IPSC: Trademark

Mark Lemley

Irrelevant Confusion

True story: in 2006 World Cup, several hundred fans of Dutch showed up wearing pants in the colors of the Dutch national team, promotion from beer company. FIFA decided this was ambush marketing, trading on the goodwill of the team without authorization—the beer company needed to stop, and the fans were engaged in TM infringement for representing affiliation with FIFA. So they took off their pants. Here (with Mark McKenna) to defend the right of soccer fans to wear pants. This story is not alone. MLB goes after Little League teams—you can’t be the Cape Cod Cubs, because there might be some confusion over whether this is a MLB team. Insinkerator gets upset when its product is shown in Heroes in a way it doesn’t like; Caterpillar and Slip ‘n Slide sue over portrayals in movies, arguing TM infringement.

This could be a rant against TM overreaching, but it isn’t. The problem isn’t owners; a lot of these cases win—courts conclude that confusion exists. What’s going on? The extension of TM claims from traditional role, confusion over product/source, to confusion over affiliation/sponsorship. TM now applies in situations where nobody could possibly be confused about the source of the product/service itself. Historically we required competition—required for passing off. That doesn’t work in the modern world, because consumers may well believe, correctly, that companies make more than one kind of product. So we have to move beyond competition at least as to confusion as to source.

We also expanded TM to cover affiliation confusion, and again there was a good reason to do so: franchising. If you go to McDonald’s, you expect some guarantee of quality, even though you understand that the owner of the store here is different than the owner 5 miles away. Similarly, Apple may outsource production, but intend to stand behind the quality of the product. Correct answer: where this dynamic operates, find consumer confusion about whether the TM owner is standing behind the product.

While sponsorship claims do make sense, they also cause harm. Generate conflict between previously coexisting mark owners in different markets, geographic or product. Poor man’s dilution: allows mark owners to sue based on claim that people will think there’s some relationship—Dell for books and Dell for computers.

Second, starts to coddle consumers, and makes it easier to confuse them. Think of grocery store shelf product placement. If you want to buy soda, you go to the soda aisle, and see competing sodas, generic cola next to Coke and Pepsi. Is that confusing consumers, suggesting a relationship? Probably not, but in a world in which we argue that any suggested relationship can be problematic, that’s a weak read on which to rest nonconfusion. Rescuecom actually says that placing generic next to Coke is TM use. Now we’re in a world in which the only question is whether 10-15% people think there’s a connection.

Third, we don’t have a very good test for measuring sponsorship confusion. The factors in the traditional test are directed to source confusion. When directed at sponsorship, they either make no sense at all or naturally point in one direction, so courts tend to ignore them. We therefore get these cases wrong. We rely too much on intent, which is malleable and incoherent.

Fourth, interfere too much with free speech. Depiction in TV shows, expressive behavior like wearing pants—not traditional trading in commerce.

Sponsorship and affiliation law does some good, but leads to a lot of bad results. Merely thinking that there’s a relationship ought not be enough to find a violation of the Lanham Act. Ought to care only if it affects consumer behavior! False advertising law is the proper model, which requires materiality. A claim of 11,500 employees when a company actually has 11,600 is provably false, but not false advertising.

TM law is an offshoot of false advertising—a subset of false statements for which we presume materiality because of the nature of those statements: I am product X/I came from this source—reasonable to presume materiality. But with sponsorship/affiliation, the tail is wagging the dog. Reasonable to presume materiality only if it’s about source or quality control (franchising, closely related goods). After that, need to prove materiality.

Evidence: it’s rare that consumers care about affiliation in evaluating product quality whether there’s co-branding; they don’t punish one brand for the misbehavior of another even if they know they’re in a co-branding relationship. Show the injury and prove it, and you can win.

Jessica Silbey: Why go to materiality in requiring a stricter definition of sponsorship? Sponsorship could tie the test back to source—a financial backer. (I think this wouldn’t work because, e.g., brands pay for placement in movies all the time.) Where does the “affiliation” language come from?

Lemley: He thinks sponsorship is a problem too. Hard to identify things that are sponsorship and things that aren’t. Official sponsor of the US Olympic Committee = pretty clear sponsorship claim. But no representation of product quality, source, or anything other than relationship. If that’s actionable (if false), how can you distinguish a claim about the relationship between Caterpillar and the Tarzan movie on any sensible basis?

McKenna: Historically, the language of affiliation comes from the practice of outsourcing to subcontractor or subsidiary. The language is just too broad.

Lemley: The advantage of materiality is flexibility: lets you prove your harm.

Ramsey: Commercial advertising or promotion is a requirement for false advertising—also for this?

Lemley: Ideally, move this into false advertising. But probably can’t, more because of the requirement of competition; instead, we’d draw an analogy—we interpreted 43(b) to have a materiality requirement, so we can also interpret 43(a) to do so. (I’d note that the competition requirement in 43(b) has also been implied.)

McKenna: Courts actually used to do this, and then they stopped, perhaps because it made it harder for plaintiffs to win.

Lemley: His view is that noncommercial speech can sometimes be actionable, though it’s rare.

McKenna: Proof of materiality is likely to have a disproportionate effect in screening out claims against noncommercial speech, actually.

Q: Any standard that’s flexible will also have linedrawing problems. How will these standards be applied—what do you mean by “guarantees the quality of”—almost sort of cheating (what happens when Dr. Pepper gets control of certain storylines in a TV show), and what do you mean by affecting purchasing decisions—what if some consumers care some of the time?

Lemley: Briefly, yes materiality is a standard and not a rule. That doesn’t bother him because the other TM principles are already standards and aren’t working. Also, we’ve got a set of false advertising cases working out these problems. Guarantee of quality = rough cut. We think TM is right to presume materiality for product/source and for franchising.

Ilanah Simon Fhima

Qualifying for Trade Mark Dilution Protection: The Fame Standard

UK/EU standards v. US: An important topic, because this is the first screen for a dilution claim. Europe protects broadly against blurring, tarnishment, and taking unfair advantage of reputation. She’s been looking at decisions, including decisions in the registries such as OHIM. Trying to get a feeling of standards being applied.

Why have a fame requirement? It’s not very consistent with the underlying theory of dilution. Reward for investment? That’s not a very attractive reason—do we really want to reward that? For some kinds of harm—detriment to repute, or tarnishment—you can only have the harm if you’re already possessed of a reputation, so maybe fame helps screen. More convincing: fame is a component of association. For the senior mark to be harmed, you need a mark well-known enough that consumers in both classes will have heard of the senior mark. That’s the most promising justification. Also historical and pragmatic reasons, especially in the US.

US rejected niche fame in favor of wide recognition among the general consuming public; courts have generally protected household names. She, as a UK citizen, recognized most of the names on the list. She was surprised to find a bit of a distinctiveness test, which was one of the motivations behind the 2006 TDRA.

EU has a niche fame standard. But we don’t have a set percentage of fame. Line of cases says mark needs to be “known,” but courts won’t say how known. You might think that this makes it easy, but it’s not necessarily. Niche fame doesn’t serve TM owners all that well in Europe: protection is available, but only in one’s own niche. Current case involves Intel: the courts have taken the approach that Intel is famous for computer processors; once the mark is used outside, e.g. for telemarketing, consumers won’t make associations across markets.

Also some uncertainty about the geographical market. Must be known in a “substantial part” of the relevant territory. Ongoing litigation on this, as there is with most TM issues in Europe.

Factors used to determine fame in the US: duration of advertising, volume of ads, geographical extent of ads, mediums of ads, unsolicited publicity, third party recognition of status (brand rankings, legal recognition), etc. Two decades is a good length of time; $20 million ad spend/year has been said to be on the low side; many cases involve worldwide marks; coverage in fashion magazines.

Sales: looking for 100s of millions in dollars and also significant volume, depending on the market—designer clothing would require less than children’s toys. Some of the factors amended out by TDRA have crept back in: duration and extent of use.

Actual recognition: what she found surprising was this wasn’t treated as vital—courts are not requiring actual survey evidence. Registration—courts have paid relatively little attention. Some attention paid to endorsement and licensing. Use by third parties has crept back in—if other people are using the mark, it won’t be standing out. General observation here and in EU: precision. A surprising number of companies haven’t been giving precise information about the mark at issue, and this never helps them.

EU doesn’t have a statutory test, comes from case law. Looks at intensity of use (range of products, frequency of use)—lower standards than US, in part because the size of the EU and its component parts is smaller than the US, but 18 million euros found insufficient in one case; geographic extent; duration of use (2/3 decades generally looked for, but Viagra was famous in a very short time). Sales: often millions required; market share thought to be particularly important, used by a bunch of courts in rejecting fame, but it’s not clear what the required share should be and one case allowed 3% market share.

Conclusion: US standard is stricter, and courts are keeping to it and introducing new strictures, which is not what happened under the FTDA. Niche market fame comes at a price for the EU, limiting TM owners’ rights.

Ramsey: What are TM registration offices doing in the EU? Several attorneys say marks are being rejected based on examiners’ conclusions that their marks would dilute other registered marks.

A: In the UK, you’re likely to have the opposite effect—won’t look at fame ex officio, wait for TM owner. Not quite sure how much to look at US registration decisions, because there are so many cases.

My Q: Registration and trade dress interact—the Reese’s Peanut Butter Cups case (under the FTDA) used lack of registration of the claimed elements of the trade dress because the alleged diluter wasn’t copying the whole trade dress, only a couple of elements. So you use registration in the US to screen out elements that in themselves aren’t famous even if they are in combination. What about the EU?

A: Hasn’t seen many trade dress cases of this sort—one that she thinks of, Kitchenaid, shouldn’t have been registered.

Mark McKenna (with Mark Lemley)

Owning Markets? Trademark Law and Market Foreclosure

Concern: confused consumers will impute bad quality back to the TM owner. Turns out that the empirical evidence on this is weak, unless the products are very closely related. Marketers—when they talk about “far” extensions, they mean going from soda to chips, while courts treat those as very closely related.

Empirical evidence for other arguments: when someone uses a known mark like APPLE for unrelated goods, that has some market preclusion effects on Apple, because it prevents Apple from moving into that market later. Usually tacked on to a free riding argument. Operates on assumption that Apple should have the right to expand into that market (under the Apple name), which is a free riding argument. Motivates courts in merchandising and keyword ad cases.

Those aren’t the concerns we talk about, but they’re doing actual work, and the empirical evidence is relatively strong compared to quality feedback. If you use a mark in a separate market, it can affect a TM owner’s ability to expand, especially if the first product fails.

These arguments don’t depend on confusion. They’re claims about market ownership. Not an accident that these arguments come up the most where claims about confusion are really tenuous or in cases where courts aren’t even trying to assess confusion—one of Posner’s defenses of dilution is just the naked free riding argument.

This set of TM cases is like the derivative work right cases in copyright. Ordinarily we make incentive arguments for derivative works. So what kind of incentives might we have here? Ordinarily in TM we talk about incentive to invest in quality of product, or quality of message about product. If there’s no representation of quality at issue, we should confront whether we need incentive. But what about incentive to invest in brand personality? Well, there’s still plenty of incentive for that—the most significant return comes in the markets in which you actually participate. There shouldn’t be a substantial effect on incentives, though in theory there could be some, based brand image in markets in which you might someday participate.

Second part: pure free riding. Use on noncompetitive goods: there’s value to the use, which is why the other party is using the mark in the other market; isn’t that a harm? Answer: no, that’s not a harm. That’s consumer surplus. There’s value in that market, and the question of whether we give it to the TM owner is the usual patent/copyright question—whether we need to give it to the TM owner to compensate for a market failure.

Efficiency story that we might be persuaded by: Our formulation, because we’re talking about claims where there’s a representation of quality, that covers circumstances where a party is preparing to enter another market. But it’s conceivable that a party actually investing ought not to have to take the chance that consumers will think they’re not responsible for quality in the extended market: a sort of ITU. But those people should have to file ITU applications. We shouldn’t just assume/give them rights in markets they’re not in.

Q: Do product and geographic markets differ?

Lemley: On the ITU point, yes. Given the system of registration, you end up controlling geographic markets in which you don’t participate.

McKenna: Geographic market is likely to be a much lower restraint in today’s market. Legitimately true that people will expect quality control across geographic boundaries. Product markets are different.

Lemley: And entry is more likely across geographic markets than it is across product markets—if I sell computers in California, I may more readily expand to New York than to DVD players.

Q: Worries that analogy to copyright is fraught. There are moral rights arguments in favor of derivative works rights, and that concedes a lot of ground on TM as property.

Lemley: We use the analogy to condemn the argument, not support it. Moral arguments are doing a lot of the work here, and pulling that out of the shadows—the sense that the defendant shouldn’t be able to make money using the mark—is the first step towards killing it: consumers aren’t hurt, the TM owner isn’t hurt, so where is the problem?

McKenna: These arguments are always tacked on to standard TM arguments, but need to be disentangled.

Fhima: In Europe, these arguments have been disentangled. Does that make it any better?

Lemley: No, because the ECJ got the argument wrong in L’Oreal. But at least we can see the court affirmatively saying that we’re stopping this practice because we don’t want consumers to be able to find out that this perfume smells like this other perfume.

McKenna: It’s a backdoor patent: if you can’t tell people that the perfumes smell alike, why make it?

Some conversation about search costs: Maybe, Lemley suggests, free riding is good by decreasing search costs—it makes a new product stand out in the market! If there’s no consumer/TM owner harm, that’s a benefit!

William McGeveran

The Trademark Fair Use Reform Act

Concerned with the effect of difficult defenses on expression. The full benefits of expression are internalized neither by the TM owner nor by the defendant (e.g., the producers of Heroes who want to show Claire Bennet’s arm getting caught in the garbage disposal). Increases the difficulty of producing speech if you have to clear rights constantly. Not to mention the chilling effect of threat letters and lawyers’ analysis—there are 10 defenses you might raise, and you might win any of them. When litigated, you usually get speech-protective results, but that doesn’t help someone faced with a C&D and the prospect of expensive litigation. It’s profitable to focus on the front end, but we should also focus on the back end, defenses operating independent of consumer perception/confusion.

Narrow proposal: develop a cleaner structure for TM fair use that is more sensitive to administrative costs. Broadly: we should think about having more per se rules as part of the solution.

First, emphasis on administrative costs. Administrative costs vary inversely with avoiding error in decided cases. TM law is biased to avoiding substantive error—complex tests with high administrative costs, mostly designed to avoid error in allowing confusion where it could be halted. Of course this also produces false positive problems. To reduce administrative costs, we need to accept more risk of error. But the error invited is false negatives, but in some cases we need to tolerate confusion to protect expression—KP Permanent.

Second, rules v. standards. All TM fair use standards used now—whether nominative or statutory fair use, First Amendment, or use of confusion doctrine—are standards: ex ante determinations, complicated, hard to predict. TM fair use is an area where ex ante rules would make sense. You have repeated fact patterns, meaning a rule has a high return on investment and that a rule provides certainty for a bunch of people. Also, free speech errors are very harmful. If you can align a rule with a most-of-the-time right result, then you don’t get much out of a more elaborate decisionmaking standard.

Third, unbundling. Confusion and defenses are usually bundled together in TM. That might help the defendant, some commentators say—puts burdens on plaintiff to show confusion. Procedurally this is terrible on defendant because it increases the expense and time of the case. It makes sense to think about defenses segregated from likely confusion. Fair use should be a threshold screen, and courts should not—as is too often done—allow confusion to be smuggled back in out of a worry about error costs.

Fourth, common law rulemaking. Courts have been chicken about using common-law decisionmaking power in TM. Descriptively, courts aren’t as aggressive as they were. (I think at the SCt level this isn’t true, and only true on certain aspects with respect to lower courts.) He therefore proposes a statute as a discussion point.

Categorical exemptions for political speech: error costs unlikely and false positives risks are too high. But would need exemption for the exemption: using a name of another group engaging in political speech. (Sorry, Jackson Browne!) Another exemption: uses in fictional works. Allow product placement deals, but free use is the rule for people who don’t want such deals. Repeat fact patterns with low risk of error and serious error costs for false positives; can be determined without reference to likely confusion.

Jim Gibson: Safe harbors sometimes drag people into them who might otherwise be more aggressive, but compared to the alternative in TM it might be a good idea. What about sensitivity to new uses? TM use wasn’t a fraught category 15 years ago, and then keywords showed up.

A: Categorical rules have to be the first line of defense, but they can’t be all of it. Antitrust: if you’re not within a bright line rule, you go to the case-by-case analysis.

Gibson: risk is that failure to fall in safe harbor is treated as informing later confusion analysis.

A: He’d also articulate more carefully what the subsequent context-specific defenses should be.

Laura Heymann: Why not model it more like §107 in copyright: giving courts specific factors. Subject-matter categories risk an expressio unius est exclusio alterius analysis: if your activity isn’t listed as a defense, then your activity can’t be ok.

A: Statutory drafting should say explicitly that such reasoning would be illegitimate. Copyright fair use is not a model of avoiding administrative costs. TM law is in worse shape than copyright, but maybe the second line of defense should look like §107. The temptation of reform is to try to fix everything, but we should take some situations cleanly off the table.

Q: relationship between bright-line rules and Rule 11? Still in interest of people to bring these lawsuits, but at least you have a possible way around it—a Rule 11 motion.

A: Hard to make an argument of objective baselessness, but clearer rules would help, and would also help respond to C&D letters briefly.

Lemley: more effective model: anti-SLAPP legislation—lawsuits get treated separately, requiring plaintiff to present case immediately without discovery. Turns out to be reasonably powerful deterrent threat.

Heymann: How do you define the market in the McKenna/Lemley analysis? Chips might not affect the soda market, and Coke might have no intent to enter the market (under the name Coke). But what if you could show that has an effect on willingness to purchase any ancillary Coca-Cola products?

McKenna: There’s some empirical support for the idea that it interferes with consumers’ willingness to try other unrelated products. Question is: what do you do with that? Is that risk in the undefined future that the TM owner might expand into a third area and might have more trouble there a reason to stop activity now? In other areas, we don’t tolerate that much speculation—e.g., antitrust. Confusion captures much of this—likely expansion.

Lemley: What you’ve described is the proper structure of a dilution by tarnishment claim: you have sold a product that people understand is not mine but nonetheless creates such bad associations that it affects my core brand.

Ramsey: Back to your search cost idea. Maybe it’s not a good idea to rely on iPhone to buy chips.

McKenna: that contradicts the argument about why it’s efficient to allow TM owners to expand into third markets, even though they won’t know anything about the product but the name. If we want people searching in the chip market on the characteristics relevant to chips, then we shouldn’t allow TM owners to expand into new markets.