Methodology & Data Sources

Primary Data Sources

PlainPatent uses USPTO PatentsView grant data and CPC classifications for the aggregates shown on this site:

• USPTO PatentsView - research platform built in partnership with USPTO providing cleaned, disambiguated bulk data exports from the patent grant database. We use the 2015–2025 grants dataset (3M+ patents). Licensed CC BY 4.0 International.
• Cooperative Patent Classification (CPC) - joint USPTO/EPO classification scheme; the canonical authority for technology-domain framing on this site.

What We Track

• Company (Assignee) Rankings: Patent counts by organization, with year-over-year trend analysis and technology diversity scores.
• CPC Technology Classifications: Cooperative Patent Classification codes identifying the technical domain of each patent. CPC is a hierarchical system developed jointly by the USPTO and European Patent Office.
• Grant-Year Aggregates: Annual counts of US utility patents by grant year, assignee, and CPC subclass.
• Technology Trends: Grant counts by CPC class over time. They describe the recorded grant dataset, not application timing, legal status, or commercial outcomes.

How We Process The Data

We download USPTO PatentsView bulk data files for patent grants, assignees, and CPC classifications, then process them into aggregate statistics:

• Assignee disambiguation is handled by PatentsView, we use their cleaned assignee IDs rather than performing our own entity resolution
• Patent grant counts reflect USPTO records without modification
• Technology diversity scores count the number of distinct CPC subgroups (4-character) in an assignee's patent portfolio
• Trend analysis computes year-over-year growth rates in patent volumes by company and technology class
• Innovation scores are computed from a composite of granted-patent volume, grant velocity, technology diversity, and claim depth

Our pipeline processes each included grant record individually, preserving the source data while computing aggregate metrics at the company and technology-class levels. No count is fabricated, interpolated, or editorially modified. The displayed counts and grant years come directly from PatentsView.

Innovation Score Methodology

PlainPatent assigns each company an Innovation Score from four derived inputs: granted-patent volume, recent versus prior grant-year volume, breadth of CPC technology classes, and average claim count per granted patent. The score is a relative ranking tool for this dataset. It does not measure application activity, legal validity, enforceability, commercial success, revenue impact, or patent quality.

Corpus placement (where each entity sits)

Every company and technology detail page states a corpus-relative placement, #N of M, computed from the same ORDER BY that powers the matching listing or rankings table. Placement is derived from live queries against this portal's USPTO PatentsView extract; it is not typed by an editor and is not a fabricated composite.

• Companies (assignees): rank by patent_count descending among all tracked assignees (same basis as /companies/ and top holders by grants). Ties break on slug ascending so ordinals are stable. Innovation Score also reports a corpus percentile (top X%).
• Technologies (CPC subclasses): rank by patent_count descending among subclasses with at least one grant (same order as /technologies/).

The according-to lead on each detail page restates that ordinal so a reader sees where the entity sits before the raw record fields.

Update Schedule

USPTO PatentsView releases updated bulk data quarterly. We update our database when new data becomes available, typically within 30 days of each PatentsView release. Between updates, newly granted patents will not appear in our database until the next refresh cycle. Historical data remains stable across updates, PatentsView does not retroactively remove records, though assignee disambiguation may be refined.

Limitations

• Patent counts reflect granted patents, not applications. Many applications are pending for 2–4 years before grant or rejection.
• A company's patent count does not directly measure innovation quality or commercial value. Some high-value inventions generate one patent; some low-value ones generate dozens.
• PatentsView assignee disambiguation is not perfect. Some patents may be misattributed to company subsidiaries or predecessor entities.
• Coverage is US patents only. Companies may have significant international patent portfolios not reflected here.
• The score uses aggregate claim counts and classifications. It does not assess claim scope, legal validity, enforceability, licensing value, or forward citations.
• CPC classification is assigned by USPTO examiners and may not perfectly capture the commercial application of a technology.

Not Affiliated

PlainPatent is not affiliated with the United States Patent and Trademark Office, PatentsView, or any government agency. We are an independent data portal presenting publicly available patent data in a more accessible format.