Accounting Denied

Rovi Guides Inc v Videotron Ltd 2022 FC 874 Lafrenière J

2,337,061 / 2,339,629 / 2,730,344 / 2,336,870

In this decision Lafrenière J held all the asserted claims of Rovi’s patents to be invalid as being anticipated and/or obvious [7] for reasons that turned entirely on the facts. Some of the claims would have been infringed had they been valid. The most interesting aspects of the decision related to remedies, and in particular the question of whether Rovi was entitled to an accounting and his assessment of a reasonable royalty. Lafrenière J’s analysis of these issues strikes me as being entirely sound and a very helpful development of the law.

The patented inventions relate generally to “interactive television program guides” [IPG]—the familiar interactive menus used to select current and upcoming programs. Television listings were originally provided in paper, then in a non-interactive linear scrolling on screen guide [9]–[16]. Interactive guides, as the name suggests, allow the user to scroll through the menu in the way that is familiar to all of us today. Rovi’s patents, which date from the late 1990s, were directed to high-level concepts. For example the 061 claims “focus on a user using a device located outside the home, such as a laptop or smartphone, with a limited IPG to schedule a recording on a device inside the user’s home” [303]. “According to Rovi, there is no pay-TV provider ‘anywhere in the world’ with an IPG who does not need a license from Rovi to operate” [552].

The defendant Videotron is a cable TV service provided operating primarily in Quebec [25]. Videotron initially took a licence from Rovi, effective April 2010 [564], primarily to avoid litigation risk [369], but declined to renew the licence, which expired in 2016, considering that there was little value in the increasingly obsolete portfolio [577], [579].

Most of this long decision is devoted to the details of the prior art and the asserted technology, but in broad terms, it was perfectly clear to everyone in the field that interactive program guides were the future, and development was limited primarily by limitations on bandwidth and hardware [154]. The 1990s were “a period of great innovation in the field of interactive digital audio-visual applications and services” [316] and the kind of conceptual innovations reflected in Rovi’s patents were natural developments as the field responded to consumer demand and improvements in bandwidth and other hardware; consequently, the claimed invention had already been proposed by others, or were obvious extensions of existing technologies. That’s all I’ll say about anticipation and obviousness, which were the main substantive issues.

Overbreadth

Before turning to remedies, I’ll note that Videotron tried yet another variant of the overbreadth argument which was really an (unsuccessful) added matter argument. Videotron produced the file history of the 870 patent which showed that hundreds of claims had been added. Videotron argued that “the inventors were not consulted about the amendments and that the 870C Claims are not described in the disclosure,” and “[a]ccording to Videotron, this raises an issue whether the claims are broader than the invention made or disclosed” [364]. Lafrenière J held that even if the inventors had nothing to do with the application after it was filed “a claim is not invalid simply because it was added without the involvement of the inventor after a patent application is filed” [369]. Even though s 38.2 was not mentioned in the decision, this was really an added matter argument, and Lafrenière J’s conclusion is consistent with O'Reilly J’s holding in Western Oilfield 2019 FC 1606 [236] that “the factual circumstances surrounding the amendments were not relevant to this issue. It is a matter of logic, not fact, whether the amendments are reasonably inferable.” Perhaps Videotron characterized the issue as being one of overbreadth rather than added matter in an attempt to avoid this rule. In any event, this represents another failed attempt at giving independent effect to overbreadth as a ground of invalidity.

Remedies

The most interesting part of the case is Lafrenière J’s discussion of remedies; while the issue was moot, given his holding on the merits, he discussed it for the sake of completeness [544]. The first issue was whether Rovi should be permitted to elect an accounting, or should rather be restricted to damages, in this case a reasonable royalty. While it is very well established that an accounting is a discretionary remedy, there isn’t a lot of jurisprudence on when an accounting should be granted in the patent context that goes beyond the accepted general rules of equity, such as laches or bad faith. It is usually desirable that the law develop incrementally, and it is convenient that Lafrenière J’s holding on the merits allowed him to discuss the issue in some detail, in obiter, in the context of a tricky case.

By way of background, the product at issue is a ‘complex product’ in the sense that the patented technology contributes only a relatively small part of the value of the product embodying the technology. This is opposed to a product, like pharmaceuticals, which are not simple in a technical sense, but rather in the sense that the patented technology accounts for a large part of the value of the product in which it is embodied. Complex products raise some special challenges in terms of remedies, as discussed in Biddle et al (eds), Patent Remedies and Complex Products: Towards a Global Consensus (Cambridge University Press, 2019) (open access).

Rovi provides products, such as the supply of IPG software and program data [549], and Rovi and its predecessors made very significant investments and contribution to the relevant technology over the years [21], [23]. Rovi also provides licensing services, “allowing customers to license patents that they own” [549]. This evidently includes licensing by customers, such as Videotron, who do not use Rovi’s products. Rovi significantly grew its patent licensing business from 2003 to 2013 [549]. Rovi’s business model is to licence its entire portfolio at a rate that does not vary based on the number of patents that Rovi has or the licensee practices. [550]

Rovi and Videotron initially entered into a 5 year licence agreement in 2010 [564]. Videotron’s primary motivation for entering into this agreement was reduction of litigation risk—“Videotron did not want patent issues, including the threat of an injunction, to distract the company from completing its important project” [567]. Videotron did not consider the market value of different features covered by the patents in the Rovi portfolio: “freedom from suit was the primary motivating factor that led Videotron to accept to pay the royalty rates that were ultimately negotiated, and not the value of any particular patents in Rovi’s portfolio” [568].

When Rovi and Videotron entered into renewal negotiations prior to the expiry of the licence agreement, Rovi wanted to nearly double the royalty rate [572]. Videotron asked for a one-year extension to allow it to assess the value of the licensed technology [572]. Videotron repeatedly asked Rovi “to identify those patents Rovi considered to have particular value to Videotron and that were specific to its platform,” but never received a satisfactory answer [575]. Videotron therefore “declined to enter into another long-term licence for what appeared to be aging patents taken from an increasingly obsolete IPG patent portfolio” [576]. In the end, Videotron was licensed through 2016.

It was against this background that Lafrenière J addressed the question of whether Rovi should be allowed to elect an accounting. There’s a bit of schizophrenia in the law relating to the patentee’s election of an accounting. It is perfectly clear that a patentee does not have a right to an accounting for profits [580] and it is true that in principle “the patentee bears the burden of proving their entitlement to an accounting of profits” [581]. But on the other hand, as Lafrenière J noted, “the Court should not decline to exercise its discretion to award an accounting of profits to a party in the absence of any compelling reasons of doing so” [580]. For this point Lafrenière J cited Philip Morris 2016 FCA 55 ¶ 8, a trademark case, which in turn cited Apotex v Bristol-Myers Squibb 2003 FCA 263, a patent case, which noted that although equitable remedies, such as an accounting, cannot be elected as of right “a discretionary remedy is not an arbitrary remedy. In the absence of proof of a bar to equitable relief, a claimant can expect to be granted the remedy it seeks in accordance with the principles governing its availability.” In practice, in a standard patent case, such as an innovator pharmaceutical company bringing an action against a generic, the successful patentee will normally be allowed to elect an accounting, subject to any traditional equitable bar (though even then there are a few cases idiosyncratically denying an accounting). In non-standard cases, the equitable analysis comes more to the fore. There is no doubt the traditional equitable considerations of laches or bad faith are applicable, but these have traditionally had a fairly high bar. But equity does allow for flexibility both in the application of traditional criteria and in developing new criteria.

In this case, Lafrenière J applied a four-factor analysis, drawn from Apotex v Bristol-Myers Squibb 2003 FCA 263 [15] (reformatted for clarity):

[581] These factors include:

(i) whether there has been undue delay in commencing or prosecuting the litigation;

(ii) the patentee’s conduct;

(iii) the infringer’s conduct;

(iv) whether the patentee practiced the invention of the patent in Canada; and

(v) complexity of calculating an accounting of profits.

The first three factors are related to the traditional grounds of laches or bad faith, though perhaps more broadly interpreted, while the last two are more specific to patents and the remedy of an accounting. I should note that neither the FCA in Apotex v Bristol-Myers Squibb nor Lafrenière J in this case suggested that these considerations are exhaustive.

(a) undue delay

There was no suggestion of undue delay by Rovi in commencing litigation [583]–[584], though there was delay by Rovi in prosecuting its patents, which Lafrenière J discussed under the next heading.

(b) Rovi’s conduct

Rovi’s conduct was an important factor in Lafrenière J’s decision to refuse to allow an election of an accounting:

[587] the evidence before me establishes that Rovi has a reputation of using hard-ball legal tactics to pressure third parties to license its patent portfolio. Rovi was known for its business model of aggregating patent portfolios, seeking licences, and relying on its prior licences and aggressive use of litigation to drive risk-avoiding businesses into deals that are consistent with its schedule of royalty fees. It would spend in the tens of millions of dollars annually on patent litigation.

Playing “hard-ball” is not close to the kind of bad faith conduct by the plaintiff that is more traditionally invoked in equity; this is perhaps an illustration of interpreting a traditional equitable ground broadly in the context of an election of an accounting.

This is not a criticism. The broader point here is that if the parties negotiating this kind of licence know that a court would grant an accounting rather than a reasonable royalty, this gives additional leverage to the patentee in the negotiations. This makes it more likely that the royalty will reflect the value of avoiding litigation, rather than the value of the technology itself. This is undesirable. The point of patents is to provide an incentive to create valuable new inventions for the benefit of society, by providing a reward commensurate with the value of that technology. This gives the inventor high-powered incentives to develop technology that consumers actually want—the better the invention, the greater the reward. Giving the patentee a reward that reflects the value of avoided litigation costs, rather than the value of the underlying technology, is therefore contrary to the purpose of patent law. So, as Professor Taylor put it, “reasonable royalties should reflect the value of patented technology rather than patent rights” (49 Ga Law Rev 79, 89, original emphasis); see also Ch 4 of Complex Products. The principle applies whether the reward is reflected in a legal remedy, such as a reasonable royalty, or is reflected in a royalty, which, after all, is negotiated ‘in the shadow of the law.’

A second factor relied on by Lafrenière J “is Rovi’s apparently deliberate strategy of delaying the prosecution of its patents,” which were issued 13–17 years after filing. This kind of submarine patenting is problematic because it means that potential licensees may be locked in to the technology when negotiations start, with the result that the patentee can leverage the threat of an injunction to extract sunk costs, a problem known as “holdup”. In this case, “Rovi’s own expert witness confirmed that patent holdup was a problem because once a potential infringer has launched a product it reduces flexibility and an opportunistic patent holder can then try to extract a larger, unreasonable licencing fee” [591]. Holdup is certainly a problem when it occurs, and the real debate is as to whether holdup is a pervasive problem, or rare and idiosyncratic. Given the evidence before him, it was certainly appropriate for Lafrenière J to take this into account.

Lafrenière J noted “When it came time to negotiate terms of renewal of the licence, Videotron repeatedly sought specific information about the value of the Rovi patents. Its attempts to substantiate and validate the terms for which Rovi wanted a multi-year contract . . . were rebuffed. It is quite apparent to me that the reason why Rovi declined to reveal to Videotron a complete list of specific patent claims it considered infringed was to prevent Videotron from designing around them. Rovi took the position that even after all four patents in suit expire, Videotron would not be free to continue carrying on its current activities as there were always other patents that Rovi would be able to assert against Videotron” [592]–[593]. This illustrates both the holdup problem and that Rovi wanted to avoid negotiating over the value of the technology itself.

In the end, Lafrenière J held that “Rovi’s questionable business practices cannot help but serve to colour my view of the value of the features that it claims in the Patents. This factor weighs heavily against Rovi.” This conclusion strikes me as entirely sound.

(c) Videotron’s conduct

Rovi argued that Videotron was a willful infringer, and this should weigh in favour of an accounting [596], relying Monsanto v Rivett 2009 FC 317 where Zinn J gave the following reasons for granting an accounting:

[23] there is no deterrent from infringing the patent if what the infringer is required to hand over is the sum he would otherwise have paid to Monsanto to buy the seed and the licence. In fact, this would almost be counter to the purpose of deterrence. It is much like saying, as the plaintiffs put it in their oral submission, “Catch me if you can”. If caught, the defendant would be required to pay the sum he would have paid to use the patent in any event. When not caught, he is left with a windfall. The accounting remedy would lack any deterrent effect if defendants could use patented technology and retain the profits from such use subject only to paying a license fee as compensation if and when they are caught.

This is an important point, but it is not complete. First, there are other disincentives to infringement, even if the potential licensee knows it would be practising the invention. An infringer will be liable for their own direct costs of litigation and some part of the patentee’s costs, as well as the indirect costs of management distraction and business uncertainty. In this case, for example, Videotron initially licenced simply to avoid the business risk from litigation. Second, a reasonable royalty is intended in principle to reflect the amount the infringer would have paid as a licensee, but, because of information problems, the reasonable royalty is unlikely to exactly reflect the royalty that would actually have been awarded, try as the court might. This means that the ‘catch-me-if-you-can’ strategy entails considerable risk. Further, if an injunction is a prospect in a sunk costs scenario, the infringer may have to settle for an amount considerably more than it would have paid for a licence. Consequently, it is not clear that an award greater than a reasonable royalty is necessary to avoid the ‘catch-me-if-you-can’ problem: see generally Ch 3 of Complex Products.

Further, the ‘catch-me-if-you-can’ problem is not really at issue if the putative licensee did not know of the patent, or believed in good faith that the patent was invalid or that it did not infringe; note that in Rivett [8] it seems clear that the defendant was an intentional infringer. An award greater than a reasonable royalty therefore risks an unnecessary chilling effect on a party who believes in good faith that they are not infringing.

Accordingly, in this case Lafrenière J considered that Videotron’s conduct did not weigh in favour an accouting:

[599] I disagree that Videotron’s conduct was inappropriate, let alone that it wilfully infringed Rovi’s patents. The evidence suggests that Videotron’s motives to refuse to renew the licence agreement with Rovi were well-founded. Its decision was based on a reasonable assessment of the necessity of the patents.

This strikes me as entirely sound.

(d) Whether Rovi practiced the invention of the patents in Canada

Videotron argued that “where the patentee does not practice their patents and only licenses them, it is ‘almost a rule of law’ to assess damages in terms of a reasonable royalty [602]. Lafrenière J considered this to be a neutral factor, as the evidence was mixed as to the extent that Rovi actually practised the invention in Canada [603], [604]. Since this issue is tricky and it turned on the facts in this case, I won’t comment further.

(e) Complexity of Calculating an Accounting of Profits

Lafrenière J noted that there is caselaw holding that the complexity of calculating an accounting may be a factor, because “the inventor is only entitled to that portion of the infringers’ profit which is causally attributable to the invention, i.e. ‘perfect compensation’” [605]. I take this to mean that complexity means uncertainty, and uncertainty means that the patentee may be over or under-compensated. However, “the calculation of damages can be just as complex as the accounting of profits” [605] This suggests it is not complexity, but relative complexity, that is important. As de Montigny J said in Philip Morris v Marlboro [2015] FCJ No 1564, cited by Lafrenière J, “the Court is essentially concerned with the proportionality of the accounting remedy in view of the length or extent of the infringing activity and the likely benefit of the accounting exercise.”

I’ll note that lost profit damages and an accounting are very similar in structure – it is the differential profit that is to be calculated, and the only difference is whether it is the patentee’s or infringer’s differential profits. The difficulty of assessing causation, and assessment of revenue and apportionment of costs, are likely to be equally difficult to calculate. Reasonable royalty damages, however, are assessed instead on the basis of a hypothetical negotiation, and so might be more or less complex that lost profit damages or an accounting, depending on the facts.

In this case, Lafrenière J held on the facts that “While the complexity of the evidence could be overcome, I am not satisfied that using any methods proposed by Rovi’s expert to calculate profits, which would be fraught with insufficient, speculative, and contradicted evidence would allow me to arrive to reliable and appropriate amount reflecting Videotron’s profits” [607], and accordingly he held this weighed against awarding an accounting. In other words, a patentee who wishes to elect an accounting should propose a sound methodology. This point strikes me as compelling.

I note that in general a reasonable royalty might also be very difficult to calculate in this type of case, where the patentee is claiming a reasonable royalty for the use of one relatively minor feature that is part of a complex product, depending eg on whether good comparables are available—though on the facts in this case Lafrenière J was able to arrive at a reasonable royalty relatively easily, as discussed below. So while I agree with Lafrenière J’s analysis in this case, in general the relevance of the complexity of calculation is a point that needs more development.

C. Reasonable Royalty

After deciding that Rovi was not entitled to an accounting, Lafrenière J went on to assess a reasonable royalty.

Videotron argued that should a particular Videotron system feature be found to infringe a valid claim, the appropriate remedy “is a one-time reasonable royalty, capped at no more than Videotron’s cost to remove or design-around the subject-matter of the relevant asserted patent claim” [612]. Lafrenière J accepted this submission. The evidence was that “the absence of such features would not impact the subscriber base for Videotron” [614] and “subscribers do not make choices around staying or leaving based on the kind of functionality at issue in this case” [615]. It is implicit that the only impact of the patented features on Videotron’s revenue is through the subscriber base and not on the rates that it would charge.

With one caveat, this approach is consistent with the standard hypothetical negotiation framework for assessing a reasonable royalty. The potential licensee’s maximum willingness to pay for a feature is the difference between the value of the product with the feature as compared with the best non-infringing alternative (NIA). In this case the NIA is not to include the feature at all. It is significant in this case that the NIA is not another unpatented technology which replicates the functionality of the patented invention; it is simply to remove the feature entirely. This illustrates that the “non-infringing alternative” is simply a way of framing the question of “what would the licensee have done if it had not infringed”? On the facts in this case, it would have removed the patented feature entirely.

There is a small caveat. Though it is not entirely clear from the brief discussion, it seems that the figure of $150,000 represents the ex post design-around cost. That is, are we assuming that the patented feature was already incorporated in the Videotron system, and $150,000 is the cost to remove it. If so, the $150,000 actually represents the holdup cost, rather than the value of the feature, and so is too high; arguably the true value was zero. With that said, Videotron’s submission, which was accepted by Lafrenière J, is not that the design-around cost reflected the value of the patented technology, but rather that the design-around cost “capped” the reasonable royalty. This is strictly correct. On the facts, a reasonable royalty reflecting the value of the invention is necessarily less than the ex post design-around cost, because the best non-infringing alternative is to remove the feature entirely. No doubt establishing the design-around cost was the simplest way for Videotron to establish a quantum, and safer than arguing that the correct royalty was zero.

The uncontroverted evidence was that the approximate cost for such a design change would have been $150,000 per feature [612], and accordingly that is the royalty that Lafrenière J would have awarded [624].