We’ve all heard of the apocryphal art gallery visitor who confidently asserted, “I don’t know if it’s art, but I know what I like.” When it comes to patent licensing, the prevailing mood seems to be, “I don’t know what a patent troll is, but I know what I hate.” The Congress, several dozen state legislatures, and even the President of the United States are wrestling with the “troll problem”, untroubled that there is no consensus or precise definition of what a troll is or exactly what they do.
Worse than trying to solve the wrong problem is denying that there is one. Yet too many in the patent licensing industry continue to hew to the “what troll problem?” position. This is clearly folly. The world is convinced that something is wrong, and is ready to [over] react. It’s no good pretending that the patent licensing industry can continue as it has for the past two centuries. Bad actors have crept in, and they run a very real risk of damaging the innovation engine of our economy. We should also give up complaining about the label “troll”. It’s marvellously evocative and firmly entrenched. Once we agree to apply this name to the true bad actors, we can concentrate on answering the question, “What is a patent troll?”
In an article in TechCrunch, Leonid Kravets of InterDigital Communications Inc. wrote that what makes a company a troll is not whether it practices the invention itself, or whether they filed the patents themselves, but whether they are “abusing” the system. Quite so. Kravets identifies low quality patents and “enforcement without a reasonable basis for assertion” as the hallmark of an abuser of the patent system – a troll. Yet this falls short of identifying what is probably the most widely agreed upon characteristic of a troll: the assertion of a patent to force a quick, nuisance-value settlement, ideally without the expense of a trial.
Many other elements of trollish behaviour flow from this basic strategy. If you’re only hoping to collect a few thousand dollars from each licensee, you’re going to need a lot of them make serious money. So a troll will place a lot of companies on notice. The corollary is that a troll cannot simply target manufacturers (e.g., of Wi-Fi chips, such as Broadcom, or of wireless access points, such as D-Link) but must descend to the bottom of the value chain, and target the product company’s individual customers, whether small companies or even individuals. For example, Innovatio IP Ventures has made a practice of suing hotels for patent infringement.
Now, if you’re a big scary troll, you need a club to intimidate people. But an extensive collection of cudgels, maces, warhammers, and goupillons actually provides little additional benefit. This is particularly true if you are a cowardly troll who prefers intimidation to combat. And because those weapons are expensive to buy and maintain, you’d actually prefer not to have too many. This brings us to the second hallmark of a troll: they own and assert a very limited number of patents – sometimes as few as one.
These two characteristics probably apply to every troll out there: They seek nuisance-value settlements for a small number of asserted patents. But there are other telltale behaviours and signs that should factor into our evolving definition of a patent troll.
Trolls are unwilling or unable to clearly identify the supposed infringement. This can be attributed to a desire to cast as wide and threatening a net as possible, but is also an unavoidable side effect of sending hundreds of notice letters: you can’t put a lot of effort into any single assertion.
A particularly aggressive species of troll will sue first and hold discussions later. This is attributable to two root causes: a desire to intimidate (a lawsuit is scarier than a demand letter) and a need to cast a wide net (with many targets in the crosshairs, one cannot spend a lot of time talking to each).
In addition, a troll often hides behind a smokescreen of shell companies and obscurely named subsidiaries, partly to conceal the widespread nature of its activities. This is sometimes an unavoidable consequence of the structure of the business deal which lay behind the sale of the patents, and should not be considered irrefutable proof of a troll.
Finally, there’s the notion of patent quality, which is referred to not only in Conversant’s Patent Licensing Principles, but also by Kravets and many others. A troll doesn’t really care about the quality of the patents. Even a low quality patent, certain not to survive re-exam, can be sufficient to frighten the recipient of a demand letter into paying. Poor quality patents are doubtless a problem for the entire patent system, but in my opinion are a red herring in this context. If it were easy to agree that a given patent is over-broad, or even invalid, the troll would obviously be toothless. Patent quality can only be proven by a time-consuming review; we’d like to spot trolls ahead of time. (Also, the review process is only a little less fallible than the original filing process. What were widely believed to be solid patents can founder.) Thus, using patent quality to identify a troll can at best be applied only retroactively, and thus is of limited use.
So here’s our seven-point guide to spotting a patent troll:
1. Seeks nuisance-value settlements
2. Targets individual customers or end-users
3. Owns and asserts a very limited number of patents
4. Unable/unwilling to clearly identify infringement
5. Sues first and holds discussions later
6. Hides behind shell companies and subsidiaries
7. Asserts poor quality patents.

The next time you hear someone accuse a company of being a troll, it may be instructive to see how it measures up against this list. But to return to the point made earlier, it may be that worrying about an exact definition is a distraction. What we really want is ways to make it harder for them to operate.

Leave a Reply

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

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>