Why I Don’t Use Extended Families When Counting Patents

One of the first posts written for the re-launched patinformatics.com site was on the use of One Document Per Invention (ODPI) as opposed to One Document Per Family (ODPF), when conducting a patent analysis project. That post provided an overview of the ODPI method, and explained that with portfolios primarily filed in the United States and Japan, there was a danger of understating the number of inventions contained within a single family. This is especially the case when the concept of an extended, or in most cases, an INPADOC family, is used to reduce a patent collection to ODPF. This post will provide a specific example of where a ODPF reduction would under represent the inventive output of a single family by looking at the claims of the document collection and classifying them by technical concept.

Due to the nature of the patent system worldwide, patents must be applied for in individual countries.  This creates a situation where a single idea might have many individual patents associated with it depending on the number of countries the applicant sought protection in. In some cases, an assignee will seek protection in dozens of countries, and thus there will be many equivalent documents associated with their invention.

In order to simplify some of the dichotomy between inventions and the various patent documents that can be associated with them, the concept of a patent family was created.  There are many ways to define a patent family, but for the purposes of this post we’ll look at the broadest one, the extended, or INPADOC patent family.  Using INPADOC as a means for reducing the number of patent documents into families is one of the most popular methods among patent analysts.

An INPADOC patent family is referred to as an “extended” family since it includes all family members that share at least one priority number.  The specific definition of  an extended INPADOC patent family from the EPO website is:

All the documents directly or indirectly linked via a priority document belong to one patent family.

It is this concept of indirect linking of priority documents which can cause INPADOC families to grow to a very large number of documents, not all of which cover the same invention, per se.  It is not unusual to see a well-developed INPADOC patent family cover several hundred individual documents when you take into account all of the priority documents, applications vs. granted patents, and the different countries where the assignee would have wanted coverage.

The following example is taken from the EPO webpage describing INPADOC patent families:

Example of Single INPADOC Family with Five Documents

Example of Single INPADOC Family with Five Documents – Click to Expand

The idea behind counting patent documents when conducting some patent analysis projects is to ensure that a fair representation of the number of inventions or the amount of work performed in a particular subject or by a specific assignee is adequately accounted for.  For the most part, analysts want to ensure that if the same invention is covered by patent documents in different counties they are not counted individually.  While country coverage can be an important indicator of the value an assignee places on an invention, it can also introduce a bias since organizations with more resources might decide to seek coverage in more countries, while a lone inventor or small company might only seek protection in a single country.  When comparing different methods for solving a technical problem, if you have one approach with companies that seek protection in a large number of countries, it might look like it has more activity than another approach where the companies involved don’t seek patents in as many countries.

So while removing multiple country filing bias is important, the use of the INPADOC or extended family method for doing this can cut too deeply, and discrete inventions can be eliminated while dealing with the multiple country issues.

As an example of this, a portfolio of patent documents covering a wearable fitness monitor was examined. Wearable fitness monitors, such as the Nike+ FuelBand and FitBit® Flex™ have become popular recently, and this area will be used for several future posts as a case technology area to study. In this post, the portfolio associated with the Jawbone Up™ was analyzed with regards to how many technical concepts were covered in the claims.

As of April, 2013, Jawbone, which files patents under the name of Aliphcom, had 80 patent applications associated with the Up™ product, filed around the world, based on data provided by Orbit.com. After removing redundant applications, filed in multiple countries, a collection of nine unique WO applications, and twenty US applications were discovered. All 80 documents were collected together in INPADOC, as a single family. In an analysis, looking at wearable fitness monitors, where the documents had been reduced to ODPF, Jawbone would look to be a very minor player with only a single family, but in fact, they have made an enormous investment in this space.

Many of the 29 documents have identical original titles and similar specifications but they differ fairly significantly when looking at their claims. To determine how many distinct technology concepts are covered by this collection the claims of each document were read and categorized. For simplicity sake, each document was only placed in a single category based on the analysis of the first claim. The chart below shows that twelve individual concepts were discovered within the collection.

Breakdown of Up Portfolio Technology Concepts from the Claims - Click to Expand

Breakdown of Up Portfolio Technology Concepts from the Claims – Click to Expand

Within the categories methods of calculating health, sleep, activity and nutrition, can be found in addition to inventions related to coatings and power management. A stacked bar chart was also used, in this case, to represent the filing date for each application to show when each concept was receiving attention.

Looking at some of the representative claims it can be seen how different they are from one another. A claim on power management, for example might read:

A wearable band comprising: a plurality of sensors;a controller coupled to the plurality of sensors;an energy storage device;a power port configured to receive power and control signals, the power port coupled to the energy storage device;a power manager comprising: a transitory power manager configured to control an application of power to one or more components of the wearable band in one or more power modes; and a power modification manager comprising a power clock controller configured to adapt a clock frequency of a clock signal configured to be applied to the controller as a function of a power mode, the power mode associated with a mode of operation.

While a claim on a protective overmolding might say:

A method, comprising: selectively applying a curable coating substantially over one or more of a plurality of elements coupled to a framework configured to be worn, the plurality of elements including at least a sensor; and selectively forming a molding substantially over a subset of the plurality of elements, the molding configured to provide a protective property.

All and all, the collection represents a particularly diverse set of individual inventions related to the development of a single product, a flexible fitness monitoring device.

While reduction of patent collections to a single INPADOC, or extended family, member is a popular method for eliminating multiple country filing biases it can often lead to situations where distinct inventions are removed, and the breadth and depth of inventive output from a single organization is understated. For this reason, alternate methods for reducing patent collections should be used, especially when comparisons between different areas or organizations will be conducted.

While there are advantages to the OPDI method, described in the previous post on this topic, it must be done manually by the analyst. There are also a few patent family methods provided by individual vendors that come close to achieving similar results:

FAMPAT Family (Questel) – Basic family plus: Applications falling outside the 12 month filing limit; Links between EP and PCT publications; Combining US Provisionals that share the same priority with US Published Applications.

Derwent Family (Thomson Reuters) – Patent Families in the World Patents Index (WPI) draw together patents covering the same invention. Their relationship is defined by the priority or application details claimed by each document.

Looking at the Aliphcom case, there were 24 FAMPAT families and 13 Derwent families associated with the collection as opposed to one INPADOC family.

Any reduction in the number of patent documents being studied will have a significant impact on the statistics that are calculated from them. Some of these reductions, such as removing redundant application publication for patents that have gone on to grant makes sense and can be accomplished fairly easily. Eliminating multiple country filing biases, on the other hand, can get complicated quickly. Traditionally, analysts have used the representative INPADOC family member method to reduce a collection to ODPF but, as discussed, this can dramatically under represent organizations that file primarily in the United States, and who use the US system of Divisionals and Continuations frequently. Looking at the claims of the documents identified from a ODPI approach shows that the diversity of inventive concepts is retained when employing it.

Comments 10

  1. Great post, Tony. I completely agree. I wonder if you can speak to any other factors, besides the US system of continuations and divisionals, which could bring additional inventive concepts into the fold of a single Inpadoc family?

    1. Hello Kristin,

      Thank you! I am happy that you liked the post.

      I think the Intellogist page on patent families does a nice job of explaining some of the examples where families can multiply. Besides the national applications, I think the inclusion of applications that were filed too late to claim priority is the next largest source of additional inventive concepts in the families.

      Overall, there is a tendency for these families to snowball, especially when the filings in other countries start to claim different applications without a consistently used priority document. I have not systematically study the issue though, so I am simply speculating.


  2. Tony,

    Good post and one that is very relevant to all of us analyzing patents for different reasons. Now a question…

    How do you handle Japanese non-families. Sometimes there will be dozens of Japanese publications with no common priority document, but nearly identical specifications? Obviously it’s a non-issue when analyzing the European or US space, but what about when it’s a worldwide analysis?

    1. Hey Dominick,

      Thanks for the comment and the good question.

      For the first pass I would include all of the JP non-families, assuming they don’t share any priority documents.

      For the next step, it would depend on whether I was looking for proof of investment or in-force patent rights as the objective of the analysis. If I am looking at investment then I would likely continue to include all of them after I sorted by application number and removed any redundant documents. The fact that the organization went to the trouble of filing the documents still has some meaning to me as an analyst. If I am looking for patent rights I would look to see how many of the JP documents have been abandoned and at what stages. A significant number of JP documents are place holders and don’t get examined so I would report on that as well.

      I hope that helps and thanks again for the question.


  3. Hi Tony,

    Thanks for the post. This could be a great presentation for the upcoming PIUG Annual Conference!


    1. Hey Suzanne,

      That is a good idea and I will likely put something together for that. I talk about OPDI all the time in my workshops but I guess I haven’t discussed it with a broader audience recently.

      Thanks for the comment and suggestion.

      Enjoy the week,

  4. Interesting read, Tony. I think the Jawbone Up example was a nice choice because it highlights the fact that quite a few initial (priority) applications are for a product and that product may have several distinct inventions and concepts that are needed to make the product. We’ve filed applications where we knew we would get restrictions and end up with several applications. I believe you can see something similar when companies are developing a whole new product platform or a new generation of a current product.

    1. Hello Jim,

      Thanks for the acknowledgement and the comments. I have heard omnibus patent applications referred to as MOAPs, or Mother Of All Patent applications, where an organization will intentionally file a single large application with the express purpose of splitting it up and continuing to shape it as prosecution continues and the market develops.

      Thanks again,

  5. On the Intellectual Property Networking LinkedIn group Alfred Yip made the following comments, that I thought were interesting:

    Thanks for the excellent write-up, Tony. As you have correctly pointed out, divisionals and also continuations in the US already indicates that patents and applications contain different inventive concepts but belong to the same INPADOC family. I completely agree with you that for specific objectives, a manual review of the claims by an analyst is required. Perhaps the analysis could also be merged with the utilization of classifications.

    1. Hello Alfred,

      Thank you for the comment. In my experience, classifications can be a mixed bag, unfortunately. Some of the time they map well to the way in which the analyst needs to categorize the subcomponents, most times though, they don’t really align.

      In addition, the way the classifications are applied by different offices is not consistent. This can also lead to potential difficulties. It is very nice when the classifications work for a project.

      Thanks again,

      In an offline discussion Alfred and I both agreed that we liked the direction that classifications were headed in with the introduction of the CPC.

Leave a Reply to Kristin Whitman Cancel reply

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