Barton Beebe gave a brief presentation on his empirical study of TM confusion cases and how the multifactor test for confusion plays out there, in contrast to Long’s paper on dilution. He only looked at reported cases, though.
Which factors are important? Which irrelevant? Do they tend in practice to stampede, that is, to be treated by judges as all pointing the same way? How are surveys used? What’s the role of inherent distinctiveness? The answers may be relevant to predicting results under the TDRA.
Highlights: 2d Circuit dominates, accounting for 33% of TM opinions, and the 9th Circuit was second with 16%. Impressive circuit variation – district courts preliminarily enjoined defendants 41% of the time in the 2d Circuit, but 69% of the time in the 9th Circuit. Of 331 district court opinions, 65 addressed survey evidence, 10% crediting that evidence and 7% ruling in the direction suggested by the survey. But in the big cases, survey evidence is presented; surveys are not evenly distributed. Only 6 cases drew an adverse inference from failure to present survey evidence. (I have direct experience with these last two statements, having worked on a huge case between enormous drug companies in which plaintiff had a dilution survey, defendant had a confusion survey, and the judge drew a negative inference from plaintiff's failure to conduct a confusion survey.)
Beebe found a curious result on inherent distinctiveness: judges use the Abercrombie spectrum about half the time, but are fairly wary of it. Actual strength v. inherent strength: Some cases found divergences between actual and inherent strength (one strong, the other weak), and in almost all of these cases, the commercial strength judgment trumped the inherent strength judgment. Beebe thinks this is a good thing. It goes to the TDRA’s fame factors, which include one empirical factor (actual strength) and a bunch of formal factors (inherent strength), which is sort of weird. Likewise with the blurring factors, one is empirical and the others are formal.
Key factors in confusion cases: Similarity is dispositive; defendant’s intent is almost dispositive. Then follows proximity of goods, actual confusion, and strength. Courts tend to stampede the factors in the direction of their decisions.
Kur: Is there any empirical data from the EU similar to Long’s work on US courts?
With respect to the Intell case, if consumers think (correctly) there are two companies with that name, so what? What harm does that do? Likewise, with the Wal-Mart case, is it surprising or troubling that consumers thinking about Wal-Mart said that the Wal*ocaust shirt made them less likely to shop there, given that they may have read the shirt as a criticism of Wal-Mart’s practices?
Hobbs: He isn’t aware of any statistics from the EC, but he’s particularly interested in the effect of disparity between goods and services. His intuition: the further apart the goods and services are, the harder it is to convince the tribunal that there’s any harm to the plaintiff or benefit to the defendant from a similar use. EC TM law won’t work as an algorithm; we have to treat it as identifying a series of factors to be weighed in any given case.
In CTM registration decisions, he detects a clear trend in favor of claimants who are seeking extended protection. But he cautions that the law in Europe doesn’t seem to be the same concept as the American idea of dilution: we are divided by a common language.
Long: Product similarity: She didn’t specifically measure that, but subjectively her sense is that people were much more likely to sue when the products were in similar or closely related markets. She’s not sure how that affected likelihood of success, though she can think of more injunctions granted in related fields than in unrelated ones. Also, courts are concerned about granting injunctions that would affect speech, as with Wal*ocaust. Courts were sensitive to that issue, but only to the extent to which the parties brought it up; free speech defenses depended on good lawyering, since courts weren’t willing to go out on a limb without support from the defendant.
Jane Ginsburg: Picking up on the Hobbs remarks, we don’t know what we mean by dilution. The Intel survey point is spot-on: knowing that two entities share the same name doesn’t show a link between them, nor does it show impaired distinctiveness unless we think that “more than one” counts as impaired distinctiveness, which is a completely circular definition.
Inherent distinctiveness is or should be relevant even though actual distinctiveness is the trump. If your mark consists of a descriptive term, you should get less protection even if it’s well-known, because the costs of extending protection to a descriptive mark are greater (even if the benefits are the same).
What fascinates her about Beebe’s study is the significance of intent, which she thinks should be irrelevant if we take consumer perception seriously. There’s always been the cute shortcut of assuming that that the defendant succeeded in its intent, but she thinks that’s facile. If intent is key, then we’re closer to Dinwoodie’s unfair competition regime perhaps than we pretend.
Jacoby: Distinctiveness is like pregnancy: either you are or you aren’t. (I’ve got to say, I wonder if the people who say there’s no such thing as being a little bit pregnant have actually been pregnant. There is a big difference between not pregnant and pregnant, but also a huge difference – medically, experientially, and legally -- between one month pregnant and nine. And while roughly 2/3rds of pregnancies will come to term in the absence of intervention, one dilutive use is not naturally going to progress to use on a zillion products.) If there’s a second Intel, then there’s a third and a fourth and you’re no longer distinct.
As for the “so what” objection: With respect to tarnishment, there is a dramatic impact from things like Wal*ocaust – as with the Anheuser-Busch “Michelob Oily” case, where people said they were much less likely to drink Michelob after seeing the parody ad. On a national level, the impact on customer patronage can be great. (Comment: No one followed up with the survey respondents and see how often they’d drunk Michelob/shopped at Wal-Mart three months after seeing the ads. Their unconsidered statements of future intent are terrible evidence of actual purchasing behavior.)
Michael Carroll: Given that dilution is always paired with confusion, is it the case that dilution works to bolster weak confusion claims – that is, is there judicial acceptance of or hostility to dilution when the plaintiff’s confusion case is weak?
Long: Courts are not necessarily hostile to dilution per se, because they don’t know what the heck dilution is. So courts are responding to particular plaintiff presentations. Early on, courts tolerated more TM owner behavior than they did later. Consistently, courts don’t like TM owners using dilution as a property-like theory. They’re also hostile to use of dilution to get a market advantage against a competitor. In the early days, there wasn’t as close a relation between the strength of confusion and dilution claims: you could have strong confusion claims and win on dilution anyway or weak confusion claims and win on dilution anyway. By the end, you get injunctions where the confusion claim is very strong or the confusion claim is very weak; you don’t get injunctions in the middle of the spectrum.
Mühlendahl: EU law seems to be developing in the direction of multifactor tests for similarity and for likely confusion. Intent isn’t a big part of the analysis because of the relative absence of discovery (not to mention the fact that rights are often assessed without use). So EU adjudicators end up relying heavily on inherent and acquired distinctiveness and abstract assessments of similarity.
Dreyfuss: Asked Jacoby to comment on the effects of changing the number of milliseconds it takes for a consumer to recognize a mark. A few nanograms of polonium just killed someone, so it’s not crazy to think milliseconds matter.
Jacoby: In consumer research, the first brand someone mentions is the one they’re most likely to buy. Anything that diminishes the speed of retrieval and makes a brand less likely to come to top of mind in a product category will correlate with decreased purchase likelihood. Consumers spend almost no time in front of a frequently purchased good – 4 seconds. Information provision does not equate to information impact; consumers don’t spend much time extracting information unless it’s a nonfrequently purchased good or service. Thus lookalikes may be enough to get purchases. It’s not the milliseconds but their ultimate correlates that are the issue, and research on that is ongoing.
Marjorie Heins: The Michelob case got her interested in IP. It seems so clearly a critique of Michelob, and TM was used to circumvent defamation law. She thinks the Wal*ocaust website had a similar purpose. It’s worth pointing out that market harm resulting from criticism should not be the kind of market harm cognizable under IP law, just as we distinguish between kinds of market harm in copyright’s fair use analysis.
Sheldon Halpern: The Michelob court said that 20% of the respondents were confused and thought that Michelob was selling Michelob Oily – we are so much more comfortable talking about likely confusion. It’s with dissimilar goods that we get very uncomfortable. (Comment: Not sure who’s the “we” here.) In the “death of a thousand cuts” theory, any other use is going to constitute dilution; it’s standardless. So what would the EU do with a book called “Just Think About It”? That’s close to Nike’s mark.
Hobbs: To win, I would have to show theft of an advertising benefit, that the trigger function of Nike's TM gives the book selling power it wouldn’t otherwise have. There is authority that merely calling another’s mark to mind – a mere mnemonic effect – is not enough to trigger liability. But we have no idea, then, where the line is to be drawn.
Bone: To Long: what about possible settlement effects? (E.g., effect in the Intell case.) Especially with the initial success spike in cases, lawyers can use the changed legal rights to get different settlements.
Long: She controlled for that by including filed but settled cases.
Bone: Unless you have a really strong view of consumer autonomy, you don’t really care about confusion but about confusion’s harms. Confusion as to affiliation is just unlikely to do harm. Factors that are probative of harm, such as mark strength and defendant's intent, might be the most important.
Hobbs: Intent isn’t necessary to liability because the consequences are the same whether intended or not, but it’s a one-way proposition. Intent can be inculpatory – who knows better than a trader the mysteries of his own trade? One asks, could this have happened by accident? If not, why was it done? That line of inquiry leads you toward particular outcomes.
Herbert Schwartz: Judges make up their mind first and then evaluate the factors and the surveys later, and TM cases are much more that way than other IP cases. Judges feel very competent to decide TM cases based on their gut reactions. Brief comment on the Hershey case, with which he was involved: That was an unusual situation, both because it was a use on similar goods and because it involved the M&M mark which was itself so strong that it dominated when added to the trade dress.
Posts
No comments:
Post a Comment