ZKK Consulting LLC are business research and analysis only. They are not legal, regulatory, medical, clinical, tax, accounting, or investment advice, and are not a guarantee of funding, regulatory clearance, reimbursement, patent outcomes, or financial results. Regulatory, patent, and reimbursement research reflects publicly available sources as of the report date. Clients remain responsible for their own regulatory, legal, clinical, and business decisions and should retain qualified professionals where determinations or representation are required. Read the full disclaimers.