Ninth Circuit Reconsiders SEO-Destroying Injunction Against DMV.Org–TrafficSchool v. EDriver (Joint Blog Post)
By Rebecca Tushnet and Eric Goldman
TrafficSchool.com, Inc. v. Edriver Inc., 2011 WL 3198226 (9th Cir. July 28, 2011)
[Over the years, Rebecca and I have blogged dozens of the same cases. However, we've never done a joint blog post until today. Rebecca and I had both blogged on the district court ruling in this case, so it made sense that both of us were going to blog on the Ninth Circuit ruling. It made even more sense for us to do it together! Rebecca's post. Warning: both Rebecca and I write long blog posts, so our combined effort is a beefy 3,800+ words.]
District court opinion discussed here. Judge Kozinski has long been a proponent of political-speech-level protection for commercial speech, and that position has its influence on this opinion, which is written in his usual highly readable style. You can also see here the growing lawsuit-limiting effect of standing doctrine, though here at least there was an actual factfinding before the judges decided whether the plaintiff had been harmed.
As Judge Kozinski explains, “Consumers visit DMV.org for help renewing driver’s licenses, buying car insurance, viewing driving records, beating traffic tickets, registering vehicles, even finding DUI/DWI attorneys. The more eyeballs DMV.org attracts, the more money defendants earn from selling sponsored links and collecting fees for referring site visitors to vendors of traffic school courses, driver’s ed lessons and other driver-related services. This seems like a legitimate and useful business, except that some visitors mistakenly believe the site is run by their state’s department of motor vehicles (DMV).”
After trial, the district court found a violation of the Lanham Act but not of California’s UCL. Its injunction required every site visitor to go through a splash screen with a disclaimer. (The Electronic Commerce & Law Reporter blurb described it as a pop-up. It’s not.)
The district court found that plaintiffs “failed to prove … that they have suffered an injury in fact and lost money or property as a result of Defendants’ actions,” and that they “provided no evidence showing a causal connection between Defendants’ actions and any harm Plaintiffs incurred.” Edriver argued that this finding proved that TrafficSchool also lacked Lanham Act standing. This was wrong because California’s UCL defines standing more narrowly (pecuniary injury and “immediate” causation) than the Lanham Act, which only requires a plaintiff to believe that it’s likely to be injured.
What about Article III standing? The UCL also defines injury in fact more narrowly than Article III does, so the loss doesn’t necessarily preclude Article III standing. The district court didn’t independently analyze Article III standing, but the court of appeals did, looking for injury in fact, causation, and redressability. The key here was injury in fact.
Article III injury exists in a false advertising case if some consumers who bought defendant’s product under a mistaken belief fostered by the defendant would otherwise have bought plaintiff’s product. This can be proven by actual market experience and probable market behavior. Probability is fine “because proving a counterfactual is never easy, and is especially difficult when the injury consists of lost sales that are predicated on the independent decisions of third parties; i.e., customers. A plaintiff who can’t produce lost sales data may therefore establish an injury by creating a chain of inferences showing how defendant’s false advertising could harm plaintiff’s business.”
TrafficSchool produced ample evidence that it competes with DMV.org for referral revenue, sometimes with the same course providers. “Sales gained by one are thus likely to come at the other’s expense. Evidence of direct competition is strong proof that plaintiffs have a stake in the outcome of the suit, so their injury isn’t ‘conjectural’ or ‘hypothetical.’ Plaintiffs also presented testimonial and survey evidence that a ‘recommended by DMV’ endorsement is an important factor in consumers’ choice of traffic schools and driver’s ed classes. It stands to reason that defendants will capture a larger share of the referral market—to plaintiffs’ detriment—if they mislead consumers into believing that DMV.org’s referrals are recommended by their state’s DMV. Plaintiffs have therefore established sufficient injury for Article III standing.”
Lanham Act standing requires (1) a commercial injury based upon a misrepresentation about a product; and (2) a competitive injury. (Footnote: a plaintiff can show both these things and still lack standing if the purpose of its false advertising suit is to enforce someone else’s statutory rights—plaintiffs here couldn’t sue on the California DMV’s trademark rights. But they sued to enforce their own right to be free of unfair competition.) Defendants argued that DMV.org was a publisher, while plaintiffs’ sites are self-promotional tools, but plaintiffs introduced evidence that they compete in the referral market.
Defendants argued that the only injury here was to the public. Though plaintiffs didn’t prove “identifiable” injury to themselves, the standard for “commercial injury” was different. Likely injury is all that’s required, not actual injury. “[W]hen plaintiff competes directly with defendant, a misrepresentation will give rise to a presumed commercial injury that is sufficient to establish standing.” Kozinski defended this presumption; advertising is about competing for consumer dollars, and misleading ads can upset competitive positions. “Moreover, the Lanham Act is at heart a consumer protection statute. Requiring proof that defendant’s ads caused plaintiff to lose sales as a prerequisite to bringing suit would frustrate its ability to act as the fabled vicarious avenger of the consuming public.” The court didn’t need to decide whether the presumption was conclusive or rebuttable because defendants didn’t provide any evidence to rebut it.
The remaining question was whether the ads were misleading. They were.
When plaintiffs sued, Californians who googled (Judge Kozinski fears no neologism, especially when it is useful) “dmv” or “drivers ed” would see DMV.org’s sponsored listings for “ca.dmv.org” or “california.dmv.org.” respectively. “While there’s nothing inherently misleading about sponsored search results, they can mislead if they are named so as to give a false impression as to the likely sponsorship of the website to which they refer.” The ca. and california. prefixes were obviously designed to suggest official affiliation. The website design also mimicked an actual DMV site, copying slogans and state symbols, and linking to DMV-related pages like applying for a license, registering a car and signing up for traffic school. The disclaimer was easy to miss because it was displayed in small font at the bottom of each page, where many consumers would never scroll.
There was also plenty of evidence of actual confusion, including two declarations from confused individuals and hundreds of emails from obviously confused consumers. “Some of these emails contained sensitive personal information that the typical consumer wouldn’t share with a commercial website”:
My boyfriend George [redacted] … got a ticket in South Carolina in 2006…. Mr. [redacted] driver’s license number is [redacted]. His date of birth is [redacted]. I have his social security number if needed, but I don’t want to put all of his personal information on this e-mail if possible…. I was told by Central Court for Lexington County in South Carolina that if we contacted the Arkansas DMV that y’all would be able to tell us what court this is in and where to pay the ticket.
Law enforcement officials and state DMV employees were also confused, seeking help with, among other things, DUI cases (again a situation where the subject undoubtedly wouldn’t like that information out in the open). Two California cities, a private law firm in Texas and a number of newspapers mistakenly linked their websites to DMV.org instead of a state DMV website.
Plaintiffs didn’t stop there, but also provided an internet survey showing that “a majority of California residents searching online for traffic schools believed that (1) DMV.org’s website was actually the California DMV’s and (2) a search engine listing for DMV.org was endorsed or sponsored by the California DMV.” There were significant flaws, including failure to use a control, but the district court found it more credible than defendants’ survey and gave it “some” weight, which was not reversible error.
Overall, there were “volumes” of evidence of competition and misleadingness. This established that the DMV.org site deceived a substantial segment of the audience, and plaintiffs also showed that a “recommended by DMV” endorsement will affect purchase decisions, causing likely injury to plaintiffs.
Remedy: the district court ordered DMV.org to present every site visitor with a splash screen stating, “YOU ARE ABOUT TO ENTER A PRIVATELY OWNED WEBSITE THAT IS NOT OWNED OR OPERATED BY ANY STATE GOVERNMENT AGENCY.” Visitors have to click to continue. Defendants argued that this was overbroad and restrained conduct not at issue in the complaint, and violated the First Amendment.
The district court reasoned that the splash screen was necessary to: (1) “remedy any confusion that consumers have already developed before visiting DMV.ORG for the first time,” (2) “remedy the public interest concerns associated with [confused visitors'] transfer of sensitive information to Defendants,” and (3) “prevent confusion among DMV.ORG’s consumers.” Defendants argued that the splash screen doesn’t do these things (I’ll have a bit to say about this shortly). Their only evidence was a declaration from DMV.org’s CEO stating that they tested several alternative disclaimers and “found them to be more effective than the splash screen in preventing consumers from emailing DMV.org with sensitive personal information.” Even crediting this self-serving declaration, defendants didn’t prove that the splash screen was ineffective, and said nothing about whether the alternative disclaimers served the other interests identified by the district court. Citing Home Box Office, Inc. v. Showtime/The Movie Channel Inc., 832 F.2d 1311 (2d Cir.1987), the court of appeals ruled that defendants hadn’t carried their “heavy burden” of showing that alternative disclaimers reduced likely confusion. The district court didn’t abuse its discretion when it concluded that the splash screen was the best way to correct the false advertising.
“Courts routinely grant permanent injunctions prohibiting deceptive advertising. Because false or misleading commercial statements aren’t constitutionally protected, such injunctions rarely raise First Amendment concerns.” However, the permanent injunction here did because it erected a barrier to all content on the website, not merely that which is deceptive. “Some of the website’s content is informational and thus fully protected, such as guides to applying for a driver’s license, buying insurance and beating traffic tickets. The informational content is commingled with truthful commercial speech, which is entitled to significant First Amendment protection. The district court was required to tailor the injunction so as to burden no more protected speech than necessary.”
The injunction was a permanent and unnecessary burden to accessing DMV.org’s First Amendment-protected content. “The splash screen forces potential visitors to take an additional navigational step, deterring some consumers from entering the website altogether,” and defendants submitted evidence that splash screens typically drive away up to a quarter of potential visitors, prevent them from tailoring the landing page, and interfere with search engine indexing and ranking.
The district court premised the injunction on its findings that defendants’ search engine marketing and natural listings, including the DMV.org domain name, caused confusion prior to viewing any content, and then identified specific misleading statements. “The splash screen is justified to remedy the harm caused by such practices so long as they continue,” and so long as it helps remedy lingering confusion, but not forever. On remand, the district court had to reconsider the duration of the splash screen “in light of any intervening changes in the website’s content and marketing practices, as well as the dissipation of the deception resulting from past practices.” If the district court continued the requirement, it needed to explain the continuing justification (comment: which would be the deceptive domain name) and what conditions defendants must satisfy in order to remove the splash screen in the future. The district court could also simply ban deceptive marketing or placing misleading statements on DMV.org.
My comment: I really don’t get this. It’s the domain name that’s at the core of the deception, something the district court pointed out—if the splash screen is justified “so long as [the deceptive practices]” continue, they will continue until there is a new domain name. Maybe if defendants only advertise the new domain name and simply redirect anyone who goes to dmv.org to the new domain, they should be allowed to remove the splash screen eventually. But for now, I don’t see why this isn’t a dilemma of defendants’ own deceptive making.
There is a potential bright spot, I suppose: the current splash screen is pretty bad (I would infer deliberately so). Take a look: Defendants have removed “Unofficial Guide to the DMV” from their license plate logo. The disclaimer runs at the very top, a couple of pixels away from the top of the browser window, in grey rather than black text (a nearly transparent tactic, pun intended!), not near the action invitation to click to continue, and not near the dmv.org logo, both of which are far more prominent to the eye. This placement clearly flunks the FTC’s guidelines for online disclosures, and defendants who already managed to mis-implement disclosures (of which more below) should not get the benefit of the doubt in this area. On remand, I sincerely hope that the revised remedy, whether splash page or not, forces the disclaimer into a prominent position.
The district court denied plaintiffs’ request for an award of profits because they provided no evidence of causation, nor did they quantify the harm they suffered. “Nothing in the Lanham Act conditions an award of profits on plaintiff’s proof of harm, and we’ve held that profits may be awarded in the absence of such proof,” though it’s an uncommon remedy without proof of harm. Profits are appropriate in false comparative advertising cases, “where it’s reasonable to presume that every dollar defendant makes has come directly out of plaintiff’s pocket.” Likewise “where ordinary damages won’t deter unlawful conduct: for example, when defendant associates its product with plaintiff’s noncompetitive product to appropriate good will or brand value.”
But neither line of cases were relevant here. Profits may be awarded only as compensation and not as a penalty, but without comparative advertising plaintiff’s injury may be a small fraction of defendants’ profits. There was no error in denying damages.
The potential attorneys’ fee award did require some rethinking, though. The district court denied fees because, while plaintiffs received an injunction, they didn’t get damages, and they had unclean hands. This was reviewed for abuse of discretion.
An exceptional Lanham Act case justifying fees usually involves conduct that was “fraudulent, deliberate, or willful.” “By examining only the relief awarded to plaintiffs, and failing to consider defendants’ conduct, the district court applied the wrong legal standard.” The district court may take failure to recover damages into account, but it should also consider that plaintiffs obtained a judgment and an injunction to ameliorate serious public harm caused by defendants’ unlawful conduct. “It would be inequitable to force plaintiffs to bear the entire cost of enjoining defendants’ willful deception when the injunction confers substantial benefits on the public.” Plaintiffs stopped the confusion (maybe) and stopped consumers from mistakenly transferring sensitive personal information to a commercial website. “The district court abused its discretion by failing to consider these substantial benefits or defendants’ bad acts in determining whether to award attorney’s fees.”
Defendants challenged the finding that their deception was willful. But there was plenty of evidence to support that, including evidence that one defendant had a pattern of registering misleading domain names. “Defendants associated their website with URLs and search terms that falsely implied DMV.org was a government site. They had in their possession hundreds of emails sent by consumers who contacted DMV.org thinking it was a state agency. And DMV.org’s director of customer service testified that he voiced concerns about these emails to senior management.”
Defendants responded that they added disclaimers, but they knew the disclaimers were ineffective, because that didn’t change the stream of emails from confused consumers. Given this knowing inaction, the district court could reasonably infer willfulness.
Moreover, the district court’s finding of unclean hands was clearly erroneous. This rested on two findings: (1) plaintiffs registered similar domain names, such as Online-DMV.org, Internet-DMV.org and cadmvtrafficschool.com; and (2) they attempted to advertise their products on DMV.org despite being aware that the site deceived the public. Neither was clear and convincing evidence of inequitable misconduct related to the subject matter of the false advertising claims. (In a footnote, Judge Kozinski commented that it wasn’t clear that the fee award was subject to equitable doctrines such as unclean hands, since the relevant statutory provision doesn’t use the term “equity.”)
Merely registering a domain name wasn’t proof of unclean hands, since registration couldn’t cause confusion unless the domain name was associated with a website visible to consumers, and plaintiffs’ ads on DMV.org ran for just six hours. There was no evidence of actual deception caused by plaintiffs’ advertising. There was an email from plaintiffs’ co-founder recommending taking an “[i]f you can’t join ‘em, shut ‘em down approach” to DMV.org. “But the doctrine is unclean hands, not impure thoughts. … And the email here actually undermines the rationale for finding unclean hands. Plaintiffs acquired information showing that defendants confused the public; using litigation to shut down a competitor who uses unfair trade practices is precisely what the Lanham Act seeks to encourage.” (I’m not sure that’s undermining, precisely.)
Remand for further proceedings.
This lawsuit is just the latest courtroom fracas involving privately operated websites that seem designed to create the appearance of official government sponsorship. Other sites in this genre include ConsumerAffairs.com and FreeCreditReport.com. As I indicated in my prior blog post on this case, the operators of DMV.org were probably too zealous about trumping up their officialness, thus destroying a promising domain name when a lighter touch would have left some short-term profits on the table but generated more profits over the long-run.
Rebecca highlights how this opinion has some important things to say about false advertising standing, consumer confusion in false advertising and attorneys’ fees in Lanham Act cases. I’m going to focus on more of the technical issues:
First, DMV.org bought keyword ads that displayed the URL “ca.dmv.org” and “california.dmv.org.” The opinion says “Defendants’ use of the “ca.” and “california.” prefixes obviously was designed to suggest an affiliation with the State of California.” As far as I can recall, this is the first time a court has explicitly treated the URL displayed in keyword ad copy as contributing to consumer confusion. I know some SEMs like to play games with the domain line in keyword ad copy. Even if Google won’t police the URL line, courts will.
Second, the opinion shreds DMV.org’s argument that consumers didn’t think it was affiliated with a government agency; it concludes that DMV.org “willfully” and “egregiously” deceived consumers because
Defendants associated their website with URLs and search terms that falsely implied DMV.org was a government site. They had in their possession hundreds of emails sent by consumers who contacted DMV.org thinking it was a state agency. And DMV.org’s director of customer service testified that he voiced concerns about these emails to senior management.
This deceptive conduct wasn’t curable with a website disclaimer: “defendants knew that the disclaimers were ineffective, because adding them didn’t end the stream of emails sent by consumers who thought they’d contacted their state DMV.” Marketing people love to “fix” any legal problems with disclaimers, but they often simultaneously resist making the disclaimers prominent enough to actually fix the problem, leaving the website no better off.
As usual, Judge Kozinski adopts a realpolitik perspective that we all know what’s going on here, and he isn’t going to countenance it. This discussion doesn’t directly link to or cite the scienter battles in online copyright and trademark cases, but it could be a little uncomfortable for defendants like Fung who rely on form-over-substance arguments.
Third, the district court ordered DMV.org to display a splash screen to incoming visitors containing a disclaimer of government affiliation. I believe the remedy of a mandatory splash screen was also unprecedented. Among other problems with the splash screen, it can hurt or destroy search engine indexing, effectively rendering the domain name worthless. Therefore, even though the district court didn’t award any damages (a ruling affirmed on appeal), the mandatory splash screen was potentially far more detrimental.
On appeal, the court partially reverses the mandatory splash screen, noting that it applied equally to deceptive and non-deceptive content that DMV.org is publishing. The court also notes the disadvantages of interfering with search engine indexing:
[The injunction] also precludes defendants from tailoring DMV.org’s landing page to make it welcoming to visitors, and interferes with the operation of search engines, making it more difficult for consumers to find the website and its protected content.
In a footnote, the court explains:
Defendants introduced unrebutted evidence that splash screens commonly interfere with the automated “spiders” that search engines deploy to “crawl” the Internet and compile the indexes of web pages they use to determine every page’s search ranking. And splash screens themselves don’t have high search rankings: Search engines commonly base these rankings on the web page’s content and the number of other pages linking to it, and splash screens lack both content and links.
This is a pretty savvy explanation of why a splash screen destroys SEO, and it’s heartening to see the court recognize the First Amendment value of SEO (at least, nondeceptive SEO).
The court gives DMV.org some very good news by giving it a chance to work its way out of the splash screen:
On remand, the district court shall reconsider the duration of the splash screen in light of any intervening changes in the website’s content and marketing practices, as well as the dissipation of the deception resulting from past practices. If the district court continues to require the splash screen, it shall explain the continuing justification for burdening the website’s protected content and what conditions defendants must satisfy in order to remove the splash screen in the future. In the alternative, or in addition, the court may permanently enjoin defendants from engaging in deceptive marketing or placing misleading statements on DMV.org.
On the other hand, as Rebecca points out, the DMV.org domain name is, itself, designed to prey on consumer interest in official DMV services. Perhaps there will be no way for DMV.org to fix this, no matter what disclaimers it uses.
One final note: this opinion was authored by Judge Kozinski, the third domain name opinion he has written in the last 13 months. (The other two are the Tabari and eVisa cases). I went back and counted a total of 9 Ninth Circuit domain name cases in that time–the others being Advertise.com, Christensen (a 2 paragraph opinion), DSPT, Balsam, Office Depot (a non-substantive affirmance) and Vericheck. Kozinski was not on the other panels, so he wrote the opinion in each of the 3 domain name cases he heard. Is all of this just random luck?
UPDATE: Something about this opinion is spurring blog posts focusing on its arcania. Tom O’Toole notes this is the second opinion using the term “googler.” Shaun Martin focuses on Kozinski’s self-citations.