About 80% of what data scientists do is preparing and cleaning messy data and data scientists studying USPTO data are no exception.
In order to compile a Business Intelligence report for any assignee, Juristat must collect all the published applications and granted patents for the assignee together to perform an analysis. Finding all the applications for a given assignee isn't quite as simple as just running an application or patent search. USPTO data can contain hundreds (or thousands) of assignee name variations for a single company including misspellings, typos, alternative names, subsidiaries, divisions, etc.
For example, Juristat's report for IBM includes applications filed on over 4,000 different versions of the name "IBM". A small sample of the variation in names for IBM is below:
Juristat's data scientists carefully curate applications to compile a single report for any given entity, collecting all of these alternative names and subsidiaries together into a single report to provide the most complete picture of an entity's patent prosecution metrics possible.
Why can't I find a company that I know has applications?
If you're searching for an assignee and finding no results, there are two possible reasons:
1. The assignee you're looking for is a subsidiary of a larger company.
By default, Juristat groups applications of a subsidiary under the parent organization for purposes of compiling a Business Intelligence report. For example, take a look at the Key Metrics table from the Business Intelligence report for Berkshire Hathaway.
The top row of the table shows Berkshire Hathaway's total number of filed and disposed applications. This includes the applications for Berkshire Hathaway's many subsidiaries, including Lubrizol Corporation, Duracell Inc., Benjamin Moore & Co., and many more.
Juristat does this because every organization is different. Some parent organizations may oversee the patent filings of their subsidiaries while others may take a more hands-off approach. Many of our clients wish to see an organization's entire patent portfolio at once; however, if requested, we are happy to separate a subsidiary into its own report.
2. The entity has relatively few patent application filings.
Even if a company is large, it doesn't necessarily mean that it files a lot of patent applications or has a large patent portfolio. For example, consider Wal-Mart Stores, Inc. Wal-Mart is one of the largest employers in the country and yet, it has relatively few granted patents. In 2015, Wal-Mart Stores, Inc. received roughly 72 patents. While that may sound like quite a few patents in isolation, compare that to IBM, our example above, which received over 7,000 patents in that same year.
As discussed above, Juristat's data scientists carefully curate applications together into a report for any given assignee. However, with over 30,000 entities receiving patents in a single year, Juristat prioritizes the curation of data for larger volume filers unless we know that a smaller filer is important to our clients.
So how do I get the data about smaller filers or subsidiaries?
You can use the "raw names" filter (located under the assignee filter) to search the un-edited names of assignees on applications. Doing that will allow you to pull up the subsidiary applications and clicking to the Charts view will allow you to view the business intelligence-style graphs on the applications belonging to that entity. Here's more on how to us the "raw names" function.
Also, we're happy to quickly compile a report for you for any entity you're interested in (most often, within 24 hours). In addition, if you're putting together information for a presentation, client pitch, or project, it's always a good idea to give Juristat 24 hours notice so we can make sure the data presented in a company's Intelligence report is as complete and up to date as possible.