March 31, 2006
Keyword Purchases Not a Trademark Use--Merck v. Mediplan Health Consulting
By Eric Goldman
Merck & Co. v. Mediplan Health Consulting, 2006 WL 800756 (SDNY Mar. 30, 2006)
About a week after Edina Realty v. TheMLSonline allowed a keyword purchase of a competitor's trademark to go to trial, we got another ruling about trademark liability for purchasing keywords at the search engines. But this case reached a different result--the court dismissed the trademark infringement claim for the keyword purchases because the purchases do not constitute a trademark "use in commerce." This sets up a direct conflict with the Edina Realty case that will require future cases to clarify the law.
This lawsuit involves various Canadian Internet pharmacies targeting US customers. Most of these pharmacies sell Merck's Zocor (a cholesterol-reducing drug) as well as a generic version. Merck sued for patent infringement, trademark infringement and dilution, and false advertising. In this post, I'm focusing just on the discussion about the pharmacies' purchase of the keyword Zocor at the search engines. The defendants moved to dismiss the trademark infringement claim for these purchases because they did not constitute trademark use in commerce.
Recall that in the Edina Realty case, the court said "Based on the plain meaning of the Lanham Act, the purchase of search terms is a use in commerce." In my previous post, I explained why the court cut corners to make this statement.
Fortunately, the Merck court correctly identifies the applicable statutory definition of trademark use in commerce (15 USC 1127(1)-(2)). This section effectively requires that consumers "see" the trademark usage for it to qualify. The Merck court, citing to the Second Circuit 1-800 Contacts ruling from last summer (as well as the other 2 district court cases involving WhenU), says that keyword purchases are invisible to consumers and therefore do not constitute a trademark use in commerce.
As a result, the defendants are able to dispose of the claims over the keyword purchases on a motion to dismiss. This is a big win for defendants generally (not necessarily for these defendants, who face other legal problems) because knocking out the case on a motion to dismiss is comparatively quick and cheap.
So this case is good news for search engine advertisers (and, as a result, search engines). However, because this ruling directly conflicts with the Edina Realty ruling, we're in legal limbo. We need more courts to weigh in before this issue becomes predictable.
Thanks to Rebecca Tushnet for spotting the case. Her commentary.
March 30, 2006
Search Engine Bias Article
By Eric Goldman
Search engines ranking algorithms create winners and losers, and some people don't like this system. Obviously the losers don't like it (see the KinderStart v. Google lawsuit as an example). More importantly, commentators with normative views about what constitutes "good" search results routinely advocate regulating the search engine ranking process in a way that advances their normative objectives.
I've now posted my latest thoughts on this topic in an article entitled "Search Engine Bias and the Demise of Search Engine Utopianism." It's coming out later this spring in the Yale Journal of Law and Technology. Regular blog readers won't be surprised that I come out swinging against the pro-regulatory forces who think they can do a better job shaping search results than search engines can do without regulatory "help." I welcome your comments.
"Due to search engines’ automated operations, people often assume that search engines display search results neutrally and without bias. However, this perception is mistaken. Like any other media company, search engines affirmatively control their users’ experiences, which has the consequence of skewing search results (a phenomenon called “search engine bias”). Some commentators believe that search engine bias is a defect requiring legislative correction. Instead, this Essay argues that search engine bias is the beneficial consequence of search engines optimizing content for their users. The Essay further argues that the most problematic aspect of search engine bias, the “winner-take-all” effect caused by top placement in search results, will be mooted by emerging personalized search technology."
March 29, 2006
Competitor's Keyword Ad Purchase May Be Trademark Infringement--Edina Realty v. TheMLSonline
By Eric Goldman
Edina Realty, Inc. v. TheMLSonline.com, 2006 WL 737064 (D. Minn. Mar. 20, 2006)
There have been many lawsuits involving keyword advertising, but in most of the reported decisions, the search engines were the defendants. There have been only a few wispy reported decisions where the keyword ad purchaser was the defendant, and I don't recall any substantive discussion of trademark law in those cases.
That's what makes this case significant. I think this is the first case substantively analyzing a purchaser's liability for buying a competitor's keyword.
This case involves two real estate companies. Edina Realty is a major old-line Minnesota real estate agency, and TheMLSonline.com appears to be an online realty agency with some full-service aspects. In that respect, the firms are competitive, although TheMLSonline.com also offers a content database of MLS realty listings from all sources, including competitor agencies.
To build traffic, TheMLSonline.com purchased keywords at Google and Yahoo such as “Edina Realty,” “Edina Reality,” “EdinaReality.com,” “EdinaRealty,” “EdinaRealty.com,” “www.EdinaReality.com” and “www.EdinaRealty.com.” (Interesting that people appear not to know how to spell “realty.”) The resulting ads contained the words “Edina Realty,” such as:
"Edina Realty™ listings--MLS Home Search
Find Edina Realty™ and Twin Cities MN MLS listings. Also includes listings from Coldwell Banker Burnet, Counselor Realty and many more. Updated daily. Request showings online
(Not surprisingly, pursuant to Google's trademark policy, Edina Realty eventually blocked text references to Edina Realty in the ad copy on Google. Under Yahoo's new trademark policy, I assume that Edina Realty now can block these purchases at Yahoo outright.)
According to the opinion, these ads were extremely successful; the court's opinion suggests that the ads produced clickthrough rates of 18-40%. (I'm skeptical that this is right; the court's language here is sufficiently vague that the court may have been trying to say something else).
Finally, TheMLSonline.com included white text on a white background ("hidden text") with phrases like "Edina Realty information presented at TheMLSonline.com."
Edina Realty sued for trademark infringement, dilution and false advertising. In this opinion, the court addressed the parties' cross-motions for summary judgment on the infringement and dilution claims.
Use in Commerce
The court (correctly) considers "use in commerce" as a threshold matter. This is noteworthy because so many other Internet trademark cases have simply ignored the TM use in commerce requirement. However, the court's discussion leaves much to be desired. The court's entire substantive discussion:
"While not a conventional "use in commerce," defendant nevertheless uses the Edina Realty mark commercially. Defendant purchases search terms that include the Edina Realty mark to generate its sponsored link advertisement. See Brookfield Communs., Inc. v. W. Coast Entm't Corp., 174 F.3d 1036, 1064 (9th Cir.1999) (finding Internet metatags to be a use in commerce). Based on the plain meaning of the Lanham Act, the purchase of search terms is a use in commerce.
Really?! The court doesn't cite to several precedent on point, such as the first Geico ruling (finding Google made a use in commerce by selling keywords) and the 1-800 Contacts 2nd circuit opinion (finding that WhenU didn't make a use in commerce by using keywords to trigger ads). Nor does the court acknowledge that the precedents are arguably contradictory.
The court also does not discuss how the federal statutory language is arguably unclear and contradictory. The Lanham Act makes two different references to "use in commerce," one arguably establishing federal jurisdiction and the other arguably requiring use on product packaging. There have been several law review articles discussing this very point (such as Uli Widmaier's and Margreth Barrett's, as well as my own) that have argued that the statute should be interpreted directly contrary to this ruling. At best, the court cut corners by saying that its conclusion is based on the "plain meaning" of the statute.
Likelihood of Confusion
Having found use in commerce, the court then moves to the standard likelihood of consumer confusion test. The plaintiff argued to cut short the LOC analysis because the various uses should constitute initial interest confusion, while the defendant argued that the court should cut short the LOC analysis because the uses constitute nominative fair use.
The court rejects both arguments. Instead, the court briefly evaluates the standard LOC factors and finds enough dispute about several factors to require a trial.
However, the court rejects the defendant's claim of nominative fair use as a matter of law, saying that the "Defendant could have done more to prevent an improper inference regarding the relationship." While this may be true, this conclusion ignores important public policy issues here. The defendant offered a version of the MLS database to the public, and it let potential searchers know about the existence of the database. In this capacity, the defendant wasn't a competitor; it was a provider of a valuable information resource. Limiting the marketing of that database service to searchers does a disservice to the competitive market generally.
The court grants summary judgment to the defendant on the dilution claim because the plaintiff didn't marshal enough evidence of actual dilution. This ruling presumably will be different after enactment of the Trademark Dilution Revision Act.
The plaintiff hasn't won its case yet. At trial, the court could find that there was no likelihood of consumer confusion at trial, similar to how the Geico v. Google court found that there was a TM use in commerce but no likelihood of consumer confusion. And undoubtedly this case will not be the final opinion on the matter.
However, until we get more data points (such as new cases or a defense win at trial in this case), I think this ruling potentially has the following two long-term implications:
1) This ruling puts increased pressure on Google's policy not to block competitor keyword ad purchases. Now that Yahoo blocks these purchases, Google effectively stands alone in the industry, so it lacks any cover provided by prevailing industry standards. More importantly, Google's position has just become legally riskier. To the extent that competitors' ad purchases constitute direct trademark infringement, Google may face an elevated risk of being deemed a contributory infringer.
2) This case offers the first solid data point (in the US) that buying competitors' trademarks as keywords, without more, could constitute trademark infringement. This could permanently reduce the market size for keyword ad sales. Some competitive keyword ad buys currently being made will be eliminated because the competitors (or search engines) decide those purchases are too risky. Stock valuations of relevant industry players should adjust downward accordingly.
UPDATE: Although this case sets new precedent in the US, there have been a fair number of analogous rulings internationally. Cedric recounts some of the action in France (partially in French).
UPDATE 2: New case, different result. See my post on Merck v. Mediplan Health Consulting, saying that keyword ad buys aren't a trademark use in commerce.
UPDATE 3: Brian Quinton wrote an article recapping the topic. In the article, he quotes Gary Angel of SEMphonic, who works in SEM. Gary explains why "using another company’s name as a search term was a sensible strategy whose time has gone," including:
* Google raised the minimum bid price on keywords across-the-board, and the minimum price now exceeds a price that gets good ROI
* click-throughs are highly qualified but low in volume because, as Gary says, "When searchers search for the company name or brand of a competitor, they’re highly inclined to click through on that competitor...That makes it very difficult to get large returns on that kind of targeted attack on your competitors."
Two observations from this. (1) Gary's quote reinforces that consumers generally will follow the trademark owner if they search on the trademark if that's what the consumer wanted. (2) Given the low margins and volume, I continue to be befuddled why trademark plaintiffs are wasting their dollars on litigation.
March 28, 2006
Animal-Related Patent Trends
By Eric Goldman
Wired runs an article on Patently Silly, one of the blogs tracking goofy patents. The article notes the implicit contradiction in the trends for patenting pet pampering products and patenting new and more effective ways of producing meat. Basically, inventors are working hard both to make pets lives more cush and make it more profitable to chop up livestock for food. As the article quotes:
For more rants about our inconsistent attitudes towards animals and about vegetarianism generally, see my vegetarian category page on my other blog.
March 24, 2006
Yahoo Loses 230 Defense for its Dating Site--Anthony v. Yahoo
By Eric Goldman
Anthony v. Yahoo! Inc., 2006 WL 708572 (N.D. Cal. March 17, 2006)
Anthony subscribed to Yahoo's dating subscription services. He claims that Yahoo creates false personal profiles and retains expired profiles to make the subscription service look more attractive to existing subscribers. He is the named plaintiff in this putative class action, claiming (among other things) that, based on this behavior, Yahoo engages false and negligent misrepresentation and deceptive/unfair trade practices.
Yahoo moves to dismiss those claims under 47 USC 230. The court rejects Yahoo's motions for two reasons. First, the complaint alleges that Yahoo creates the false profiles, so Yahoo (and not a third party) is the information content provider of those profiles.
Second, Yahoo sends subscribers the expired profiles. Though the profiles are created by its users, Yahoo (not the users) create the false perception that the profiles are valid. Thus, Yahoo can't claim the 230 immunity for creating that impression.
Normally, I never meet a 230 defense that I don't like. So it may surprise you that I think the court rightly rejected the 230 claim. Certainly, 230 on its face does not apply to the allegation that Yahoo created profiles itself. Note to plaintiffs: you should always be able to survive a 230 motion to dismiss by claiming that the online service provider itself solely created the content in question. (See the analogous ruling in the Hy Cite case). Of course, these allegations are subject to Rule 11, so plaintiffs need to be able to support these factual assertions.
The liability for false impressions about expired profiles presents a closer question. I can see some room for mischief if plaintiffs can claim that they are suing the service provider based on some false metadata where the underlying data was user-generated and covered by 230. However, if Yahoo falsely claims that profiles are valid when it knows they are expired, Yahoo--and not its users--is taking the legally significant act and should bear the consequences accordingly.
As usual, this ruling on the motion to dismiss doesn't represent the end of Yahoo's defenses. Therefore, we can't make any inferences about the plaintiffs' ultimate likelihood of success.
March 23, 2006
By Eric Goldman
Gratis Internet runs several websites that promise free stuff (like free iPods) in exchange for consumers signing up for subscription trials. The trials are initially free but then convert to paid subscriptions. The idea is that many consumers will either like the subscriptions or be duped into keeping the subscriptions against their will. For an example of how even very intelligent people can be trapped by these free trials, see my colleague Christine's story (and the update).
Along the way, Gratis made a variety of privacy promises to consumers. Of specific relevance here, Gratis promised that it would never resell the consumers' email addresses. However, as it turns out, Gratis allegedly may have done precisely that.
More interesting to me is Spitzer's action against Datran Media, one of the buyers of email addresses from Gratis. Last week, Spitzer's office announced a settlement with Datran that included a $1.1 million check.
But I think there's a more fundamental lesson to learn. This case reinforces that it's very hard to legitimately buy/sell email addresses. At minimum, I think buyers need to do thorough diligence of the email addresses' origins, and it's hard to find legitimate email addresses that were completely acquired without restriction on transfer or resale. Then, under CAN-SPAM, the email addresses have to be filtered out for any opt-outs that the buyer has received in the past. And then, it's hard to get bulk emails through the email service providers/IAPs, especially if the sender can't claim some type of relationship with or authorization from the recipients.
All told, I just don't understand how legitimate companies think that email addresses can be flipped like commodities. The practice may never have been legitimate, but I see it as a completely dead practice today.
UPDATE: Dan Solove weighs in on the case. I generally agree with Dan's analysis, except that I think we need to know more about Datran's scienter. This result is defensible only if the scienter level was high enough.
UPDATE 2: Chris Hoofnagle calls the case "one of the biggest cases for consumer privacy ever."
March 22, 2006
CDT Report on Adware Advertising
By Eric Goldman
The Center for Democracy and Technology has released "Following the Money: How Advertising Dollars Encourage Nuisance and Harmful Adware and What Can be Done to Reverse the Trend."
The report details the complex web of relationships between advertisers, agencies, adware vendors and software vendors that can lead to big brand advertisers having their ads unwittingly delivered via adware. This isn't really news; Ben Edelman (and others) have demonstrated this for some time, and last year PCWorld ran a good article on this very topic.
The real question is: what to do about it? FTC Commissioner Leibowitz's solution is to shame the advertisers. The plaintiff lawyers are trying to hold the advertisers legally responsible. In its report, CDT offers an alternative: advertisers should adopt and enforce advertising placement policies. Again, this isn't really new; TRUSTe has launched a "trusted download program" that is designed to serve as a proxy for advertisers' policies.
The most interesting part of the report is where CDT describes how it contacted 18 advertisers who ran ads via 180solutions. Only 7 replied to CDT. 2 advertisers adopted advertising policies in response to the contact, but the other 5 had policies that were breached by running ads on 180solutions.
Personally, I don't find this very surprising. It is expensive for companies to monitor vendor behavior/contract compliance in all contexts. Let me offer a personal example.
When I was in-house counsel, I had effectively zero ability to monitor our vendors' compliance with our contract provisions. I could write the prettiest contract with the terms that we really wanted, but the contract goes into someone's desk drawer and the terms are effectively forgotten. To overcome this, we would have needed to make monitoring part of someone's job, but that is a very expensive solution (especially when we were a very small company with too much to do and not enough people to do it).
At the same time, when we were someone else's vendor, I built and disseminated various charts to outline our promises/responsibilities. However, ensuring my co-workers' compliance with those responsibilities was difficult/impossible. One of my worst realizations as in-house counsel is that I needed to clone myself and look over the shoulders of each and every employee of our company, because each person's decisions affected our compliance with our contractual obligations. Clearly, this isn't scalable, and my efforts to scale using employee education did little to improve the situation.
The lesson I learned, then, is that it's easy to put policies in place, but ensuring adherence to those policies by employees and third party vendors is hard--perhaps impossible. And there was absolutely no hope of policing third party behavior unless it was someone's job responsibility.
In the adware advertising context, most advertisers are not going to be willing to incur significant costs to police their vendors. They'd rather pour those dollars into actual advertising, and other media don't require such costs. So smart advertisers would either (a) redirect their ad dollars into enforcement-free media, or (b) cut corners and minimize/skip the monitoring process. We've already seen what most advertisers choose under the current model.
This brings me to my most fundamental confusion about adware advertising. Every advertising-supported media vendor has the capacity to break the law, and yet we don't expect advertisers to be the policemen of those media. On the contrary, we often want advertisers to subsidize the costs of producing editorial content for the positive externalities created by that content. (See Edwin Baker's Advertising and a Democratic Press.)
So what's so unique about adware that we are creating a form of cyberspace exceptionalism? Some adware is legitimately installed and some isn't; expecting adware advertisers to incur the monitoring/enforcement costs to figure out which is which (especially when there's no direct relationship) makes no more sense to me than expecting newspaper advertisers to determine if newspapers commit trespass when they deliver their newspapers. For more on this point, see my CNET editorial.
March 21, 2006
Google Sued Over Rankings--KinderStart.com v. Google
By Eric Goldman
KinderStart.com LLC v. Google, Inc., No. C 06-2057 RS (N.D. Cal. complaint filed March 17, 2006)
Google has been sued for downgrading the PageRank of websites in contravention of its stated "objective" policies. In KinderStart's case, they got kicked out of Google in March 2005 and immediately lost 70% of their traffic. Google is now 0.01% of KinderStart's referral traffic.
As a result, KinderStart, and similarly situated plaintiffs, are trying to form a class action lawsuit to get money from Google for the PageRank downgrades. The complaint alleges seven causes of action:
Free speech violations under the US and California Constitutions.
This claim is a loser on its face because Google is not a state actor. See the cites in my previous post about why ICANN isn't a state actor. The complaint doesn't even try to allege specifically that Google is a state actor. It does claim that Google is an "essential facility," but this argument has no hope. It has been soundly rejected in the context of domain names and email, and search results are no more essential than those.
While California's free speech protections are broader than the US Constitution's, I don't think California's protections extend anywhere close enough to cover Google's actions.
Sherman Act Section 2 (antitrust/monopolization).
This is also a loser claim. I don't see how the plaintiffs can establish that Google has a monopoly position. If anything, the search engine market remains dynamic and competitive with several large players, as the statistics clearly illustrate.
California Business and Professions Code Sec. 17200
I'm not an expert in this law, but this claim looks less frivolous than the previous two claims. In particular, the complaint marshals some non-trivial evidence that Google has falsely described its ranking algorithm operations. For example, the complaint (particularly paragraphs 37-38) describes some disclosures in Google's FAQs that say Google doesn't censor or manipulate search results, but the facts suggest that Google in fact does so. (For example, Google just recently had to change its censorship FAQ after launching its China service).
(To be clear, Google is hardly alone. Indeed, just yesterday it was reported that Amazon manually hacks the search results for "abortion." Algorithmically, this search produces a prompt "did you mean adoption?," and Amazon manually deleted that prompt).
In this respect, Google has a duality: Many people at Google still think that search engine operations are as automated and pure as they were in 1999. However, in the intervening 7 years, Google has made lots of manual changes and hacks that run contrary to its claim of automation and objectivity. I document some of these in my Deregulating Relevancy article, and I will make a more organized presentation of this claim in my imminently forthcoming article on search engine bias. Google may ultimately pay for its delay in realizing that it is a media company and makes editorial choices just like any other media company.
To the extent the 17200 claim is based on false advertising, Google also won't be able to claim a first amendment defense like it did in the Search King case. In that case, Google claimed that its PageRank decisions were constitutionally protected opinions. If the 17200 claim is based on Google's false descriptions of its ranking algorithms, then the claim may be beyond constitutional insulation.
Other Causes of Action
* Unfair Practices under Cal. B&P 17040. This is an anti-price discrimination statute, although it's written more broadly to cover other types of commercial discriminations. This is an interesting claim and I don't know this statute well. I suspect that courts will be very reluctant to apply this statute to media companies' publication decisions; there may even be constitutional limits on doing so.
* Breach of Implied Covenant of Good Faith/Fair Dealing. This is a bizarre claim. It claims that AdSense participants allocate scarce web page space to AdSense ads, and Google breaches its obligations by reducing traffic that might see these ads. But think about this--the claim is that Google breaches a duty to web publishers by refusing to pay the publishers for traffic that Google itself sourced for the publishers! I will be very surprised if a court believes that Google refusing to pay for its own traffic is "bad faith." I also think that Google's express contract terms may trump any applicable implied covenant.
* Defamation. The complaint claims that Google defames a website by giving it a zero PageRank when the mathematical algorithms should have given it a higher number. This claim is also a loser for several reasons. First, as the Search King case found, PageRank should be a protected opinion. Second, there's no actual defamation because no one knows exactly how PageRank is computed, so there's no way for a third party to draw any inferences about what a zero PageRank means. (However, Google makes its life more difficult by claiming in the toolbar that PageRank is a characterization of the website's "importance;" Google should tone down this description of what PageRank means). Third, because Google can change its ranking algorithm at any time, there's nothing to prevent Google from assigning a zero PageRank on any basis it chooses.
* Negligent Interference with Prospective Economic Advantage. This is also based on a claim that Google owes traffic to AdSense participants. Among other reasons why this claim will fail, I think Google's express contract terms should negate any such expectations.
As you can see, I think this lawsuit has low merit; most of its claims will be knocked out on a motion to dismiss. However, the 17200 claim could survive because of Google's sloppiness in its marketing disclosures, but I'm still not convinced that the 17200 claim is meritorious.
Personally, I find this lawsuit fascinating from a theoretical perspective. At its core, the lawsuit presumes that there is a single way Google should run its ranking algorithm, and Google should suffer penalties for any deviations. Intrinsically, this lawsuit is about diminishing Google's ability to run its algorithm as it sees fit (including mismanaging the algorithm).
This pro-regulatory argument is more common than we might think. Rep. Smith has effectively advanced such an argument in proposing the Global Online Freedom Act to regulate ranking algorithms. And plenty of academics have jumped on the pro-regulatory bandwagon. The seminal pro-regulation article is by Introna and Nissenbaum, but they are hardly alone. Just recently, Frank Pasquale has posted a work-in-progress, "Rankings, Reductionism, and Responsibility," arguing for various controls over ranking algorithm decisions.
The arguments in favor of regulating search engine ranking algorithms are wrong, wrong, WRONG. I'll make a more complete case in my search engine bias article, but for now, consider the argument as applied in the non-Internet context of newspapers. Newspapers have absolute freedom to select which articles to publish and what order to present them in. Not only is this the basic value proposition of newspapers (i.e., why we buy/read newspapers), but newspaper editorial decisions are constitutionally protected (see Miami Herald v. Tornillo) and cannot be overridden.
So why should search engines be treated differently? Search engines have to tell the truth (i.e., if they claim not to manually hack search results, they can't do a manual hack), but in all other respects, they should have the same freedoms as other publishers. If you can't imagine legally requiring newspapers to reorder their articles, then it should strike you as equally bizarre that some argue that search engines should reorder results or be liable for their ordering choices.
As for KinderStart, there is a lesson to learn: it's precarious to build a business on free search engine traffic. My advice is to enjoy the ride so long as it lasts, but don't plan on the good times lasting forever, and definitely don't expect a ton of sympathy when the gravy train ends.
March 12, 2006
Do-Not-Contact Registries Proliferating?
By Eric Goldman
There has been a groundswell of action at the state level to implement do-not-contact registries of various kinds. Two of particular note:
1) Don't-spam-my-kid Registries. Two states (Utah and Michigan) already have laws enabling do-not-spam registries for kids' email addresses. Now, according to BNA (registration required), six more states are considering hopping on the bandwagon (including Wisconsin)--despite the FTC's public announcement that it thinks these registries are a bad idea, plus the pending lawsuit against Utah's registry.
I have yet to see how the implementation authenticates that the registered email address is actually used by a kid. Without that authentication, these registries have the potential to be backdoor across-the-board do-not-email registries. If so, draping the do-not-email registries in the "protect the kids" flag inhibits robust discussion about the real effects of the registry.
2) Don't-junk-mail-me Registries. At least 3 states (Missouri, New York and Illinois) have introduced laws to create do-not-mail registries analogous to the very popular do-not-call registries. For reasons that I've not entirely understood, the delivery of junk mail is lightly-regulated compared to telemarketing and email marketing. Perhaps that is ready to change. In my Coasean Analysis of Marketing paper, I will point out the policy deficiencies of "do-not-contact" registries generally, so I find the possible proliferation of such registries troubling.
March 11, 2006
FTC Extends COPPA Without Changes...and New FTC RSS Feeds
By Eric Goldman
The FTC has extended the COPPA rule unchanged. Most significantly, the rule continues to preclude non-authenticated email as a way of obtaining parental consent.
I don't spend a lot of time thinking about COPPA much any more. I teach my Cyberlaw class that they should advise clients to avoid being governed by COPPA at all costs. COPPA makes it expensive to provide online interactivity tools to kids; but by definition, kids don't have any online purchasing power. So it's hard to profit from providing robust online tools to kids.
Having said that, I am constantly surprised by the number of websites I see that should be COPPA-compliant but don't appear to make any effort to comply with it. I think the FTC could find plenty of targets if it decided to sweep for COPPA violations.
On a separate note: While checking out the FTC's website, I discovered that the FTC quietly has launched RSS feeds. Terrific news!
March 10, 2006
AP on Internet Porn Regulation
By Eric Goldman
Associated Press runs a retrospective/status report on Congress' battle against Internet porn. Believe it or not, this battle recently entered its second decade--Congress' first effort, the Communications Decency Act, was passed in February 1996. After that law was struck down, Congress passed the Child Online Protection Act in 1998, a law that has already resulted in two US Supreme Court opinions, led to a major flap between Google and the DOJ, and is now scheduled for trial in October. As the detritus of Congress' efforts spills out to new fronts, I can't fathom just how much legal work has been spurred by Congress' fixation on Internet porn.
Meanwhile, as the article points out, everyone has moved past children's access to online porn to even more problematic issues like online sexual predators, adware/spyware concerns and copyright liability for P2P file sharing. As a result, if the trial court upholds the law, there will be massive shock and panic as the Internet industry contemplates how to comply with a law that virtually everyone has been ignoring for 8 years.
ART Act Indictment for Distribution of Pre-Release Music
By Eric Goldman
Robert Thomas of Milwaukee and Jared Bowser of Jacksonville have been indicted under the ART Act for pre-releasing four songs by Ryan Adams in August 2005, a month before the commercial release of Adams' album. Allegedly, the songs originally came from a legitimate pre-release copy sent to the media. This may be the first prosecution under the pre-release restrictions under the ART Act (there has also been at least one indictment under the camcordering provision of the ART Act).
March 09, 2006
Smoking Gun on Movie Budgets
By Eric Goldman
The Smoking Gun runs an interesting feature on movie budgets. Among the goodies include the entire line-by-line budget for the movie "The Village," and various excerpts from other budgets of movies by M. Night Shyamalan. Very interesting reading if you're interested in Hollywood accounting...or if you're curious about movie star paydays and perks.
March 08, 2006
Lane's Gifts Click Fraud Lawsuit Near Settlement
By Eric Goldman
Google announced that it is near a settlement in the Lane's Gifts class action lawsuit over click fraud. The way I read the announcement, the parties have agreed to the terms, so the settlement just needs the judge's approval.
As for the substantive terms, Google will allow any advertiser to claim credit for click fraud, an expansion of its normal policy that such credit requests must be made in 60 days. Google then will give "makegoods" to advertisers who experienced click fraud. Google's total exposure (makegoods plus attorneys' fees) is $90M.
From my perspective, this sounds like an outstanding outcome for Google. Because this was a class action lawsuit, the settlement will bind all advertisers (except those advertisers who opt-out, which I doubt many will do). So with the settlement, Google effectively eliminates its entire legacy liability for click fraud for $90M, only a portion of which is out-of-pocket cash. Given that the click fraud has been estimated in the billions, a $90M settlement sounds like an excellent deal.
Unfortunately, the settlement doesn't appear to resolve the basic question of what constitutes click fraud and when search engines are on the hook for it. On that front, Google still will have an advertiser relations issue that needs further attention.
I've repeatedly blogged about click fraud issues.
UPDATE: Danny Sullivan has done some excellent investigative work, including interviews with both plaintiff's counsel and Google's in-house counsel.
UPDATE 2: John Battelle weighs in, saying that the "settlement is a major victory for Google....I think the folks at Google are pleased as punch with the deal."
UPDATE 3: Google has posted this FAQ on click fraud. It says, "Some invalid clicks do make it through our filters, but we believe the amount is very small." The post also deconstructs the report claiming that click fraud is 30% of clicks.
UPDATE 4: The AIT lawyers say they are going to continue pressing their litigation. We need to see the settlement terms before we can fully understand the interaction between the two suits. I know it's obvious, but I should also point out that the judge might not approve the settlement, so the case isn't done until it's done.
Pasquale on Copyright Externalities
By Eric Goldman
Frank Pasquale has posted an article called The Law and Economics of Information Overload Externalities to SSRN. I'm not sure the title really captures the article, so don't make the decision to read/not read it based on the title.
The article reminds us that the world cannot be simply divided into copyright producers and copyright producers. At minimum, there's a third category of copyright intermediaries (let call them categorizers) that match producers with consumers. Search engines are the most prominent example of categorizers, and Google Book Search is the paradigmatic matching dilemma.
Accordingly to Frank, categorizers play an essential role in society because copyright's production incentive cntributes to information overload, which hinders access to all copyrighted works. However, categorizers invariably make copies in the categorization process, so copyright law can prevent categorizers from providing their social benefits. Accordingly, Frank argues that copyright law should take the social benefits of categorization into account.
"Environmental laws are designed to reduce negative externalities (such as pollution) that harm the natural environment. Copyright law should adjust the rights of content creators in order to compensate for the ways they reduce the usefulness of the information environment as a whole. Every new work created contributes to the store of expression, but also makes it more difficult to find whatever work one wants. Such search costs have been well-documented in information economics. Copyright law should take information overload externalities like search costs into account in its treatment of alleged copyright infringers whose work merely attempts to index, organize, categorize, or review works by providing small samples of them. They are not free riding off the labor of copyrightholders, but rather are creating the types of navigational tools and filters that help consumers make sense of the ocean of expression copyrightholders have created.
By modeling information overload as an externality imposed by copyrighted works, this article attempts to provide a new economic justification for more favorable legal treatment of categorizers, indexers, and reviewers. Information overload is an unintended negative consequence of copyright law’s success in incentivizing the production and distribution of expression. If courts grant content owners the right to veto categorizers’ efforts to make sense of given fields of expression, they will only exacerbate the problem. Designed to promote the progress of the arts and sciences, copyright doctrine should privilege the efforts of those who make that progress accessible and understandable. Categorizers fill both those vital roles."
UPDATE: Brett Frischmann has some interesting comments.
March 06, 2006
Congress Is Lovin' the Internet...to Death?
By Eric Goldman
Congress has an unresolved love-hate attitude towards the Internet. Through the 1990s, Congress frequently said that the Internet should be left alone from a regulatory standpoint. Indeed, in some cases, Congress affirmatively deregulated the Internet; 47 USC 230 and the Internet Tax Freedom Act come to mind.
However, Congress is irresistibly drawn to Internet regulation. Every Congressional session, members of Congress propose literally hundreds of laws to regulate some aspect of the Internet. Obviously, not all of these laws pass, but the sheer volume is evidence of the seductive lure of Internet regulation. Congress just can’t control itself!
I was working through the piles on my desk yesterday and I came across three recently proposed laws that demonstrate this irresistibility. All three laws reflect legislative opportunism to capitalize on hot media issues; all three laws reflect a certain idealism for how markets should function; and all three laws would have radical (and possibly crippling) effects on the Internet.
1) Eliminate Warehousing of Consumer Internet Data Act of 2006, HR 4731 (Introduced Feb. 8 by Rep. Markey).
Rep. Markey promised this law in response to the DOJ-Google flap. The premise is simple enough: online companies should flush their databases of personal data so the DOJ can't abuse its power to get that data. This animating principle translates into the following operative provision:
"An owner of an Internet website shall destroy, within a reasonable period of time, any data containing personal information if the information is no longer necessary for the purpose for which it was collected or any other legitimate business purpose, or there are no pending requests or orders for access to such information pursuant to a court order." The definition of personal information is suitably broad--first/last name qualify, as does an email address.
I don’t like this law’s expansive sweep. It would govern many seemingly-unimportant websites, such as my blogs (which allow users to submit both their first/last name and their email addresses). In some cases (like mine), I can’t flush personal data because it’s in the hands of my service providers.
Further, ironically this law doesn't even correct the DOJ-Google situation. First, the data requested by the DOJ wasn't personal data as defined by the law. Second, and more importantly, Google arguably has a legitimate business purpose to keep every scrap of data it ever lays its hands on (after all, how can you organize the world's information if you have to flush some of it down the drain?). Given that many businesses can claim a continuing benefit from keeping personal data, this law won't get that data flushed. Instead, I think it merely creates weird/unexpected technical headaches.
2) Internet Non-Discrimination Act of 2006, S. 2360 (Introduced March 2, 2006 by Sen. Wyden)
This law follows on the hot topic of net neutrality, or a “two-tier” Internet, which is also linked to AOL’s implementation of Goodmail’s certified email program. The law’s basic premise is simple: data transit vendors should not discriminate between bits—each bit should get processed equally. This gets codified in a list of restrictions about the products/services that a covered entity can (or can’t) offer.
I'm dubious about the theoretical underpinnings of this law, but for now my objection to the law is far more tactical. The law restricts the behavior of "network operators," which is anyone who "provides communications directly to a subscriber." I think the law was intended to govern the provision of Internet access/connectivity. But, as drafted, I think the law covers everyone who moves data from one point to another--this should include every website that provides user-to-user communication, including email service providers, instant message providers, blog providers, "email this page to a friend" providers, etc., etc. In other words, virtually the entire Internet.
This drafting error is, in theory, fixable. The law could just define the covered entities as Internet access providers more carefully. However, I don't think this is an easy fix. I think there is no clear distinctions between the various "layers" (content v. application v. transport); at least, the distinctions aren't definable statutorily.
Worse, it significantly restricts beneficial intermediary behavior, such as blocking incoming spam. The law acknowledges this consequence and says that the governed entities can block spam if the consumers are notified and have a chance to disable the application. So whereas AOL might kill almost all incoming spam at the server level, the law would take this choice out of AOL’s hands. I’m not sure what consequences result from that, but my heart tells me it’s expensive for AOL/the consumer and it could lead to weird and unexpected results.
In effect, this law would place most of the Internet under the oversight of an administrative agency (the FCC). The Internet had thrived without FCC oversight for a while now. I’m having a hard time believing that turning the Internet into a comprehensively-regulated industry would be a good thing.
3) Global Online Freedom Act of 2006, HR 4780 (Introduced Feb. 16, 2006 by Rep. Smith).
This law builds off the Google/Yahoo-China flap. It has various proposals designed to get China and other repressive countries to stop censoring Internet content.
Specifically, the law would create a new administrative agency called the "Office of Global Internet Freedom." This title alone is disconcerting; it sounds like something out of Orwell or Kafka. Indeed, like any good dystopian view of bureaucracy, the OGIF would free the Internet by telling its citizens what they can't do.
In this case, the OGIF would help generate a list of bad censorship-loving countries. On the initial list are China, Iran and Vietnam. All US search engines or content hosts cannot locate those functions in the bad countries. (Note that the definition of search engines or content hosts covers anyone who has a search tool on their website or permits users to generate content). Search engines also cannot change their filtering based on requests from bad countries. Content hosts also can't help "Internet jamming" and can't disclose personal data at the request of bad countries.
This law seems terribly misguided. It’s as if Rep. Smith thinks that Google is so amazing that countries will change their censorship laws just to get Google’s services. But we know better. Chinese entrepreneurs will have no problem providing competent yet censorable search services. I’m sorry to be the bearer of bad news, Rep. Smith, but embargoing Google isn’t going to bring down the current Chinese government.
Meanwhile, this law represents a dangerous step towards government regulation of search engine operations. I know that pro-regulation forces would love to have the chance to regulatorily inculcate their normative values into search engine algorithms; this law represents a first step along that path. However, I think there’s little chance that government fiat will improve search engine coverage or relevancy. Instead, I think there’s a much better chance that government intervention in search engine operations will degrade search engines’ usefulness to consumers.
Reading these 3 laws in succession, two 1980s songs came to mind. First, Congress “just can’t get enough” regulation of the Internet. However, "If you love somebody, set them free." We’ll see just how much Congress loves the Internet as it wrestles with these bills.
March 04, 2006
San Francisco Presentation, March 15 12:30 pm
By Eric Goldman
I'm presenting my latest article, A Coasean Analysis of Marketing, at University of San Francisco on March 15. The talk is free and open to the public.
Although the title suggests that the talk will be heavy on economics theory, I'm giving a "economics-free" version of the talk. As an added incentive, I suspect the Q&A will be fun as audience members hammer me for my pro-spam, pro-adware arguments.
What: "A Coasean Analysis of Marketing," alternatively titled "Regulating the Distribution of Marketing"
When: Wednesday, March 15, 12:30-1:30 pm
Where: University of San Francisco School of Law, 2130 Fulton Street, San Francisco. (Official university directions). Go to Kendrick 102.
How: RSVP to Julia Dunbar [firstname.lastname@example.org]
If you're in the Bay Area, it would be a delight to see you there!
March 03, 2006
NCSoft Sued in South Korea for ID Theft
By Eric Goldman
NCSoft has been sued in South Korea for allowing users to improperly register Lineage/Lineage 2 accounts in other people's official Korean ID number (I'm inferring this is similar to a social security number). More than 3,500 people have joined the class action so far, although the affected number is in the hundreds of thousands.
Based on this report, my understanding is that an organized crime ring stole a large number of Korean IDs from a third party shopping website, used those IDs to create Lineage accounts, used Chinese gold farmers to manufacture in-world wealth, and then converted that to physical-world wealth.
Assuming this is true, I don't immediately understand how NCSoft could be liable to the people whose IDs were stolen. It's not clear that NCSoft played any role in the initial ID theft, and so far the news reports indicate that the people whose IDs were stolen have not suffered any damage. If NCSoft had no role in the initial theft and the people whose IDs were stolen suffered no damages, I'm having a hard time seeing how this is NCSoft's problem. Certainly, in the US, I can't see how the plaintiffs in this situation could state a valid cause of action.
As a result, this lawsuit smells fishy. The organizers run a case auction service that matches victims with lawyers for all types of lawsuits. (From their website: "Case Auction is a bidding system of which clients find out lawyers to handle his/her case through the auction.") Could this lawsuit just be a traffic driver for the website?
Alternatively, lawyers just may be trying to capitalize on consumer outrage. I'm inferring from news reports that NCSoft collected the ID number unnecessarily and consumers are ticked about the security breach and its possible implications (even if no damages were caused here). If the analogy is that an online service provider collected social security numbers are part of their authentication process, I see why some people would want answers about the necessity of such data collection. This article recaps some of the controversy.
Thanks to Matt Goeden for pointing this out. More coverage at Terra Nova.
March 02, 2006
Trademark Dilution Revision Act Update
By Eric Goldman
It's been a while since I've blogged on the Trademark Dilution Revision Act, HR 683 (now the Trademark Dilution Revision Act of 2006). The bill keeps chugging along. In February, it passed the Senate Judiciary Committee unanimously (draft as passed), and it's now on the Senate's legislative calendar. I'm not clear enough on Congress' internal mechanics to tell if this means it will pass this session, but the opposition to this bill doesn't appear to have raised insurmountable hurdles.
UPDATES: The law has passed. My comments on the Trademark Dilution Revision Act of 2006 as passed.
March 01, 2006
Symantec v. Hotbar Lawsuit Settles
By Eric Goldman
Symantec Corp. v. Hotbar.com, Inc., Case No. C05-02309 (notice of dismissal Feb. 1, 2006)
In June, Symantec sought a declaratory judgment that Symantec could characterize Hotbar's software as "adware." This lawsuit has settled. According to ComputerWorld, "the deal calls for Symantec to dismiss its suit, but continue to classify Hotbar's program files as low-risk adware."
It's good that this particular matter got resolved, but legal disputes over characterizing software as adware/spyware will not go away any time soon. As a result, there would be significant benefit to creating some unambiguous breathing room for anti-spyware vendors to make their characterizations without having to seek declaratory judgments to do so.
Hat tip to Geist's Internet Law News for catching this.