Not all patents are created equal: The Wizard of Oz and patent portfolios

In the 1939 film The Wizard of Oz, the title character presents an imposing image of power and authority, designed in part to intimidate. Of course, we know that the great and powerful wizard turns out to be a relatively meek and modest man standing behind a curtain. However, there is a thematic parallel in the US patent system. A number of companies use patent filings to create an imposing appearance, designed to present an image of power and authority, partially to intimidate others.

By various media accounts, their tactics are working, as these companies generate a variety of press releases and headlines touting their relative innovation rates compared to those of their peers. In reality, these rates – as represented in the US patent data set – may be well below the market average. Moreover, like the wizard, these companies benefit economically from creating an appearance of patent depth and breadth, especially as it relates to multi-party licensing and cross-licensing negotiations for patent portfolios involving SEPs, which carry licensing obligations requiring FRAND terms.

FRAND manipulation results in part from the reality that the courts attribute more weight to patent counts (or patent family counts) than inventive value contribution, while simultaneously ignoring important attributes of the underlying patents. This is because the analysis is hard and labour intensive. Yet billions of dollars in value exchange depend on the outcomes of such analyses, the cost of which pales in comparison to the dollars spent on FRAND litigation.

While a detailed analysis of the associated equitable deficiencies in the approach used by the courts and the tactics that FRAND licensees can use to manipulate outcomes is beyond the scope of this discussion, this article highlights several patent portfolio development strategies that companies can use to influence licensing negotiations and to create the appearance of having a larger portfolio, thereby increasing their bargaining power in a FRAND negotiation when calculating balancing payments.

None of the patent analytics tools in the market today illuminate this type of behaviour, as the developers of these tools suffer from the same basic limitation – a lack of specific domain knowledge about the tactics that companies can employ to create the illusion of innovation. Moreover, deep learning and other artificial intelligence (AI) techniques are largely unhelpful for the same reason – the designers lack domain knowledge about the available strategies to play the system. As a result, the AI model endpoints are flawed and the models solve for the wrong outcomes. Thus, deep learning and AI models filled with prejudiced data produce prejudiced results (ie, garbage in, garbage out).

This article explores this topic further and highlights the tactics that companies can use to outsmart the analytics tools that plague the market today, as well as the general patent system, in order to create an economic advantage in licensing negotiations.

Not all patents are equal

It is important to understand that not all patents are created equal. Until designers and users of patent analytics tools accept this basic tenet, existing toolsets will continue to provide low marginal utility. First, it is important to understand that there are a variety of different patent types. The major types include the following:

• Parent patents: parent patents represent a patent that was the subject of a new, independent application, which did not derive from another non-provisional application in the same jurisdiction. A parent patent has an original specification and priority date that does not include non-provisional applications in its priority chain in the jurisdiction. The earliest effective filing date for this type of patent is generally the application date. The priority of the patent derives either from a set of one or more provisional patents or from one or more patents with some form of foreign priority claim. The economic life of a parent patent is generally independent of any other existing patents (except for parent patents whose term is limited by terminal disclaimers).
• Continuation patents: continuation patents derive from applications filed by an applicant that wants to disclose claims to an invention in a co-pending patent application (the parent application). Continuation patents:
  • introduce no new matter;
  • have the same specification as their respective parent applications (apart from the claims);
  • claim the priority of the parent; and
  • generally co-terminate with the patent (ie, expire at the same time as the parent patent, not accounting for patent term extensions or terminal disclaimers).
• Continuation-in-part (CIP) patents: CIP patents derive from applications filed by an applicant that wants additional claims to an invention; however, the applicant may not have disclosed those additional inventions in a co-pending patent application (the parent application). CIP patents:
  • do not necessarily have the same specification as their respective parent application;
  • claim the priority based on the priority of the parent (except for new matter that is not added in the parent application’s specification); and
  • generally co-terminate with the patent.
• Divisional patents: a divisional patent claims priority based on the filing date of the parent application but differs from a continuation or CIP patent in that it claims a distinct or independent invention bifurcated from the parent patent (a patent’s invention is a unitary inventive concept). Divisional patents:
  • have the same specification as their respective parent application (apart from the claims);
  • claim the priority of the parent; and
  • generally co-terminate with the patent.

     

    Table 1. Growth rate

    	Parent	Divisional	Continuation	Continuation-in-part
    Growth rate	(8.82%)	(15.05%)	151.6%	(35.38%)

    Source: Pellegrino & Associates

As a result, the economic life (measured from the grant date to the expiration date) of a continuation patent, a CIP patent or a divisional patent (collectively derivative patents in this article) – excluding patent-term extensions – is generally shorter than that of a parent patent.

Of the four main patent types, most patents granted in the United States are parent parents. Figure 1 captures the composition of patents issued in the United States in 2017.

As Figure 1 indicates, approximately 73% of patents granted by the USPTO in 2017 were parent patents. The next most popular type was the continuation patent, followed by CIP patents and divisional patents. While there are other types of patent (eg, reissue patents and substitution patents), the issuance of these is rare (less than 0.5%). However, these ratios have not always been constant. Analysing USPTO patent grants between 2000 and 2017, the data reveals that the composition of patents has changed over time.

Figure 2 shows that the emphasis on parent patents has dropped over the past two decades; however, the rate of deceleration began to increase in 2010, which corresponds with the beginning of the smartphone wars. Further, since 2000, the number of continuation patents in particular has grown disproportionately compared to the other major patent types, as Table 1 demonstrates.

Between 2000 and 2017, the number of parent patent applications dropped by 8.82%, while divisional and CIP patents experienced even larger losses. However, the popularity of continuation patents surged over the same period. Several motivations help to explain this phenomenon.

First, continuation patents are generally less expensive to prosecute than parent, CIP and divisional patents. The prosecuting agent uses the same specification as the parent asset; therefore, less time is spent drafting the initial application and prosecuting the application through the USPTO.

Second, continuation patents take less time to prosecute; therefore, they are an economical way to bulk up a patent portfolio quickly. Often, the same examiner examines the continuation patent and the parent patent. Thus, the process is more efficient, as the examiner and prosecuting agent may have an established rapport. This plays out in the general duration of prosecution efforts – as demonstrated by the 2017 issued patent pendency data.

Figure 4 reveals that, of the grant data analysed, continuation patent applications generally issue one year sooner (373 days on average) than a parent patent application. Moreover, continuation patent applications publish 329 days sooner on average; thus, the market has notice of the published continuation application almost one year before a parent application. This advance warning can be helpful in establishing the damages period in the case of infringement.

Third, continuation patents are generally easier to obtain. On average, a continuation patent will have:

• 2.66% fewer non-final rejections;
• 26.29% fewer final rejections;
• 26.69% fewer requests for continued examination;
• 0.99% fewer 101 rejections;
• 33.93% fewer 102 rejections;
• 37.07% fewer 103 rejections;
• 37.00% fewer 112 rejections;
• 47.72% fewer double patenting rejections;
• 48.34% fewer drawing objections;
• 28.11% fewer informality objections; and
• 23.48% fewer specification objections.

   

  Table 2. Pendency rates

  Pendency	Patent count	Average days’ pendency	Average days to publication
  Parent	256,763	1,115	484
  Divisional	19,177	748	148
  Continuation	62,785	742	155
  Continuation-in-part	14,322	982	216

  Source: Pellegrino & Associates

There are strategic reasons for maintaining pending continuation, CIP or divisional applications on patent families. For example, patent buyers generally like to acquire portfolios that have assets with pending patent applications in order to further shape and refine the portfolio after acquisition. Moreover, pending patent applications create uncertainty around the final scope of protection for a given patent family. Such uncertainty may serve to deter competitors in the market. Further, open continuations are an effective solution to ameliorating post-grant challenges using the inter partes review process. Few rational actors will file an inter partes review on a patent estate with open prosecution, as the patent owner can create additional filings in response to arguments presented in the review process, neutralising the business reason for filing the review in the first place.

Nevertheless, for companies that do not sell large blocks of patents or that do not find their patents challenged in post-grant proceedings (eg, many of today’s Big Tech companies), the ultimate utility of these strategies is lower than that of smaller counterparts. In summary, without going into the open market to acquire patents en masse (as many companies have over the years), the easiest and most cost-effective way to bulk up a portfolio when developing patents organically in an organisation is to use continuation patents.

Despite the prosecution advantages, continuation patents come with costs. First, the patent term is shorter. The average continuation patent granted in 2017 will have approximately 5,279 days of exclusivity or 14.5 years on average if the patent owner maintains the patent fully through the third maintenance payment. This period differs dramatically from the typical exclusivity period for a parent patent, which in 2017 was approximately 6,246 days or 17.11 years on average. Parent patents thus have on average 2.5 years’ more term than continuation patents. All else the same, a patent with a longer term is worth more than a patent with a shorter term. By extension, patent portfolios that comprise mostly continuation, divisional or CIP patents will have truncated economic lives.

Consider, for example, the case of Digimarc Corporation and competitor Symbol Technologies, LLC. Both companies generated a reasonably large portfolio of US assets under management between 1 January 2000 and 31 December 2017. Digimarc has roughly 1,050 grants; Symbol has roughly 1,350 grants. Figure 3 compares the composition of these grants for each company’s respective portfolio.

As the data indicates, parent patents account for 205 of the 1,037 grants listed in the Digimarc portfolio – 19.76% of the total portfolio size. The balance of the portfolio comprises derivative assets that generally co-terminate with the parent assets in the portfolio. Symbol has a different portfolio composition; parent patents account for 1,102 of the 1,356 grants listed in the portfolio – 81.26% of the total portfolio size. Derivative assets comprise a much smaller portion of the Symbol portfolio. Further, each of the parent patents in the Symbol portfolio has an independent economic life without premature expiration. In contrast, premature expiration has a significant effect on the long-range viability of the Digimarc portfolio, as evidenced by the patent mortality chart in Figure 4.

The chart captures the remaining percentage of each company’s patents after accounting for patent expiration over time. Normalising the data to the percentage of the remaining patents in the portfolio after expiration for each respective company controls for the relative differences in the portfolio size. The chart ultimately indicates that the composition of the portfolio affects its economic life. In fact, the rate of decay between 2018 and the date on which most of the patents expire in 2036 is steeper for Digimarc than Symbol. When looking at the difference in portfolio size between the two companies, Symbol will have around 30% more patents between 2023 and 2032 than Digimarc. All else the same, when a company’s business model depends on the size and longevity of its patent portfolio, it is better to have a portfolio composition like Symbol’s, as opposed to Digimarc’s.

There is an operating metric, known as the integration ratio, that helps to identify the differences between the composition of two patent portfolios. The integration ratio for a patent portfolio is a key indicator of the depth and breadth of a patent portfolio within its respective patent families. The value will range between 0 and 1. The calculation of the integration ratio works by calculating the ratio of parent patent assets in the portfolio to the total patent assets in the portfolio. Integration ratios that are closer to 1 tend to indicate patent portfolios that are horizontally aligned. Horizontally aligned portfolios are generally broader than vertically aligned patent portfolios, which have an integration ratio closer to 0. Vertically aligned patent portfolios tend to cover one or more key concepts in depth, yet those concepts may have little application outside of a narrow focus. For example, a semiconductor manufacturing technology for building touchscreens, which has two parent patent applications and 50 derivative patent applications, may indicate a patent portfolio describing significant expertise in building touchscreens. Yet, a process patent or manufacturing technology for building any semiconductor (eg, a memory chip or microprocessor), which has 40 parent patent applications and 10 derivative patent applications, may indicate a much broader (and possibly more valuable) overall patent portfolio, as the economic footprint of the portfolio is larger. Digimarc’s integration ratio is 0.19, indicating a vertically aligned patent portfolio. Symbol’s integration ratio is 0.81, indicating a horizontally aligned patent portfolio. Naturally, this type of analysis extends into particular art units. A company can have an integration ratio that is more vertically aligned across an entire portfolio, but more horizontally aligned within a particular art unit.

Terminal disclaimers

Another strategy that companies employ to bulk up portfolios is to file multiple patents on the same invention, albeit with small changes. The USPTO may grant a similar patent to the applicant; however, this generally involves the applicant first addressing a double patenting rejection from the USPTO. In general, the USPTO will reject the later application with a statutory double patenting rejection (ie, the later patent application has the same claim as an earlier issued patent) or a non-statutory double patenting rejection (ie, the claims are similar enough to cover the same exact invention disclosed in the earlier patent). The applicant must reword claims to overcome a statutory double patenting rejection. To obviate the non-statutory double patenting rejection, the applicant must file a terminal disclaimer, which is a tacit admission that if the USPTO grants the later patent, it will terminate no later than the date of the earlier patent. This mechanism prevents a patent owner from extending protection beyond the earlier patent’s term. The economic term of the later patent is shorter than that of the earlier patent.

In the context of patent portfolio development, it is a less obvious, albeit common technique – especially among various large, well-known patent producers – to grow the size of the portfolio by filing numerous applications for the same invention with minor claims changes and to address the USPTO’s double patenting rejections with terminal disclaimers. The patent owner receives the benefit of the granted patents; however, the economic life (measured from the grant date to the net expiration date) of a patent with a terminal disclaimer is generally shorter than that of a patent without a terminal disclaimer. A host of other gremlins will emerge when a patent owner attempts to transact in or enforce patents with terminal disclaimers; however, these are outside the scope of this article. Nevertheless, the strategy is reasonably common, as scores of large, recognisable companies have patent portfolios in which at least 33% of the patents have terminal disclaimers. This is a relatively recent phenomenon.

Analysing USPTO patent grants between 2005 (the earliest data available in the data set with embedded terminal disclaimer signals) and 2017, the data indicates that the use of terminal disclaimers for patents – regardless of patent type – has grown dramatically over time.

As Figure 5 indicates, the rate of patents that are subject to double patenting rejections and the use of terminal disclaimers has more than quintupled over the past decade. As was the case with the patent type data, the increase in the use of terminal disclaimers has accelerated significantly and corresponds with the beginning of the smartphone wars.

Moreover, analysing the rate of patents with terminal disclaimers as a percentage of total patents granted, it is clear that the increase in the number of patents with terminal disclaimers does not attribute solely to the growth in the number of patents issued by the USPTO.

Figure 6 shows that patents with terminal disclaimers have grown from roughly 5.5% of all patents granted in 2005 to 13.26% of all patents granted in 2017, representing a total growth of around 141% over 12 years.

Patents with terminal disclaimers also have a host of other assorted nuances to manage. First, as is the case with continuation, CIP and divisional patents, the patent term for a patent with a terminal disclaimer is shorter. The average patent granted in 2017 with a terminal disclaimer will have approximately 5,305 days of exclusivity or 14.5 years on average if the patent owner maintains the patent fully through the third maintenance payment. As discussed above, the exclusivity period for a patent with a terminal disclaimer differs dramatically from the typical exclusivity period for a parent patent, which in 2017 was 6,246 days or 17.11 years on average. Parent patents have roughly 2.5 years’ more term than a patent with a terminal disclaimer. On the grounds that a patent with a longer term is worth more than a patent with a shorter term, patent portfolios containing a significant portion of patents with terminal disclaimers will have truncated economic lives, all else the same.

Consider, for example, the case of Dolby International AB and competitor DTS, Inc. Both companies generated a reasonably large portfolio of US assets under management between 1 January 2000 and 31 December 2017. Dolby has roughly 379 grants, of which 113 have terminal disclaimers – about 29.81% of the total portfolio. DTS has roughly 105 grants, of which six have terminal disclaimers – about 5.7% of the total portfolio.

Dolby has another entity with a larger patent portfolio, which does not rely as much on patents with terminal disclaimers. However, for the purpose of this article, the mortality comparisons are the controlling factor in the analysis; therefore, the focus is on the international entity instead. More than 50% of the patents granted to Dolby International AB in 2017 had an associated terminal disclaimer, while none of those issued to DTS in 2017 had a terminal disclaimer.

Table 3. Breakdown of top 25 US patent producers’ total 2017 grants

Source: Pellegrino & Associates

Patents with terminal disclaimers comprise a much smaller portion of the DTS portfolio. Thus, each of the non-disclaimed patents in the DTS portfolio has an independent economic life without premature expiration. In contrast, premature expiration has a significant effect on the long-range viability of the Dolby portfolio, as is evidenced by the patent mortality chart in Figure 7.

The chart captures the remaining percentage of each company’s patents after accounting for patent expiration over time. Normalising the data to the percentage of the remaining patents in the portfolio after expiration for each company highlights the relative differences in the portfolio sizes. The chart ultimately indicates that terminal disclaimers affect the economic life of the portfolio. On average, DTS will have 22% more patents between 2023 and 2032 than Dolby. When a company’s business model depends on the size and longevity of its patent portfolio, it is better to have a portfolio composition like DTS’s, as opposed to Dolby’s.

Table 4. Number of top 25 US patent producers’ total 2017 grants and parent patents

A company analysis

The data in this article indicates that companies are increasingly using continuation, CIP and divisional patents, as well as patents with terminal disclaimers. Whether a company is using continuation, CIP and divisional patents or patents with terminal disclaimers to bulk up its portfolio and create the illusion of power is fact specific to each company and industry. However, the well-known maxim “you get what you reward” applies. A company that rewards its managers for producing patents, regardless of patent quality, can expect to obtain an array of patents. A company that discriminates with its patent resources and can still produce a significant portfolio without relying on a number of derivative assets or assets with terminal disclaimers may be the mark of a well-run patent operation in a company.

Table 3 contains data for the top 25 US patent producers in 2017.

As the table indicates, IBM is the largest US patent producer and has maintained this position for several decades. However, the raw patent count does not tell the true story of what is happening at IBM. Despite increasing patent counts year over year, IBM’s original innovation rate in the patent system has been decreasing for years. This is understandable, as IBM divested its PC business in 2004 and its semiconductor manufacturing division and x86 server business in 2014. These divisions were large patent engines for the company. In 2017 37% of IBM’s patent grants had an associated terminal disclaimer to obviate a double patenting rejection and only 56% of IBM’s patents were parent assets. Coupled with the fact that IBM abandoned 47% of its patents between 1 January 2000 and 31 December 2017, the data indicates that IBM’s current patenting strategy focuses more on raw patent counts than original innovation.

Other companies that have a large number of patents with terminal disclaimers and that rely largely on continuation, CIP and divisional patents include Google, Huawei, LG, Microsoft and Sony. The largest standout in the crowd is Huawei. While it has put up impressive patent counts in the past several years, only 19% of Huawei’s patents are parent assets. In 2017 the company authored only 289 original specifications, despite generating five times as many patents in the same period.

In contrast, companies that stand out for scarcely relying on derivative assets or patents with terminal disclaimers include Canon, Toyota, GE, Fujitsu, Panasonic, BOE Technology Group, Hyundai, Robert Bosch and General Motors.

Reordering these companies according to the number of parent patents, which is the closest analogue to standalone or independent inventions (although parent patents can have terminal disclaimers to other patents), we find that the order of the top 25 innovators in the United States changes.

As Table 4 indicates, Samsung dethrones IBM for the top spot. While there is a general reshuffling among many of the companies (eg, Qualcomm rises to fifth position, Microsoft drops to eleventh and Google drops to ninth), Huawei falls to the bottom of the list. In fact, expanding beyond the top 25, Huawei’s position in the rankings drops all the way to 99th place.

The astute reader may notice that the strategies outlined in this article are reasonably sector specific. Companies that participate in the development of wireless communications equipment (including handsets) or computer hardware or software where FRAND licensing requirements exist tend to rely more on derivative assets and assets with terminal disclaimers than companies in other scientific areas. For example, none of the automobile companies outlined in the 2017 top 25 US patent producers tend to use patenting strategies that rely on derivative assets and assets with terminal disclaimers.

FRAND effect

So why does this matter? Many companies produce smartphones, computers and other electronic products that employ standards with FRAND licensing requirements. As discussed above, some companies benefit economically from creating the appearance of patent depth and breadth because the courts attribute more weight to patent counts (or patent family counts) than inventive value contribution, while simultaneously ignoring important attributes of the underlying patents.

A detailed analysis of the associated equitable deficiencies in the approach used by the courts and the tactics that FRAND licensees can use to manipulate outcomes is beyond the scope of this article; however, the practical effects of such activities can alter royalty payments and balancing payments worth millions of dollars per year. A smartphone company that generates $20 billion in revenue in one year from the sale of phones can manipulate the general allocation of royalty balancing payments with manipulations to its patent portfolio in order to enlarge the appearance of the portfolio, but not the inventive scope.

The instrument of manipulation is the general formula used to calculate a royalty rate to a licensor based on the proportional share of patents that the licensor provides to standard. That formula is outlined in Figure 8.

Unfortunately, there is no standard, industry-accepted methodology for calculating the contribution to the standard. Some practitioners prefer to use disclosures or contributions to a technical standard as a measure of calculating contribution to the standard. Others prefer to use various forms of patent counting (eg, patent family counting or raw patent counting). A fundamental tenet of the common calculation methods employed today is that each patent carries the same weight, which – as this article demonstrates – is likely a faulty assumption. Not all patents are created equal. A second fundamental tenet of the common calculation methods employed today is that all contributions to a technical standard carry the same technical and economic significance to the technical standard, even though many elements of technical standards are boundary conditions rarely utilised. These two basic tenets provide three means through which parties can manipulate the calculation of a royalty rate determination.

First, in a licensing negotiation between a company and a given counterparty, the company can overcount its patents, driving its numerator higher in the formula where it receives royalty payments, perhaps using several of the tactics outlined in this article. This increases the payments that the counterparty must make to the company. Second, that same company can then undercount the patents of the counterparty, driving the counterparty’s numerator lower in its royalty calculation, thereby reducing the payments that the company must make to the counterparty. The combination of these two tactics reduces the net balancing payments that the company may have to pay to the counterparty or increase the net balancing payments that the counterparty must pay to the company.

Third, and more subtly, a company can propose a variety of features for integration into a technical standard that addresses boundary conditions in the daily operation of a device. This increases the total number of patents in the standard and the total number of patents attributable to the contributing company, even though the market value of the patents addressing boundary conditions may be of little significance to the overall standard. The effect of this tactic is to dilute the economic impact of the major elements of the technical standard and the subsequent payments to the contributors of those elements. If a company creates enough noise around the standard-setting process, it can dilute the value of the key innovations in a technical standard.

Conclusion

The focus of this article was to illuminate several tactics that companies can use to play the analytics tools on the market today and the general patent system to their economic advantage in licensing negotiations. The examples illustrate how companies can quickly and organically grow their portfolios using continuation, CIP and divisional patents several years faster than companies that focus on filing original patent applications. Moreover, companies can file multiple patents on the same invention with small changes and receive issued patents from the USPTO, subject to terminal disclaimers. Taking into account general growth in the US patent system, the empirical evidence demonstrates that these tactics are a relatively recent development from the past decade.

No matter which strategy a company employs to bulk up its portfolio, the bulking-up activity truncates the economic life of the portfolio by 2.5 years on average and the transacting and enforcement of patents within the portfolio carries greater nuance and risk. It behooves licensing professionals to consider the types of patent that they are transacting in and to use those patent types to help set value expectations. Licensing professionals should be aware that the types of patent that a company owns can influence calculations associated with FRAND licensing and the associated net balancing payments due under a cross-licence.