Patent Agent (AGT) Normalization

You might be surprised to learn how inconsistent agent (AGT) data is in raw patent records. Large law firms file patents on behalf of their clients using, in some cases, hundreds of variants of their name, and even small firms will often have a dozen or more.  It is nearly impossible to analyze patent attorney activity without normalization.

One practice that is fairly common is to append the firm name with the prosecuting attorney or agent. From a database perspective, each of these variants is treated separately, so doing any kind of meaningful analysis on agent data is virtually impossible.

To illustrate, a large firm such as “OBLON MCCLELLAND MAIER & NEUSTADT” has 327 (and counting) variants of their name, and Oblon doesn’t even let their attorneys append the firm name with their own. Other firms like “SUGHRUE MION PLLC” have 323 variants many of which have a person’s name appended to them like SUGHRUE MION PLLC BRIAN HANNON W. In these cases, the AGT names are all normalized into the firm name “SUGHRUE MION PLLC.”

Agent data is normalized for all international patent data collections.

Updated Field Codes

The AGT field code now searches the normalized agent field.

The new AGT_ORIG field code searches the un-normalized version of the agent field. AGT_ORIG is useful to find patents with a specific firm/attorney combination that may have been destroyed in the normalization process. All AGT filters and facets are based on the normalized variant.

What’s the Outcome?

Now it is possible to analyze what attorney firms are working for which companies and get results you can rely on. Consider the following charts run in AcclaimIP:

Top patent firms prosecuting patents for Google Inc

Top patent firms prosecuting patents for Google Inc

You have to remember that 51.4 percent of Google’s patent portfolio has been acquired. As a result, before running the chart, make sure you only chart Google’s organic patents, since these are the most relevant to predicting Google’s hiring decisions. The inverse analysis or a combined analysis (using multi-series charting) is also interesting, however.

Longitudinal study of patent prosecution activity of the top 10 patenting firms used by Google Inc.

Longitudinal study of patent prosecution activity of the top 10 patenting firms used by Google Inc.


The chart above used the multi-series function to chart Google’s top 10 agents over time. Overall, the number of organic Google patents is growing tremendously, but not all their large attorney firms are enjoying that same growth in the account.

Are there any Gotchas?

Yes, agent data is not a required field on patents. Some patents have no data in their AGT field. However so far in 2015, 92.4 percent of all granted patents contain AGT data. Patent owners such as Philips, Verizon, Marvell, Covidien, Siemens, and Western Digital, among others, often publish documents with no AGT data. Do you suppose these are prosecuted in-house?

Following is a chart illustrating the completeness of AGT data from 1996 through today. As you can see, the percentage of patents with data has been consistent for the past 20 years–in the low to mid 90th percentile.

Chart shows the percentage of patents with and without data in the AGT field.

Chart shows the percentage of patents with and without data in the AGT field.

Also recognize that AGT data is no longer published on US patent applications. The most recent US application that contains AGT data was published on February 24, 2011. Why Should You Care You’d be surprised how often I show these analyses to our law firm clients and they are shocked what the data tells them. They might be growing in a large account like Google, but in terms of percentage they are losing out on new business and losing hundreds of thousands of dollars of potential revenue!

Until you understand your situation in each of your accounts and prospects, you don’t know how to find a solution and grow your business. These analyses are very easy to create, and there is no reason every attorney firm can’t have complete visibility into all their clients and prospects. Deeper Analysis with Multi-Dimensional Matrices Do you get business from a large customer in some technologies, but not others, even though your firm is fully capable of prosecuting the patents?

Matrix analyses found in the Analyst version expose, not only how many patents you and each of your competitors prosecute, but also in which technical areas you are relatively strong or weak. An accurate matrix of this type is only possible with normalized AGT data. Many of our large patent owner customers use AGT matrices to understand which firms are getting what business in which technologies. Matrix landscapes provide a simple, always up-to-date view of AGT activity in multiple dimensions.

Who Gets Normalized AGT Data

We’ve added the normalized AGT field and the related charts and filters to both the Searcher & Analyst seats. Of course, some AGT analyses are only possible with additional features in the Analyst version, such as multi-series charting, evolution charting and detailed technology-agent matrices.

New Charts

You’ll find a new “Chart by Agent” chart in the Searcher and Analyst versions. The Agent Evolution chart is a Analyst-only feature (as are all evolution charts), and multi-series charting with AGT field is a Analyst feature as well. The filters and facets use the new normalized AGT field, and are available in all versions of AcclaimIP.


Take a close look at AGT data in AcclaimIP. I think you’ll be amazed what you can learn about your clients or your attorney firms!

Post navigation