All Citations Are Not the Same – Exploring Examiner Citations from US Patent Documents – Part 2 – Using Thomson Innovation

In the previous post, an introduction to citations associated with US patent documents was provided. The process of isolating examiner citations from Thomson Innovation will be the subject of part 2 in this series.

To demonstrate the process of differentiating examiner from applicant citations, forward citations associated with Allure Energy were used as an example. Allure Energy was discussed recently on this blog since they filed suit against Nest Labs for infringing their recently granted patent for a “auto-adaptable energy management apparatus”, or a thermostat. Forward citations, to the Allure Energy portfolio, refers to patent documents that were filed after the ones from Allure Energy and are cited by the applicants, or the corresponding examiners, on either an IDS or Form 892.

A search on Thomson Innovation for Allure Energy US documents generates ten granted patents and twenty-nine pre-grant applications for a total of 39 documents. 28 of these document have at least one forward patent citation associated with them according to the database. As mentioned, most search systems provide the ability to locate forward citations for a collection of patent documents and many provide displays that differentiate examiner citations from ones coming from the applicant. In this example, data from Thomson Innovation was used, since one of the fields available for export, Citing References Details – Patents, puts all of the available data on these citations in a single cell that can be managed in Excel and then manipulated in Google Refine to provide a comprehensive picture of which patent documents, and their corresponding applicants, are citing the Allure Energy portfolio. Looking at the data in this field from one Allure Energy document, US8,082,065, we find the following:

US20120029713A1,US, ,7 (Pre-search),2010-08-02,GEN ELECTRIC | US20120116597A1,US, ,7 (Pre-search),2010-11-09,GEN ELECTRIC | US8386087B2,US, ,0 (Examiner),2010-08-02,GEN ELECTRIC,SPICER LUCAS BRYANT,BESORE JOHN K,WORTHINGTON TIMOTHY DALE,FINCH MICHAEL FRANCIS

The ‘065 patent has three forward citations associated with it (so far) and, in this case, they are separated by a vertical bar (|) between each entry. Within each entry we find six pieces of information associated with each of the three citations. These are separated by a comma and represent the citing patent number, the citing patent country, the type of citation (examiner or applicant) for international reports (blank in this example) and for the US, the citing patent application date and the citing patent applicant. This is a tremendous amount of information associated with a single patent document and it is all captured in a single cell when data is exported in MS Excel format from Innovation. This will be teased apart and cleaned up using Google (Open) Refine but first redundant patent applications and their citations have to be reconciled with the corresponding granted patents that have developed from them.

Previously, the need to consider citations to pre-grant applications when counting citations to grant patents in the US was discussed. In the case of the ten granted Allure Energy patents, all ten have pre-grant applications associated with them, in only one, of the ten cases, both the granted patent and the corresponding pre-grant application don’t have any forward citations. Three of the granted patents have no citations themselves but their corresponding pre-grant applications do. In the case of these nine granted patents the data associated with the pre-grant application needs to be added to the data associated with the grant so that it represents the accumulated citations for the full lifecycle of the document. The table below provides the details of the citations associated with the various versions of the granted patents:

Citations from Allure Energy US Grants and Corresponding Applications – Click to enlarge

If the application number is exported with the publication number and the Cited Reference Details – Patent it can be sorted on in MS Excel so that the pre-grant applications are in the row next to the corresponding granted patents. At this point, the analyst can simply copy the data from the Cited Reference Details – Patent cell for the pre-grant application and paste it on to the end of the data associated with the corresponding granted patent. Remember though, to include a horizontal bar after the data in the granted patent cell before pasting in the data from the pre-grant application. If we use the ‘065 patent, referred to earlier as an example of this, we see that it went from three forward citations on its own to eleven when reconciled with the pre-grant application, and the Cited Reference Details – Patent cell now contains the following:

US20120029713A1,US, ,7 (Pre-search),2010-08-02,GEN ELECTRIC | US20120116597A1,US, ,7 (Pre-search),2010-11-09,GEN ELECTRIC | US8386087B2,US, ,0 (Examiner),2010-08-02,GEN ELECTRIC,SPICER LUCAS BRYANT,BESORE JOHN K,WORTHINGTON TIMOTHY DALE,FINCH MICHAEL FRANCIS | US20110122798A1,US, ,7 (Pre-search),2009-11-24,SILVER SPRING NETWORKS INC | US20110257804A1,US, ,7 (Pre-search),2010-04-14,RAYTHEON CO | US20120023225A1,US, ,7 (Pre-search),2009-07-20,  | US20120130513A1,US, ,7 (Pre-search),2010-11-18,VERIZON PATENT & LICENSING INC | US20120198083A1,US, ,7 (Pre-search),2011-01-27,OPENPEAK INC | US8375118B2,US, ,0 (Examiner),2010-11-18,VERIZON PATENT & LICENSING INC | US8396602B2,US, ,11 (Examiner),2009-07-20,ALLURE ENERGY INC,IMES KEVIN R,HOLLISTER JAMES | WO2012163901A1,WO,X, ,2011-06-02,NOKIA SIEMENS NETWORKS GMBH,DE LA RUE MICHAEL,SINGH ABHINAV,GULAK MACIEJ,NAKAMURAKARE MANUEL,SOEIMA MARCO AURELIO,KJ BOBY

The reconciled Allure Energy forward citation data can now be taken into Refine, as shown in a previous post on expanding patent families from, but this time the multi-value data in the Cited Reference Details – Patent cell will be manipulated by doing the following:

Split multi-valued cells in column Citing References Details – Patents: In this case the separator is the vertical bar | and this will create new rows for each entry in the cell. The collection started with 22 rows, one for each of the patent documents from Allure Energy explored, but after this operation the collection has expanded to 147 rows, signifying that there are 147 forward citations in total associated with these documents.

Splitting multi-value cells using Open Refine – Click to enlarge

Fill down cells in column Publication Number: We need to fill down in our table so we can run a Pivot Table later to associate the citing patents back to the document in the Allure Energy collection that generated the citation.

Fill down operation in Open Refine – Click to enlarge

Split column Citing References Details – Patents by separator: We want to get access to the type of forward citation, examiner or applicant, in our collection and do to so we need to isolate the citation type data into its own column. This is done by splitting the data into a new column using the comma as a separator.

Splitting multi-value cell content into several columns in Open Refine – Click to enlarge

At this point, for the sake of making the resulting spreadsheet generated easier to understand, the new columns created with the split column command are renamed to represent what data they contain. By default, Refine simply places a number at the end of the original column name to signify that a new column was generated.

  • Rename column Citing References Details – Patents 1 to Citing References Details – PN
  • Rename column Citing References Details – Patents 2 to Citing References Details – Patents PC
  • Rename column Citing References Details – Patents 4 to Citing References Details – Patents Citation Type
  • Rename column Citing References Details – Patents 5 to Citing References Details – Patents AD
  • Rename column Citing References Details – Patents 6 to Citing References Details – Patents Citing Assignee

Some additional columns that were created, since in some of the rows the citing assignee data contained entries for more than one assignee were removed as well. Technically, this doesn’t need to be done and the columns can simply be ignored but for the sake of completeness they have been removed in this example. A copy of the code which can be used to perform this same operation on a new collection is included:

Code for Cited Reference Patent Details Transform

This code can be reused on any collection brought into Refine by clicking on the Undo/Redo tab in the upper left corner, pressing Apply… and then pasting the code into the window. In this fashion, the same method for preparing data from Thomson Innovation for analyzing examiner citations can be re-done quickly and without going through all of the steps manually.

Use Redo/Undo to apply JSON code – Click to enlarge

Now that the multiple values in the Citing References Details – Patents have been segregated into new rows, based on the discrete entries, and into new columns, based on the types of data in the field a little bit of cleanup needs to be done on one of the new columns to get it ready for further analysis. As discussed in the original post on Refine, the Citing References Details – Patents Citing Assignee column needs to be cleaned up using the same clustering algorithms used in that original article. Once this is accomplished, looking at the text facet window, sorted by count we find the following:

Blank assignees in citing assignees analysis

Notice at the bottom of the list that 35 of the 147 entries in this collection have “blank” listed for our citing assignee. This happens since a number of the forward citations are coming from a US pre-grant application, where an assignee has not been provided in this field. In reality, since some of these documents cite more than one Allure Energy document, there is actually only twelve assignees that need to be looked up. Once the data has been exported out of Refine, a quick trip back to Innovation, to search for the publication numbers associated with these blank records, will need to be done to discover if Thomson has information on who the likely assignee might be or if the these documents are associated with an inventor.

Once the citing assignees are reconciled a number of interesting observations about the Allure Energy portfolio and who is citing it can be made. Looking at the type of rejections, and whether they are coming from an examiner or not, the 147 forward citations can be classified as follows:

Citation TypeNumber of Citations
7 (Pre-search)90
0 (Examiner)33
11 (Examiner)14
1 (Applicant)3
A (Examiner)2
Y (Examiner)2
12 (Other)1
X (Examiner)1
XP (Examiner)1

Citation type 7 refers to citations on Form 892 for pre-grant applications that have not issued yet. They are listed as being Pre-search on Innovation but since they come from Form 892 they have to be coming from the examiner. Looking at the totals for examiner vs. Applicant it can be seen that 144 of the 147 citations are coming from the examiner. 90 of these are associated with pre-grant applications and 30 are associated with US granted patents. The A, Y, X and XP types are coming from a handful of cases where the Allure Energy documents were used in an International Search Report from WO documents.

Next, let’s look at the companies that are having Allure documents cited against them by an examiner. This can be accomplished by filtering out the three applicant citations (the applicants in these three cases are Nest, General Electric and Rockwell for those of you who are curious) from our collection and looking at the citing assignees count.

Citing CompanyNumber of Citations
Allure Energy37
Aruba Networks4
Assigned to Individual3
Hunt Energy2
LG Electronics2
Millennial Net2
Rockwell Automation2
Solar Chief2

Allure Energy has their own art cited against them most frequently but coming in a close second is Verizon. GE, Sony, Aruba and IBM are also seeing Allure Energy documents being cited against them. Nest, who is the object of the Allure Energy patent infringement suit also makes an appearance on the list.

These lists are easily generated by manipulating a Pivot Table with this data in MS Excel. It is also possible to layer information in a Pivot Table which allows an examination of the top citing assignees and whether it is granted patents or pre-grant applications that are having Allure Energy documents cited against them.

Pivot Table exploring Citing Assignee and Citation Type – Click to enlarge

Verizon has an even split of 18 type 7 (pre-grant applications) and 18 type 0 (granted patents) citations from Allure Energy documents while Sony and IBM only have type 7 citations. GE and Aruba have more pre-grant applications citing Allure but there are some granted patents as well.

The Allure Energy portfolio is only a few years old but it has already generated 110 forward citations, if the self-citations are not counted. The vast majority of these are coming from examiners, as opposed to simply being listed on an IDS that could contain a large number of potentially suspect references. As discussed previously, examiners normally have very good reasons for mentioning a document during the examination process, so there is a high likelihood that the citing patent documents are closely related and pre-dated by the Allure Energy documents. It can also be seen that a reasonable number of large companies are running up against the Allure Energy portfolio during the course of prosecuting their own patents in this space. Finding these companies and identifying the specific documents of interest in this case can be a powerful source of information when looking at potential partners for the Allure Energy technology. Examining forward citations, in general, can be useful when considering out-licensing, but identifying the forward citations coming from the examiners specifically, provides an even higher likelihood of relevance.

Leave a Reply

Your email address will not be published. Required fields are marked *