Policy makers must realise not all licensing-based businesses are the same

The moniker ‘patent troll’ is increasingly widely used – often as a scare term. However, while the label describes a variety of patent-holding companies, it fails to differentiate between those engaged in abusive behaviour and those behaving perfectly reasonably

Media reports are full of stories about patent trolls. According to conventional wisdom, patent trolls have a nasty habit of sending letters to big companies, small companies and even retailers, threatening lawsuits if an exorbitant licensing fee is not paid for the use of some everyday technology. We also hear that these lawsuits are based on a bunch of patents, most of them probably invalid, which the patent troll bought from some poor failing company. Rather than seeking to license the technology right away, patent trolls are said to wait until companies are already making and selling products that incorporate the technology, making it difficult to turn down their licensing demand. Of course, if a patent troll does file a lawsuit, we are told that it will likely accept a quick settlement to avoid having the patent’s validity tested. President Obama bluntly described patent trolls as companies that “hijack somebody else’s idea” to “extort some money”.

If you accept the mainstream story, this bad behaviour by patent trolls is a serious problem. Academic work by Professor Colleen Chien estimates that over 60% of lawsuits in 2012 were brought by patent trolls; Professors Jim Bessen and Mike Meurer have claimed that defending against patent troll lawsuits cost companies over $29 billion in 2011. However, there is a serious problem with these accounts, and one that is not being reported in the media. The patent trolls that Chien, Bessen and Meurer and other scholars include in their research are not based on a company’s behaviour. Rather, the research takes a very broad definition of ‘patent troll’ which includes any entity that can be considered non-practising, such as a university, an individual inventor who does not manufacture a patented innovation, a start-up forced to sue a larger company which stole its patented innovation before it had a chance to get it off the ground, large licensing companies such as IBM and even manufacturing companies that license patented innovation outside their primary areas of business.

Defining the problem

There is no standard definition of ‘patent troll’ and the term remains a fast-moving target. The Federal Trade Commission has defined ‘patent trolls’ as “firms whose business model primarily focuses on purchasing and asserting patents”. Chien, now a senior adviser for intellectual property and innovation in the White House Office of Science and Technology Policy, has described patent trolls (or ‘patent assertion entities’ (PAEs), a term she coined) as “businesses that assert patents as their primary business model” – she excludes universities and start-ups. However, Bessen and Meurer do include universities and start-ups in their definition of a ‘patent troll’ or what they call a ‘non-practising entity’ (NPE).

This is not merely a semantic debate. The definition makes a significant difference to how big the actual problem is. Consider the difference between the number of lawsuits that Chien attributes to patent trolls and the findings of the Government Accountability Office (GAO).

Chien’s figure is based on lawsuits in 2012. The GAO’s number (19%) is based on lawsuits from 2007 to 2011; its data found a low of 17% of all lawsuits filed in 2007 and a high of 24% filed in 2011, but that the increase was not statistically significant. In any case, it is significantly lower than Chien’s. The GAO notes that the discrepancy between its data and other higher figures is due to definitional differences.

Regardless of whether it is a big problem, as academics claim, or a smaller problem, as the GAO data supports, the definition of ‘patent troll’ includes no reference to bad behaviour.

The deluge of unfounded demand letters and nuisance lawsuits of a very few prototypical patent trolls has driven Congress and the courts to try to reform patent law to hinder these activities. While there are excellent reasons for wanting to deter the worst of them, the reforms – like the definition of ‘patent troll’ – are not being tailored to affect only those companies that are behaving badly; instead, the proposed legislation and judicial decisions are aimed at a category of litigant, NPEs, rather than the behaviour to be stopped. This is happening as increasing numbers of companies find themselves categorised as NPEs, PAEs, patent monetisation entities (PMEs) or just plain patent trolls – even if they are making or selling something or providing a service, and are not engaging in any abusive litigation or sending demanding letters. So while the number of bad actors is small, the number of companies licensing their technology is increasing, meaning that academics can claim that the patent troll problem is growing out of hand. It is getting to the point that having any sort of patent licensing programme can put a company at risk of being labelled a patent troll. It certainly makes it more likely that any reform efforts may have unintended consequences.

In addition to realising that the patent troll problem may not be as large as previously claimed and reported by the media, it is also important to start thinking about patent trolls based on specific and articulated bad behaviour, rather than ill-defined intuitions about litigation models or other status-based labels. New research in legal scholarship is seeking to disrupt these definitions of ‘patent trolls’ and instead look at the companies behind the label. In the first of a series of planned academic articles, I look at one particular type of company that is commonly labelled a ‘patent troll’ – what I call the ‘formerly manufacturing entity’. These are companies that used to make or sell something, or provide a service, but for any number of reasons have shifted towards a licence-based business model. Some examples of formerly manufacturing entities include Conversant (previously MOSAID), which has moved to a strictly development and licensing model; and General Electric, which still makes numerous products, but also obtains a significant revenue stream by licensing its extensive patent portfolio. Both of these companies, as well as others that I discuss below, have been unfairly labelled patent trolls by the media, as well as by most academic research.

Figure 1. Compare and contrast: percentage of US patent cases brought by NPEs/PAEs, according to two studies


Focusing on behaviour

To begin defining ‘patent trolls’ based on their bad behaviour, rather than their licensing-based business model, it is helpful to look at what exactly is problematic about the actions of patent trolls. It is claimed that patent trolls hinder invention and innovation – the development and improvement of new ideas into useful products – as well as commercialisation, the process of bringing these useful products to the marketplace. The story goes that instead of engaging in productive behaviour, patent trolls unjustly seek royalty payments through abusive litigation tactics for patents on technologies they often did not create, but instead purchased. Their business model is allegedly based on ambulance-chasing-style legal action, rather than on productivity or creativity.

Patent trolls are regularly condemned for their licensing and litigation behaviour. The patents that they acquire are alleged to be of questionable validity and to have broad claims that cover existing technology. They are then said to use these broad and weak patents to threaten litigation and seek licences from companies making use of the technology. In addition, they are accused of preying on small companies and start-ups that cannot afford to battle them in the courtroom. Perhaps most egregiously – and the behaviour that earned them the colourful moniker – patent trolls are said to ‘lurk under the bridge’, delaying licensing negotiations or the enforcement of their patents and hoping to trap a company after it has already sunk time and money into developing and marketing products or services. Such delays increase the value of the technology to the alleged infringer, affording the patent troll a better negotiation position for a larger licence fee or settlement. Additionally, with respect to litigation, it is often said that, unlike other companies, patent trolls are not deterred by the possibility of a reciprocal lawsuit (ie, they have no fear that their products will be enjoined based on a defendant’s countersuit). At worst, their patent will be found invalid.

Many of the companies unfairly labelled as patent trolls also facilitate innovation and commercialisation by market-making in innovative technology

These behaviours are alleged to amount to a shakedown or a tax on doing business, which results in less money being invested in R&D or in the processes necessary to bring new products to market. Of course, companies that make or sell products also engage in licensing and litigating their IP portfolios, but they – allegedly – do not employ the same abusive tactics.

On the flipside, there are many reasons to believe that the vast majority of NPEs, PAEs, PMEs and other companies that have been labelled as ‘patent trolls’ actually facilitate commercialisation (and, indirectly, innovation) by performing important match-making and market-making functions. While it is easy for big companies to find each other and make deals – for instance, the 2012 Microsoft purchase of patents from AOL, followed by a subsequent sale of a portion of that portfolio to Facebook – it is not as easy if one or more of the companies is small or the patent is owned by the inventor. Further, if an individual or very small firm seeks to license its patents or other intellectual property, there is no credible threat of litigation to back up the licensing request because the individual or small firm will generally lack the resources to commence and sustain a lawsuit against a bigger company. In other cases, the individual or small firm may be uninterested or unable to commercialise the invention. In these situations, it is beneficial to have a company that can step in to match-make or connect inventors which are unable or do not wish to commercialise their invention with willing and able commercialisers. These companies, acting as intermediaries, are in a better position because they offer a credible threat of litigation. The patented technology ends up in the hands of companies which will bring it to market, thus promoting commercialisation; and the inventor or small firm is rewarded with resources it otherwise would not have obtained, allowing for further invention and innovation.

Many of the companies unfairly labelled as patent trolls also facilitate innovation and commercialisation by market-making in innovative technology. Specifically, these companies serve as a centralised market where patents can be bought when the seller is ready and sold when the buyer is ready. This adds liquidity to IP assets, making them more valuable and enabling companies to get products into the hands of consumers more quickly and efficiently. Further, in their capacity as practised intermediaries, these companies have experience and specialisation in assessing patent value. Patent valuation is a difficult endeavour and individual inventors and small companies can rarely obtain sufficient knowledge or practice to be good evaluators, leaving them to undervalue or overvalue their patents when seeking to license them on their own.

Formerly manufacturing entities

Congress, the courts, academics and the media rarely differentiate between the myriad of NPEs, PAEs, PMEs and other companies that they lump together as patent trolls. As long as a company is not practising its patented technology – in other words, is not making, selling or doing something covered by its patents – it can be called a harmful patent troll. The only occasional exception has been for universities, which generally do not produce products or provide services. A few empirical studies have tried to classify potential patent trolls into various classes for counting purposes. I suggest that research needs to go further and actually look at the behaviour of the many different types of company that are commonly labelled as ‘patent trolls’, so that Congress and the courts can more carefully tailor their reform efforts.

One example worth looking at is formerly manufacturing entities. An examination of a few of these reveals little in the way of the harmful behaviour that Congress and the courts are trying to deter.

Despite stories to the contrary, many so-called ‘patent trolls’ were not formed as licensing firms. Rather, they began life as companies based around technologies that they hoped to manufacture and sell. Some were simply unsuccessful at commercialising their inventions – even Thomas Edison licensed his inventions to others who were better equipped to manufacture and sell the products embodying his technology. For a more modern example, another often-named patent troll, NTP, turned to licensing only after repeated failed attempts at manufacturing and selling its own products. Other companies may have succeeded, but changed to a licensing business model for any number of reasons, including foreign competition, disaggregation of the supply chain, changes in technological standards or preferences, or extraordinary amounts of unchecked infringement by others in the market.

Conversant is a prototypical formerly manufacturing entity, which began life as a semiconductor technology company in the 1970s. Its engineers invented and it then manufactured the circuit technology used in dynamic random access memory chips, as well as equipment for debugging prototype chips. In the mid-1990s the company realised that others were using its patented technology and thus it began a licensing programme in addition to developing and manufacturing. In 2006 it restructured to focus on IP licensing. Since then, it has obtained additional patents – some through purchase and some through in-house development. Conversant is often called a patent troll based on its extensive IP portfolio and tremendously successful licensing programme.

Soverain Software is another example of a formerly manufacturing entity; while it still offers a product, the company has increased its focus on licensing its intellectual property. The company began in 1993 as Open Market, at the point that the Internet as we know it came into existence and began to reflect capabilities for uses beyond those of university research and defence. The next year, Open Market developed some of the initial e-commerce technology and filed for patents on those inventions – these issued in 1998. From 1996 to 2000 Open Market’s product, Transact, was a leader in the new and burgeoning e-commerce field. As the 2000s approached, larger companies entered the field, while Open Market was restructured and then eventually acquired by Soverain Software in 2003. A number of original inventors remain at the company and it continues to offer and support the Transact product. Soverain Software has also been repeatedly criticised for being a patent troll due to its efforts to enforce the patents it acquired, some of which courts have subsequently found invalid.

These vignettes are simply two examples; many other companies have histories similar to Conversant or Soverain Software. Common threads run through these stories. In particular, formerly manufacturing entities tend to share three common characteristics that distinguish them from the rhetorical trolls that have captured the attention of media, Congress and the courts. First, they had or still have a product that is available on the market – Conversant invented, made and sold actual chips; Soverain’s predecessor invented Transact and the product is still available. Second, these companies operated in the business world and were subject to the business ethos that goes with being part of an industry – Conversant was, and still is, part of the chip industry as it is still developing technology in the space; Soverain Software, although it did not develop the technology, has retained a number of the employees who were part of the original development team for the technology. Third, these companies continue to develop and/or support their technology.

Formerly manufacturing entities in general are unlikely to engage in the types of abusive behaviour that are routinely attributed to patent trolls

These stories also illustrate how formerly manufacturing entities in general are unlikely to engage in the types of abusive behaviour that are routinely attributed to patent trolls. First, formerly manufacturing entities do not delay knowledge of a patented technology in the hope of sneaking up and trapping unwary companies. Conversant was manufacturing its chips; it approached infringers for licences after the company itself had already made the technology known. Similarly, Soverain Software obtained the patents it had been seeking to license after many companies had already become users of the Transact product; not only was the technology known, but it was heralded in the media as the online world was opening for business.

Second, because formerly manufacturing entities have in the past brought products to market, this experience gives them important knowledge and market specialisation skill sets. Thus, they are in a better position to evaluate the technology they are licensing, as well as the costs and risks associated with commercialising it.

Third, although formerly manufacturing entities may not fear a reciprocal patent infringement lawsuit, they do face other risks and deterrents. These risks are multifaceted both in the law and in commerce. One legal risk is that their patent(s) can be invalidated either in court or through one of the many post-issuance review programmes at the US Patent and Trademark Office. There are also other legal tools that can be used against them, such as antitrust complaints. Finally, they face an important social norm deterrent to bringing frivolous lawsuits – business reputation. When the business model of the company shifts, undoubtedly some of its culture and ethos remains. In the case of Soverain Software, employees of the company when it was functioning under a manufacturing regime remained even after the switch to a licensing business model. Conversant has gone so far as to publish a document that describes its commitment to ethical licensing behaviour.

Fourth, formerly manufacturing entities are unlikely to engage in delay. In general, these companies were trying to commercialise the invention themselves at the time the patent was being prosecuted. At that point in time, there is little to be gained by delay, as the property rights afforded to the patentee are available only after the patent is granted. Formerly manufacturing entities are also unlikely to delay in negotiating licences or bringing a lawsuit, especially since their revenue stream is linked to success in obtaining licensing fees or court-issued awards of reasonable royalties. In fact, many of these companies, including Conversant, attempted to license their technology while still operating as a manufacturing company. However, because the technology was already known, there was little likelihood of catching an infringer unaware.

In addition to not engaging in abusive behaviour, formerly manufacturing entities provide unique benefits in the match-making and market-making areas. In particular, because they either did or attempted to commercialise the technology themselves, they are in a better position to know its true value. Patent trolls are often condemned for seeking a quick settlement of cases by seeking low-ball royalties or damages that are not worth the trouble of fighting in court. However, a knowledgeable formerly manufacturing entity could be seeking a low figure precisely because it knows that that is what the technology is worth.

Behaviour, not business model

The constantly shifting definition of ‘patent troll’ – as this term is used by the media, academics, judges, members of Congress and even the president – includes many different types of NPEs, PAEs, PMEs and other companies. It is thus increasingly important to look at what sorts of behaviour are driving the concerns, rather than whether the company fits within an overbroad, sensationalist category. If the concern is t0 stop deceptive demand letters and classic ambulance-chasing-style litigation tactics, then revisions to the patent system must be tailored to address these specific issues. Otherwise, ongoing revisions will likely harm all sorts of companies in the innovation industries – including formerly manufacturing entities and others – that are not engaging in bad behaviour and that instead promote the discovery and commercialisation of patented innovation.

Action plan

‘Patent troll’ is such a commonplace term in current IP-related parlance. However, the actual meaning of the term is not well defined – and those that use it should recognise that:

  • the label ‘patent troll’ covers all sorts of companies that are seeking to license their IP portfolios, but are not engaging in any abusive behaviour; and
  • if patent reform is not narrowly tailored towards specific bad behaviours by actual bad actors, it will likely do more harm than good.

Kristen Osenga is a professor of law at the University of Richmond School of Law, Virginia

This piece is based on an academic article forthcoming in the University of Connecticut Law Review, for which research support was made possible by a Thomas Edison Innovation Fellowship grant from the George Mason University School of Law. The author is not affiliated with any of the companies discussed in the article

Unlock unlimited access to all IAM content