Search Engine Liability for Selling Keywords Redux–800-JR Cigar v. GoTo.com
By Eric Goldman
800-JR Cigar, Inc. v. GoTo.com, Inc., 2006 WL 1971659 (D. N.J. July 13, 2006)
This is a confusing but important case about search engine liability for selling trademarked keywords. Because of the opinion’s convoluted nature, there’s no clear winner. However, the plaintiffs have to be happy with this opinion, and if I were Yahoo/Overture, I’d be looking to settle this case rather than continue before this judge.
Overview
JR Cigars is apparently a well-known discount seller of cigars. GoTo.com (now Overture, now part of the Yahoo empire) sold keywords including “JR Cigar,” “J R Cigar,” “J & R Cigar,” “J-R Cigar,” “JRCigars.com,” and “800 JR Cigar” to various JR competitors. JR sued the competitors and GoTo in 2000 (6 years ago!). Subsequently, the competitors all settled, leaving GoTo as the sole remaining defendant. In this ruling, JR Cigars seeks summary judgment on the trademark infringement and dilution claims, and GoTo seeks summary judgment on that and two ancillary claims.
The outcome is a little complicated, but here’s the high level overview. The judge denies JR’s motions, saying that some factual issues require a trial. Further (and this is part of the weirdness), the court says that because the various competitors have dropped out, JR should proceed only under a contributory infringement theory—not the direct infringement theory litigated in this ruling. Meanwhile, the court grants GoTo’s summary judgment on the ancillary claims, but surprisingly rejects a 230 defense as the rationale.
Superficially, GoTo won this ruling by defeating JR’s SJ motions and winning some of its SJ motions. However, I think this case is a net loss for GoTo because GoTo lost on some key issues and because the court appears very sympathetic to JR’s claims. In particular, the judge appears to have accepted the plaintiff’s argument that displaying keyword-triggered ads can “divert” searchers (a point I vehemently disagree with in my big paper, where I argue that courts cannot determine “diversion” without, at minimum, reviewing the ad text that consumers responded to). As a result, if I were GoTo, I’d settle this case before it gets worse.
To better understand the nuances of this complex decision, let’s look at some details.
Trademark Infringement
Closely following the GEICO v. Google precedent, the court says that GoTo’s sale of keywords constituted a trademark use in commerce. Specifically, the court says:
GoTo makes trademark use of the JR marks in three ways. First, by accepting bids from those competitors of JR desiring to pay for prominence in search results, GoTo trades on the value of the marks. Second, by ranking its paid advertisers before any “natural” listings in a search results list, GoTo has injected itself into the marketplace, acting as a conduit to steer potential customers away from JR to JR’s competitors. Finally, through the Search Term Suggestion Tool, GoTo identifies those of JR’s marks which are effective search terms and markets them to JR’s competitors. Presumably, the more money advertisers bid and the more frequently advertisers include JR’s trademarks among their selected search terms, the more advertising income GoTo is likely to gain.
I could take serious issue with a number of aspects of this analysis, but ignoring the details, the court is just following the GEICO precedent (and the uncited American Blinds precedent). However, it is noteworthy that the court specifically called out GoTo’s keyword suggestion tool—this is not the first time courts have not looked at this tool favorably. In response to GoTo’s argument that the tool operates automatically, the court’s response is cutting: “the Search Term Suggestion Tool permits GoTo to channel advertisers directly to JR’s trademarks by demonstrating quantitatively the potential for successful advertising, thereby implicitly recommending those terms to advertisers.” From my perspective, these tools just seem like huge liability traps. As a result, if I ran a search engine, I would reevaluate the cost-benefit equation of the keyword suggestion tool.
Having found the predicate trademark use in commerce, the court next tries to apply the standard multi-factor test to determine if the use creates a likelihood of consumer confusion. I say “tries” because GoTo isn’t advertising its goods and services using JR’s trademarks (one of the reasons why the court got the use in commerce analysis wrong), so the factors don’t really fit. The court frequently has to say “I know that GoTo sells search advertising and JR sells cigars, but they are really in the same business.” It doesn’t work.
As part of the analysis, the court considers the initial interest confusion doctrine. I can’t say enough bad things about this doctrine and how it leads courts astray. (I’ve already said plenty–about 37,000 words–deconstructing the initial interest confusion doctrine in my big paper). In the Third Circuit (which covers New Jersey), the leading IIC case is Checkpoint. As IIC cases go, the Checkpoint case was comparatively restrained because it clearly limits IIC to competitive situations. Nevertheless, the court has no problem extending Checkpoint to GoTo, who does not compete with JR in any way, because it treats GoTo as the proxy of JR’s competitors buying the ads. As further “evidence” of initial interest confusion, the court is also impressed with some evidence of diversion marshaled by JR, such as claims by JR that there were 1,000 diverted customers and a 24% diversion rate.
Ultimately, out of the multi-factor likelihood of confusion test, the court finds that four factors favor JR (mark similarity; mark strength; overlap in targeted customers; product relatedness), none favor GoTo, and four require a trial (evidence of actual confusion; length of time without confusion; intent, and various aspects of the initial interest confusion doctrine). (A couple of Third Circuit factors were irrelevant/indeterminate).
Due to these open factual issues, the court defers consideration of GoTo’s fair use defense, saying it too requires a trial on the facts. However, in passing, it offers GoTo some hope, saying that fair use probably applies “where GoTo permits bids on JR marks for purposes of comparative advertising, resale of JR’s products, or the provision of information about JR or its products.” This appears to implicitly endorse much of GoTo’s trademark policy, which requires some ad relevance to the keyword. However, the court continued that “fairness would dissipate, and protection under a fair use defense would be lost, if GoTo wrongfully participated in someone else’s infringing use.” I suspect the latter reference supports the court’s preference to analyze these claims under a contributory theory, not a direct theory.
Dilution
The court finds that JR Cigars is a famous mark. The court rejects out of hand GoTo’s argument that it has not made a commercial use in commerce. Then, the court says that there are factual issues about whether actual dilution has occurred because “confusion, actual or likely, is one factor bearing on the dilution analysis,” so the court punts on this issue pending the trial on the likelihood of consumer confusion issue. Equating actual dilution with likelihood of consumer confusion is, at best, unorthodox.
Other Claims
JR also sued GoTo for violating the federal Telemarketing and Consumer Fraud and Abuse Prevention Act and state consumer anti-fraud law. GoTo defends on 47 USC 230. The court rejects that defense because it doesn’t think GoTo qualifies as an interactive computer service. This conclusion is completely inconsistent with the extensive precedent that the court missed—most recently in the DiMeo v. Max case, but there have been plenty of others. The court also says that 230 doesn’t protect against the “fraud” of using trademarks in the bidding process. Because I don’t understand how GoTo’s acts meet a prima facie claim of these acts, I don’t understand where the fraud is. And, if the objectionable conduct is based on advertiser-supplied keywords and search listings, 230 should apply to any non-IP claim.
Nevertheless, the court dismisses both claims because (1) JR lacks standing to enforce the consumer anti-fraud statute and (2) the Internet should not be governed by a telemarketing statute just because telephone wires are used to make connections.
Why is JR Cigars Still Litigating this Case?
In my recent talk on web marketing issues, I made the possibly controversial statement that most keyword purchase cases aren’t worth litigating on a cost-benefit basis. My thinking is that companies can spend their money on marketing or on lawyers, and it’s far more likely that they will get better value (more new customers, better ROI) from marketing than from litigating.
This case seems like a textbook example of this argument. In the 2 years when GoTo was selling the applicable keywords, GoTo made total revenues of $345 from them. $345!!! And some of that was for non-cigar advertisers on the “JR” term and clicks to a site operated by plaintiff’s lawyers. Take those away, and there were not a lot of diverted customers that JR Cigars is litigating over. (JR offered some evidence that it lost 1,000 diverted customers in the 2 year period, but I’m fairly confident that this number is overstated because there’s no way to measure diversion without seeing ad copy). Further, with GoTo’s (lack of) revenues, JR will have difficulty winning big damages (and only a modest chance of the court granting attorney’s fees). Meanwhile, JR has been litigating this case for 6 years. I’m 100% confident that JR would have made a lot more profit pouring its litigation expenses and managerial time into marketing, not litigating.
Conclusion
I hate Internet cases like this where the case has been growing mold for years. Usually the law has completely changed between the initial pleadings and the ruling (and sometimes between the motions for summary judgment and the ruling), creating the possibility of very anachronistic rulings. Especially with respect to the 230 discussion and some of the factual discussion about GoTo’s business practices, this case really shows its age.
As for the trademark issues, up to this point, this case resembles the GEICO v. Google case–selling trademarked keywords is a trademark use in commerce, but it may not create a likelihood of consumer confusion. It will be interesting to see where this case goes next (if it doesn’t settle), especially if the parties really explore the contributory infringement angle.
With respect to the contributory infringement claim, it will be interesting to see if this court considers the Merck precedent (another case not cited), which held that advertisers were not making a trademark use in commerce in buying the keywords. If so, GoTo can’t commit contributory infringement because there’s no predicate direct infringer.
Meanwhile, this case reinforces the completely unsettled nature of keyword purchases and trademark law. Combined with the other two search keyword cases this year that reached directly opposite results to each other (Edina Realty and Merck v. Mediplan), there really is no way of predicting how the next case is going to come out. In some ways, the law applicable to search engines selling keywords has not advanced in any meaningful way from 2000 when this case was first filed.