February 28, 2005
Nielsen-NetRatings Search Engine Loyalty Study
Nielsen-NetRatings released its latest study showing that searchers use multiple search engines. For example, the study says “58 percent of Google searchers also visited at least one of the other top two search engines, MSN Search and Yahoo! Search… Nearly 71 percent of those who searched at Yahoo! also visited at least one of the other top two search engines, and 70 percent of those who searched at MSN also tried their luck at one or both of the other two.” I’ve recently been experimenting myself while watching the search engines index this new blog. While Google has the best sorting algorithm, I’ve noticed that it has been acting a little strange recently. Meanwhile, Yahoo’s and MSN’s sorting algorithm have been noticeably inferior, and they generally seem to index fewer pages than Google—BUT, they have been delivering results that Google has missed entirely. So the lesson remains clear to me. Although Google remains the juggernaut, it’s not complete, and other search engines can be a useful supplement.
February 27, 2005
Good GQ Article on Google
February 25, 2005
Associated Press on Spam Filters
Associated Press article about IAP spam filters, specifically going after Verizon’s policies. There’s not really much new in this article, but it does give me a chance to make a few points about spam filters:
* any type of filtering mechanism, in any medium, will create Type I/Type II errors. It would be a mistake to overreact to the fact that spam filters block some wanted messages. This statement says it all: “Bruce Gingery, a security consultant in Cheyenne, Wyo., says users should simply get used to losing mail.” As much as we don’t like this, there is no perfectly reliable communications medium, and that includes email.
* there is no meaningful accountability for an IAP’s choice of spam filters. IAPs rarely or never disclose their filtering techniques, so subscribers have no idea what spam filter is being used, and IAPs may change these filters from day-to-day without any notice. We ran into this issue in the late 1990s with IAPs who subscribed to the Realtime Blackhole List, a very crude list that blocked lots of wanted emails based on suspect criteria. Despite this, I find it hard to believe that many subscribers really want IAPs to disclose their practices. I would imagine that such disclosures would be read and acted upon about as often as privacy policies or EULAs are.
* We as email users are schizophrenic about IAP server-level spam filtering. On the one hand, we want to make choices ourselves about what emails we get or don’t get. As a result, we’ll drop an IAP using overzealous filters. On the other hand, subscribers routinely complain about too much spam, so we reward IAPs for being more aggressive about spam filtering. Plus, IAPs simply have to do server-level filtering to manage their servers and keep prices competitive. It’s easy to see how IAPs get caught in the middle.
* Blocking emails from an entire country (like Verizon does, although IAPs have been doing this for years) overreacts to the spam problem.
February 24, 2005
Dealer Management Systems v. Design Automotive Group
Dealer Management Systems v. Design Automotive Group, Ill. Ct. App. Jan. 2005. Vendor sells $20k of software and $15k of services using an unsigned purchase order. Customer doesn’t pay. Vendor sues. Customer defends claiming Statute of Frauds. Court concludes that this transaction is a UCC sale of goods, citing a large number of cases treating software as a UCC good and going out of its way to include customized software as part of a good. As a result, UCC SOF applies and vendor loses.
Another Follow-up on Microsoft Antispyware Misclassification
Search Engine "Impression Spam"
Click fraud has been well-discussed among SEMs/SEOs, but Rob McGann runs a good article in ClickZ calling attention to “impression spam.” This occurs because Google considers click-through rate in its sorting algorithm for paid ads. If an ad gets a low click-through rate, Google degrades placement. Thus, someone trying to game the system can generate lots of “fake” impressions, increasing the number of impressions without an increase in click-throughs, and drive down the ranking of that ad. The gamer could then take advantage of this by getting top positioning by paying a lower CPC. Google has taken some heat for its approach to click fraud; ultimately, it will need to build more game-proof systems if it wants to keep advertisers happy.
Block v. eBay
Block v. eBay, #105CV035930 (Cal. Superior Ct. filed Feb. 17, 2005). Important new class action lawsuit against eBay. Internet News and Reuters have focused on the claim that eBay acts as a shill bidder by exhorting high bidders to up their maximum bid. If the bidder actually responds to the request, then eBay raises the bid even if no other higher bidders emerge.
This claim is somewhat interesting, but I’m much more interested in the claims that eBay is an auction house governed by California’s Auction Act. As far as I can recall (let me know if my memory is fading), this is the first lawsuit against eBay making this claim. (In the Gentry case, the plaintiffs claimed eBay was a seller of sports memorabilia, and I believe some governments have made inquiries to eBay). Whether eBay is an auction house/auctioneer or not is a bet-your-business issue that eBay simply cannot afford to lose. Therefore, it was a high-risk move for Lerach Coughlin to put this issue on the table. This may make eBay more willing to settle; but it also might mean that eBay will invest whatever resources it takes to win this point.
I’m sure seeing this class action claim makes eBay wish that it could have retained the mandatory arbitration clause that implicitly got struck down in the Comb v. PayPal case.
Penenberg on WSJ and Search Engines
Adam Penenberg argues that the Wall Street Journal eventually will become irrelevant because its subscription model (1) makes its content invisible to the search engines, and (2) prevents bloggers from linking to it. I think he’s right. Our assessments of branding are increasingly becoming measured by online buzz and by having the brand reinforced by good Google placement. If the WSJ gets neither, over time this will degrade the brand; at minimum, it will allow others to build strong competitive brands using the buzz/search engine methodology.
February 23, 2005
Fracas over iDownload C&D Letters
iDownload has been sending C&D letters to people who have characterized iSearch software as “malware” or “spyware.” See postings at Spyware Warrior and CastleCops. I’ve blogged on the misclassification problem before, and I remain deeply troubled that valuable software functionality will be prevented from coming to market in an overzealous pursuit of “spyware.” On the other hand, these C&Ds strike me as completely inappropriate. First, there are no well-accepted definitions of “spyware” or “malware.” Second, these characterizations may be protected opinions. See, e.g., Search King Inc. v. Google Technology, Inc., No. CIV-02-1457-M (W.D. Okla. Jan. 13, 2003) and Search King Inc. v. Google Technology, Inc., 2003 WL 21464568 (W.D. Okla. 2003). iDownload might consider spending its energies educating consumers rather than trying to conform discussion to meet its standards.
Search Engine Strategies Session on Legal Issues
Grant Crowell summarizes the Search Engine Strategies panel on legal issues, including some rather inoffensive quotes from me. I had a lot more outrageous things to say on the trademark topic, but they didn’t appear to make the cut.
Dating Search Engine
Lycos is planning to launch a meta search engine for dating websites.
February 22, 2005
Trademark Dilution Revision Act of 2005
The Trademark Dilution Revision Act of 2005 (HR 683) was introduced February 9, with a hearing of the House Committee on the Judiciary, Subcommittee on Courts, the Internet, and Intellectual Property on February 16. The law appears to be a trademark owner’s wish list, most specifically to overturn the Moseley case by allowing an injunction based on likelihood of dilution, not actual dilution. Even more disconcerting to me is the proposal to create a cause of action against defendants who “willfully intended to trade on the recognition of the famous mark” or “willfully intended to trade on the reputation of the famous mark.” These seem to implicitly codify the goodwill misappropriation/initial interest confusion/"diversion" cases that have made trademark law so plaintiff-favorable.
UPDATE: The law has passed. My comments on the Trademark Dilution Revision Act of 2006 as passed.
More on Dutch Search Engine Lawsuit Against Microsoft
Michael Geist’s ILN reports that Microsoft has altered its software and issued an apology in response to the Dutch search engine’s lawsuit over being labeled spyware.
February 21, 2005
RIAA Settlement Tactics
Great article describing the RIAA’s tactics in trying to demand settlements with alleged file-sharers. I have had some conversations with people who have been RIAA targets, and this article is dead-on. The RIAA generally doesn’t bend or compromise beyond its pro forma package—they want their money and the law is behind them, so they are inflexible once someone is targeted. I do think it’s interesting that the article mentions a de facto requirement of at least 500 songs to catch the RIAA’s attention. I wonder if that’s universally true.
Lawsuit Over Spyware Label
Anthony Greco allegedly becomes the first person arrested for “spim” (instant messaging spam). But the LA Times article continues: “MySpace spokesman Bennet Ratcliff explained that the company's messaging system is technologically more like regular e-mail than instant messaging, and users must click on a message to see its full content.” Typical sensationalism.
Alan Perkins' Response on Black Hat/White Hat
Alan Perkins replied to my previous post on black hat/white hat issues in SEM. (I’m quoting his email with his permission). He took issue with my placing search quality responsibility solely on search engines. He writes:
“Why hold a search engine solely responsible? For example, if a searcher uses a very broad query ("cars") when they are looking for a very detailed answer ("Used sports cars less than $20000 within 30 miles of my current location") should the search engine be responsible for the quality of the result?”
I think this is a great example of my point. Obviously, if the searcher had a very detailed query, the searcher has poorly selected a generic search term. However, we know that many searchers do a poor job selecting search terms. Search engines can differentiate by catering to these needs better than others. What searchers want is a search engine that reads their minds—that realizes that the search term “car” means “Used sports cars less than $20000 within 30 miles of my current location.” Searchers will reward search engines that do a better job of mind-reading. This may require search engines to draw more information from the searcher than a single decontextualized search term (such as by aggregating personal information from other online behavior). But however they do it, search engines will be rewarded or punished for mind-reading.
He also takes issue with various forms of publisher cloaking. He writes:
“If the publishers dupes the search engine by delivering content to the search engine that its searchers won't see, then potentially the search engine is delivering poorer quality results to its searchers, even if the publisher is managing to convert at such a level that there is profit in the deception….I agree that the search engine is responsible for the quality of its results if it sees what its searchers see.”
If cloaking can distort search engine relevancy, then the search engine’s technology and processes are defective. Of course, search engines have a variety of tools to combat abusive cloaking, and they use them at least sometimes. So search engines will find a way to weed out many types of abusive cloaking or will be punished by unhappy searchers who don’t find what they are looking for. However, if the searcher finds that cloaking gets the searcher to a result they want, I’m still not clear why this isn’t a good thing. We also know that cloaking is not always bad—indeed, as Danny Sullivan has pointed out, Google has approved cloaking for its Google Scholars tool.
Thanks to Alan for the thought-provoking response.
The Long Tail
I recently caught up with Chris Anderson’s excellent Wired article The Long Tail from October 2004. This article discusses the consequences of disaggregating copyrighted content from the physical manufacturing/distribution/retailing chain that historically has been required to get that content to market. Because of disaggregation, niche content can become economical to distribute, and he gives some great case studies of this phenomenon in action. In particular, he makes a persuasive case that we all have minority interests that are not economical to explore today but can become economical in the future.
However, his article doesn’t do a thorough job of explaining how consumers will find the niche content they want. He cursorily references Amazon’s recommendation engine and some other editorial/filtering processes, but we still have not seen the emergence of a widely-used process that reliably orders a pool of heterogeneous content based on taste preferences. Until then, it remains very difficult to satisfy our niche interests.
Google's AutoLink tool
Google’s AutoLink tool has been in the news. AutoLink picks certain words on a web publisher’s page and turns those into hyperlinks of Google’s choosing.
I’m struck by how often this idea has come up. I remember a client pitching this idea to me back in 1998. Then, there are tools like IntelliTxt, which has run trials (with permission) on sites like Forbes.com. And, of course, who can forget Microsoft’s Smart Tags?
From a legal standpoint, AutoLink looks questionable. The tool modifies publisher’s web pages by adding hypertext links without the publisher's consent. While this modification isn’t a huge change, I could still see some (many?) courts treating them as unauthorized derivative works. Honestly, it seems like a fairly routine copyright infringement. Google appears to be trying to position this as a situation where it’s merely acting as an agent for user instructions, but I’ve just recently blogged on how courts frequently slice through that argument pretty quickly.
More interesting to me is that the idea of using web words as search terms seems to keep coming up, for what is likely a very good reason—because consumers find these links valuable. Though we can’t ignore the copyrights of web publishers, I’m also troubled that copyright law might stand in the way of tools that consumers might find valuable and might expressly choose to use for themselves.
I have expanded my discussion about copyright infringement here.
February 18, 2005
Spyware, Researchware, Trackware, Greyware...What to Ware?
ClickZ’s Rob McGann reports that comScore’s tracking software has been on a spyware “rollercoaster.” CA called it spyware, then changed its mind, and has changed its mind yet again. comScore wants a new category for the software: “researchware.” Computer Associates did create a new category called “trackware,” which is still spyware to CA, and put comScore’s software in it. However, CA is kind enough to permit comScore to “appeal” the decision in CA’s private courtroom.
David Nason of eAcceleration calls trackware “greyware” (presumably because it’s neither black nor white?). He continues: “There is anti-spyware software that is arguably spyware. There is anti-spyware software that is not spyware that gets identified as spyware and removed by competitors. When it gets down to the consumer level, it's hard for people to know what to believe.”
Reading that quote makes my head hurt! Meanwhile, I’m proposing a new category called euphemismware.
UPDATE: Suzi has some fun with nomenclature--ultimately championing "sneakware."
More on Termination of Virtual World Accounts
Greg Lastowka of Terra Nova blogged on my previous post about termination of virtual worlds accounts. If you’re not familiar with Terra Nova, it is the leading blog on virtual world issues. Good comments over there too. Check out the discussion.
February 17, 2005
Black Hat/White Hat SEO
Great description by Andrew Goodman of the Black Hat/White Hat panel at Search Engine Strategies Chicago. (I attended the panel and Andrew captures the event perfectly). You should also read Alan Perkins’ follow-up “Ethical Search Engine Optimization Explained,” where he makes a number of interesting points.
Personally, I find the Black Hat/White Hat debate fascinating but misdirected. I think the debate should focus on who bears ultimate responsibility for delivering what searchers want. I would hold search engines solely responsible for search quality, because they have full market incentives to maximize relevancy, and they can control the technology and rules/practices that determines what a searcher sees. Because I put the responsibility on search engines, I don’t think publishers or SEOs have any ethical responsibilities to help the search engines deliver relevant results. This is not to say that publishers can do what they want without consequence; there may be legal limits (like false advertising), and there may be private rules promulgated by the search engines. However, I’m not troubled by the tricks publishers use to game the search engines; if those tricks succeed, I view that as the search engine’s failing.
The moralistic wing of the SEO/SEM community would hold publishers liable for bad behavior, although there remains deep divisions over what constitutes bad behavior. I think pointing the fingers at publishers is counter-productive. Either traffic converts or it doesn’t. If it converts, didn’t the searchers have a successful search? We want publishers to reach this result. If traffic doesn’t convert, then why is the publisher/SEO wasting its time? So long as publishers don’t engage in false advertising, searchers will make the ultimate call of whether publisher content is relevant or not.
Alan Perkins’ ethical guidelines go even further, putting some ethical responsibilities on searchers. While this sounds noble in spirit, it is too idealistic to be useful. Until searchers are adequately taught how to search, we must treat searchers based on how they actually behave, not as we wish they would behave.
February 16, 2005
Copyright Case on Photos and Descriptions
SMC Promotions, Inc. v. SMC Promotions, 2005 WL 292492 (C.D. Cal. Feb. 7, 2005). Plaintiff is a product distributor that relies on affiliate marketing. Plaintiff grants a copyright license to affiliates to use catalog photos and descriptions on the affiliates’ websites, but restricts the licensee’s ability to delegate or authorize another person to exercise the right.
Defendant offers to build websites for plaintiff’s affiliates. To make it easy for the affiliates, the defendant creates a database of plaintiff’s copyrighted photos and descriptions that it will integrate into each affiliate’s website. To avoid the obvious copyright problem when it aggregated the photos and descriptions, its user agreement with each affiliate says that defendant is acting as the affiliate’s agent for purposes of operating the affiliate’s website.
Thus, the argument goes, the user has a license to the photos and descriptions, and the defendant can merely rely upon that. This contract appears to position the defendant as a passive implementer of its users’ instructions. Unfortunately, this type of legal jujitsu rarely works, and it failed here. The court says that the plaintiff’s license to affiliates can’t be stretched that far, so the defendant loses the copyright claim.
The court also finds that defendant engaged in trademark infringement. The domains sellsmc.com, smcforums.com and smcpromotions.com are all deemed to create initial interest confusion, irrespective of the content or any disclaimers at the websites.
Copyright in Tattoos
Tattoo artist claims copyright in the tattoo he put on Rasheed Wallace. Note to kids out there contemplating getting a tattoo: don’t do it, but if you do, at least practice safe tattoo-ing and get an adequate copyright license from your tattoo vendor.
This isn't the first time that there has been questions about Rasheed's commercialization of tattoos.
Ninth Circuit En Banc ruling in Gator.com v. LL Bean
Gator.com v. LL Bean, 2005 WL 351228 (9th Cir. Feb. 15, 2005). In 2001, LL Bean sent a C&D to Gator (now Claria). Gator responded by suing for a declaratory judgment. The district court dismissed Gator’s lawsuit for lack of personal jurisdiction over LL Bean. The Ninth Circuit reversed, and then granted an en banc rehearing.
So far, so good. Now things get weird. The parties brief the jurisdiction issue and make oral arguments. Then they reach a “confidential” settlement that “does not provide for the dismissal” of the Ninth Circuit appeal. The Ninth Circuit asks for a copy of the settlement agreement. The parties submit the agreement “under seal.” However, the Ninth Circuit learns that no court has ever agreed that the settlement was confidential. Therefore, the Ninth Circuit says “it is appropriate for us to disclose the settlement agreement's content because the outcome of our mootness inquiry hinges upon those specifics.”
Out goes the confidentiality seal. The Ninth Circuit proceeds to dish on the settlement terms:
Gator got a 3 month “wind-down” period where it could display up to 25 pop-ups per month over the LL Bean site. After that, the LL Bean site is off-limits to Gator. Gator also paid money (unfortunately not revealed) to LL Bean, and LL Bean waived all claims relating to the pop-ups.
Finally, the parties make a “side bet” on the 9th Circuit’s ruling. If LL Bean wins the ruling, it gets an extra $10k; otherwise no money is exchanged. Although the court’s discussion seems to suggest that other litigants have used side bets in the past as a way to preserve a lawsuit post-settlement, I find the approach uncomfortable. Gambling on the lawsuit’s outcome is illegal; why isn’t this?
The en banc majority concludes that LL Bean’s release of liability moots the appeal and dismisses the lawsuit. A minority would have let the lawsuit continue because the side wager was enough to provide an actual case or controversy. From my perspective, the majority got it right.
Of course, this leaves the adverse Ninth Circuit jurisdictional ruling against LL Bean as precedent.
House Subcommittee Approves SPY Act
The Subcommittee on Commerce, Trade, and Consumer Protection of the House Committee on Energy and Commerce has approved the SPY Act with a modification to clarify that cookies are not covered.
A blog worth watching
There are a lot of good blogs out there, and at some point I’ll provide a list of the ones I check regularly. But today, by accident, I stumbled across the Guiding Rights blog by Mark Partridge and immediately added it to my RSS reader. Looks like a low volume but high quality contribution to the blogosphere.
February 15, 2005
Claria Goes Mainstream
Claria is planning to deliver ads by buying excess inventory and reselling the inventory at a higher price by using behavioral targeting. This seems to turn Claria into a pretty run-of-the-mill ad network. Is this a good thing or a bad thing?
Microsoft Anti-Spyware Tool
Microsoft will be giving away its anti-spyware tools. I’m not sure this will be a good thing. Given the inherent subjectivity of the definition of “spyware,” the last thing I want is Microsoft making that decision for me. Do you Microsoft will be particularly quick to label software as “spyware” if it might also have the ancillary effect of competing with Microsoft’s offerings? It reminds me of the age-old dilemma: who watches the watchers?
February 14, 2005
State Anti-Spyware Laws
Ben Edelman has put together this very helpful page on state anti-spyware laws.
Goodlatte Reintroduces Anti-Spyware Law
Goodlatte has reintroduced his anti-spyware bill. As of today, I didn’t see a bill number yet. Compared to the other anti-spyware bills, Goodlatte’s bill was relatively inoffensive. The basic structure was to criminalize unauthorized placement of a software program on a computer to comment another federal criminal offense, obtain or transmit personal information, or impair the computer’s security protections. I find this relatively unobjectionable mostly because it seems to overlap substantially (completely?) with the Computer Fraud & Abuse Act. If anyone figured out what Goodlatte’s bill would criminalize that wasn’t already criminal under the CFAA, I’d be grateful to hear from you.
More on TRUSTe's Seal Revocation
A follow up article to my previous post on TRUSTe. It looks like the problem will be solved soon by the licensee paying some money for additional employee training. This reinforces the question: just how much value is a TRUSTe license worth?
New UETA case
Lamle v. Mattel, 394 F.3d 1355 (Fed. Cir. Jan. 7 2005) finds that an email sent before the enactment of UETA satisfies the Statute of Frauds, and the email certainly would satisfy the SOF under UETA.
"Rich Internet Applications" and Spyware
Bob Tedeschi runs a good article on “rich Internet applications.” This technology is a small applet that is downloaded to the user’s computer to facilitate getting the user to the right place. For example, the software will monitor a user trying to check out from an e-commerce site and will request a correction if the user types an invalid zip code. Using this technology dramatically improves checkout rates—the article cites how TJMaxx.com had 50% more customers complete the checkout process using this technology.
So far, so good. But isn’t “rich Internet applications” a synonym for “spyware”? The software is surreptitiously downloaded to the user’s computer, watches their every move and varies its content displays based on user behavior (including, in some cases, based on users’ personal data). Poorly drafted anti-spyware laws have the risk of making these types of programs illegal or heavily regulated (i.e., lots of disclaimers/additional screens of disclosures before the user can gain the benefit). As a result, these types of technologies expose exactly how regulators are out-of-sync with the market and consumer behavior. In the zeal to be the toughest on spyware, there’s a real risk the regulators will merely force us back to hard-to-complete e-commerce checkout processes.
Paul Boutin on Typo Traffic
Paul Boutin discusses the problem of searchers making typos when typing domain names into the address bar. He runs through the typical litany of gripes about product efforts to solve this problem—Microsoft’s “blatant ploy” to promote MSN search when IE users look for non-existent pages, Network Solutions’ attempt to capitalize on its “monopoly” with its SiteFinder product, and the new “mutating germ,” Paxfire, that allows IAPs to redirect typo traffic to an ad page.
But what, exactly, is the problem? In other words, what’s worse: getting a useless 404 page, or getting potentially helpful ads or content that match your interests? Perhaps these repeated and ongoing efforts to deliver ads to typo pages are not the result of a sinister plot but instead reflect efforts to improve the search experience.
Boutin offers a typical engineering-oriented solution: A software program that lets him customize the response he gets when he makes a typo. I suspect the demand for this is not great—most people will not invest the time to train the software when the fix (hit the back button) for making a typo is so minimal. Meanwhile, I’ve got a MUCH easier solution for him. Don’t use the browser address bar at all. Instead, start every search effort at a search engine. Looking for Slate? Search for Slate at Google. That way, you’re much less likely to run into unexpected ads (or worse) if, in fact, you make a typo.
(Thanks to Marty Schwimmer for calling attention to the article).
February 13, 2005
Termination of Accounts in Virtual Worlds
I attended Santa Clara’s Rules & Borders conference on Friday, and the subject of virtual worlds came up extensively. One issue in particular is continuing to vex me. Some virtual world participants invest significant time and money in their online characters—earning (or otherwise obtaining) virtual money or items, gaining experience/levels, creating or customizing online “property” (such as houses or widgets that are for sale in the virtual economy) and forming social networks. In some cases, participants purchase items, cash or characters with real cash either directly from the provider or in a secondary market like eBay. All of these “investments” can be lost or diminished if a virtual world provider terminates the individual’s account or otherwise changes the rules or environment of the world.
Intuitively, it seems problematic that participants can lose their investments. On the other hand, most (all?) EULAs says a provider can terminate accounts or change rules at any time. Let’s assume that the EULA is properly formed as a contract (i.e., there are mutual manifestations of assent). With a EULA like this, there seems to be a conundrum: the participant has invested time/money in a world in which the fundamental ground rules appear to provide no protection for those investments. Given the shaky foundation, why should we care if participants in fact lose those investments?
I’ve been trying to think of other circumstances where a customer’s investments in a vendor relationship are protected even if the contract clearly says that the vendor may terminate the contract for convenience. I have been able to think of only two examples of this:
· franchise agreements. Franchising programs involve the sale of a business and share a lot of similarities with other sales of securities. As a result, many state laws limit franchisors’ rights to terminate the franchise as a way to protect the franchisees’ investments in the business.
· distributor protection laws. Some states restrict a manufacturer’s rights to terminate distributors of their products. Like franchise laws, these distributor protection laws implicitly protect investments made by distributor in the distribution business.
There are, of course, other ways in which a contracting party’s termination rights may be limited beyond the contract terms, such as rescission rights for fraud or other remedies for marketing misrepresentations. In addition, there may be statutory limits on a vendor’s ability to fire a customer for illegitimate reasons (such as racial or gender discrimination).
One other analogy comes to mind, even though it’s not a classic vendor/customer relationship—the laws limiting employment termination at will. In some (increasingly rare) cases, laws limit an employer’s right to terminate an employee at will. While these laws do not explicitly protect an employee’s “investment” in the relationship, they do so implicitly (protecting out-of-pocket and opportunity costs incurred by the employee by joining the employer).
I am sure there are other situations that limit a vendor’s ability to terminate a customer for convenience. I’m opening comments on this post—if you can think of any, please speak up.
However, for now I’m struck by how rarely the law protects a customer’s investment in a contract relationship when the relationship says that the contract may be terminated for convenience. The general approach seems to be that customers who make those investments without a solid contract footing are making speculative investments.
The recent Second Circuit opinion in Hall v. Earthlink Networks provides a good case study. Hall opened up a personal Earthlink account (not to be used for business purposes). Earthlink’s backbone provider erroneously notified Earthlink that Hall’s account was sending spam, so Earthlink terminated the account and put the account on the spam watch list. Hall’s lawsuit for breach of contract and breach of implied covenant of good faith and fair dealing got zero traction (although, interestingly, the court didn’t even discuss the EULA to dismiss the complaint).
As the case illustrates, Hall was told not to use his email account for business purposes, did so anyways, had his economic expectations frustrated when Earthlink did exactly what it said it could do in its EULA, and walked away empty-handed in court. Is there any reason why should we treat virtual world participants any differently? If the EULA clearly says that all investments the participant makes are at their risk, I’m struggling to think of any reason why we should be sympathetic when, in fact, the participants lose those investments.
Avoiding Attention Distractions
Katie Hafner writes a good article on the difficulties we have avoiding distractions when we use our computer. It’s so true! There are so many temptations and messages competing for our attention. I know I struggle with this—there’s always something new going on somewhere in my of my email/web accounts, and it takes tremendous self-restraint not to prioritize those above more difficult/time-consuming but ultimately higher-value projects.
Katie talks about some of the technological efforts to prioritize computer-delivered announcements and alerts. If these efforts have any hope of succeeding, they need to learn our behavior and intelligently parse through a message’s contents. These intelligence-based efforts are exactly the type of technological developments that are threatened by the current emotional overreaction to spam and adware.
Randall Stross Whines About Spam
Randall Stross writes an article whining about spam as pollution and complaining that the recipient pays for spam. Haven’t we heard this argument before? As a matter of fact, we have. Often. A long time ago (his own article cites one of the early academic efforts attacking spam economics--from 13 years ago!). So why is this news?
I deconstruct some of Randall’s anti-spam rants in my article on spam harms. I will be more directly attacking the argument about “recipient pays” because of attention consumption in my next major article. For a high-level overview, see my short book chapter on data mining and attention consumption.
February 11, 2005
Ben Edelman's Response to My Post on Utah's Spyware Law
I got an email from Ben Edelman in response to my earlier posting on Utah’s anti-spyware law. If you don’t know Ben, you should. Ben has done some first-rate empirical research on the Internet, and I cited several of his research projects in my Internet search paper. Ben also is a leading crusader against spyware and adware, so we have crossed swords in the past.
Ben suggested that I overstated my argument that the proposed Utah amendments make adware illegal. Ben is right—I did overstate. As Ben reminded me, some types of adware are not covered, such as adware that delivers untargeted ads or that displays trademark-triggered ads on a delayed basis (he think this would not satisfy the “contemporaneous” requirement, though that may depend on judicial interpretation of the word).
While I may have gotten carried away in my earlier post, I think my core point stands. While some adware is not covered by the law, I think the law outlaws the only types that have commercial viability. For example, the law still allows adware displaying untargeted advertising, but who wants that? Most/all of the adware programs that delivered ads on a poorly targeted basis are dead—AllAdvantage is a leading example, but they are just one of dozens/hundreds of browser bars that failed. Consumers (and advertisers) have little interest in software that just throws up random ads. If that’s what the Utah law allows, gee thanks (for nothing)—there’s no one left in that space, and isn’t that a good thing?
Instead, consumers do want software that infers consumer interests through behavior. This is why trademark-triggering adware gets high clickthrough and conversion rates. This is also the wave of the future—in the future, software will try to read our minds through the words we type and our online activities. These types of useful software programs are precisely what the Utah amendment targets.
The more I think about this law, the more depraved I think it is. I did a thought exercise that clarified for me exactly why this law is so objectionable. Let’s start with a simple question. This law is basically a trademark law, so why doesn’t Utah just amend its state trademark laws instead of proposing a new law?
This would be easy to do. Utah could simply pass a law saying that a trademark “use” occurs when a trademark is used to trigger ads. This approach has a chance of surviving the dormant commerce clause. State trademark laws presumably aren’t immune from a DCC inquiry, but given the long-standing state regulation of trademarks, courts will likely be far more deferential about upholding a state trademark law than a sui generis law. Also, defining “use” would have a secondary “benefit” of giving plaintiffs the right to go after all types of trademark triggering, including search engines.
However, defining trademark “use” would still require plaintiffs to show likelihood of consumer confusion. But plaintiffs who would take advantage of Utah’s proposed law don’t want to be required to show consumer confusion because they probably can’t. Doesn’t this reinforce the problem with the proposed Utah amendments? The proposed law requires zero showing of consumer confusion. In fact, defendants lose under the law even if no consumer is ever confused about anything. As I mentioned before, this law applies even if consumers expressly and unambiguously want the software to trigger ads based on online behavior.
Even so, Utah could still use existing trademark law by saying that it is presumptively confusing to consumers to trigger ads based on third party trademarks. Doesn’t this accomplish the goal?
The short answer is no. Why not? Its principal advocates include 1-800 Contacts and Overstock.com—both companies with extremely weak “trademarks.” They do not want to rely on trademark law. They want protection for words/phrases that aren’t trademarks. This explains why the law covers “registered domain names” in addition to trademarks. Protecting domain names is the key to this bill—strike that, and there’s no reason why its advocates need this law. In other words, this law is really about creating sui generis property rights for the few Utah companies that chose weakly-trademarkable domain names.
As a result, this law isn’t about protecting consumers from software that harms their computer. This law isn’t about protecting consumer privacy. This law isn’t about protecting consumers from being misled when making purchases. This law is about taking choices out of consumer hands. This law is about restricting competition to increase profits for a few Utah companies. This law represents everything that’s wrong with our current legislative system.
February 10, 2005
Is TRUSTe Irrelevant?
TRUSTe pulled its logo from some websites for breaking its rules for the first time in 2 years. The article focuses on TRUSTe’s refusal to specify the rules violation, but this seems to miss the point, and widely. In 1999 and 2000, virtually every important website had a TRUSTe logo, and the logo had the potential to really shape consumer expectations and behavior. Now, I rarely see the logo, nor do I care if I see it or don’t. If my experience is typical (and I think it is), then TRUSTe has become irrelevant. I think a better angle for the story would have been—why does anyone care that TRUSTe yanked the logo?
February 09, 2005
More on Utah Spyware Law Amendments
I finally had a chance to look at the proposed amendments to the Utah Spyware Control Act. They are much worse than I imagined! The law talks about spyware but instead makes adware illegal in Utah--regardless of how the software is installed, any disclosures made by the adware vendor, and any consents given by consumers. In other words, a consumer may expressly and unambiguously want the software but the Utah law would deny them the opportunity to have it.
The proposed law represents a frontal assault on the use of keywords to deliver content to consumers. The law prohibits client software from using any trademark or domain name to trigger ads. As far as I can tell, the law permits any trademark owner to veto any sale of their trademark, regardless of context (i.e., Apple Computers could veto any ad triggered by the word Apple, even if the ad was selling fruit). To ameliorate this, the statute says: “This chapter does not preclude any person accused of violating this chapter from asserting any fair use or other defense that is available to persons alleged to have engaged in trademark infringement.” But I don’t understand how these defenses would work. Plaintiffs would not be suing for trademark infringement, so how can a defendant claim trademark law defenses? The law also protects against the use of non-trademarked domain names to trigger ads, so what trademark defenses are available against those?
The law gives a private cause of action to trademark owners but not consumers. This law is not about protecting consumers, and anyone who might claim otherwise is lying. This is pure anti-competitive protectionist rent-seeking behavior by a few trademark owners in Utah (like 1-800 Contacts and Overstock.com) using legislators as their shills. Shame on all of them.
Pornmeister Liable for Affiliate’s Spam
A settlement involving Sobonito Investments, a Cyprus pornmeister that used an affiliate marketing scheme to generate traffic. The affiliates engaged in bad email practices, and the advertiser was put on the hook. This situation is clearly covered by CAN-SPAM but this settlement involved incidents before the law passed. While I suspect there were in fact closer ties between Sobonito and its affiliates (a common practice in the porn industry), the idea that an advertiser should be liable for acts of independent third parties is a deeply troubling one—and I fear an overapplication of the CAN-SPAM provision on this point, which could have dramatic and negative effects on many types of socially-beneficial marketing.
French Lawsuit Involving "Search Engine Optimization" Techniques
I haven’t seen much discussion on the recent French opinion Societe Kaligona v. Societe Dreamnex, French Court of Appeals, Jan. 12, 2005. (I’m working from a BNA report—subscription required). The case involves two competing Internet sex shops. Dreamnex’s site saw a drop in traffic, which it attributed to lower rankings in French search engines. It then discovered that Kaligona had stolen some of its “search optimization techniques.” The BNA write-up is a little fuzzy about what this means. There’s reference to Kaligona stealing keywords and metatags, although later there’s reference to Kaligona taking a “detailed catalogue” of the site’s goods and services. The appeals court rejected any copyright claim in these “search optimization techniques” but still found against the defendant based on (according to BNA) unfair competition/parasitic commercial behavior/usurpation of creative rights. I really don’t understand these claims. According to the BNA report:
“the court did find that Kaligona's reproduction of Dreamnex's search engine optimization techniques represented a "deliberate" attempt to claim its rival's spot atop the search engine rankings, and thus cause "confusion" among would-be visitors.”
This could be a garden variety metatag case, but the write-up seems to suggest something more interesting. If anyone has or can get a good English translation, please let me know and I’ll post it.
Coca-Cola v. Purdy Permanent Injunction
The court issued a permanent injunction in the long-running Coca-Cola v. Purdy case. If you’re not familiar with Purdy, he is an anti-abortion activist. He targeted various publishers that he believed were pro-choice, plus a number of other companies that he thought should take an anti-abortion stance. Here’s one article recapping some of the story. He then registered various domain names including the trademarks of his targets, and then he redirected the domain names to anti-abortion websites (in some cases, causing web browsers to unexpectedly encounter some uncomfortable photos at the anti-abortion sites).
Not surprisingly, the court was sympathetic to the trademark owners. Even though Purdy has a political point to make, and many courts will consider his gripe rights, his techniques have been tough for the court to accept—-especially after he registered yet more domain names following an initial injunction.
Unfortunately, the new ruling illustrates why bad facts make bad law. Consider, for example, how the court resolves some factors of the likelihood of confusion test:
· Competitive proximity. An anti-abortion activist “competing” with major newspaper publishers and some big branded food companies like Pepsi and McDonalds? The court’s response? The parties compete “because the Defendant intended to redirect Plaintiffs’ audience and customers to view content of their choosing.” Treating a “diversion” as competition guts this factor.
· Intent to pass off goods. Of course the defendant isn’t selling any goods at all. He is trying to communicate his political message, and unquestionably there was little likelihood that any web browser would have assumed that the horrible photos were the “goods” of the trademark owner. The court’s response? Purdy registered the domain name with a bad faith intent to profit by diluting the marks and by “relying on Plaintiffs’ good names and goodwill to achieve the personal gain of promoting their messages, generating publicity and raising money for supported causes.” Bad faith intent to profit is important, of course, to an ACPA inquiry, and the court had already ruled against Purdy on that claim—-but inexplicably reuses the same language for the trademark infringement analysis. The references to using goodwill to promote a message might have relevance under trademark law, but they surely do not evidence that Purdy was trying to pass off goods as if they were the plaintiffs’.
· Purchaser care. This inquiry should properly focus on the care exercised by searchers for each of the plaintiffs’ respective products, which would require the court to treat each plaintiff differently. Instead, the court merely focused on the pre-purchase information marketplace, allowing the court to assess how careful Internet searchers are (rather than how careful buyers are when disgorging themselves of their hard-earned money). Framed this way, the court quotes a 2000 case for the proposition that:
“In the internet context, in particular, entering a website takes little effort - usually one click from a linked site or a search engine’s list; thus, Web surfers are more likely to be confused as to the ownership of a web site than traditional patrons of a brick-and-mortar store would be of a store’s ownership.”
I don’t know if this was ever true. Even if it was, I think we need a lot more empirical evidence to support this statement today in 2005.
Even though the court took too many liberties to reach a convenient conclusion, there remains no question that courts will continue to find ways to curtail aggressive marketing techniques by gripers. It seems like the gripers can still have their say, but not by automatically redirecting web browsers to pictures of aborted fetuses.
"They Don't Own the Jews"
Does a “Jewish Rock and Roll Hall of Fame” infringe the “trademark” owned by the Cleveland “Rock and Roll Hall of Fame”? Given the large number of “Hall of Fames” in the world, one might think there would be better issues to litigate. I think this quote sums it up perfectly:
Jeffrey Goldberg, Washington correspondent for the New Yorker, said, "Speaking as a layman, I don't think they own rock and roll and I don't think they own the phrase 'Hall of Fame' and I know for sure they don't own the Jews."
February 08, 2005
Google and Keyword/Domain Name Convergence
Bob Tedeschi speculates about why Google became a registrar. He discusses a number of good reasons but missed perhaps the most obvious. To Google, there’s no difference between selling domain names and selling keywords: both are merely ways for companies to use keywords to generate traffic. I make this point about convergence between domain names and keywords in my article on Internet search.
Amicus Briefs in MGM v. Grokster
If you haven’t checked it out, 28 amici briefs have been filed so far in the MGM v. Grokster case, including three different briefs led by law professors. I find it particularly interesting that Sen. Leahy and Hatch submitted a brief. It seems like they can influence the outcome in a lot of ways, so why did they need to resort to an amicus brief? I would expect a healthy number of briefs supporting Grokster on March 1. What will the court do with this deluge of submissions?
Utah/Urquhart on an Anti-Spyware Crusade Again
Utah is reworking its anti-spyware bill. I need to see the legislation but this news report sounds ominous. It reminds me of the expert report from the LICRA v. Yahoo case, where the experts predicted that we would be bombarded with queries about our geography as we traipsed around the Internet, making Internet browsing far less seamless than it is today (and for some, raising troubling issues about privacy disclosures). Having recently looked at Washington’s proposed anti-spyware law, and wondering how that would survive the dormant commerce clause, I am dubious that any state-based anti-spyware effort will improve social welfare one bit.
Shopbots and Nash Equilibrium
Interesting News.com article by Paul Festa discussing a report by Prof. Michael Baye of Indiana University on why shopbots have not led to homogeneous prices by all retailers. The basic argument: retailer are constantly changing prices to avoid letting competitors compete against a static target.
While the claim that we have reached a Nash Equilibrium makes the argument sexy, there is something missing in this argument (at least as summarized by Festa). If retailers are constantly changing their prices, consumers should be flocking to shopbots because they can provide time-sensitive and hard-to-aggregate information (i.e., which retailer is dropping its shorts on pricing today). This should give the shopbots significant power over retailers and lead to consumers constantly rewarding the low-price leader at the exclusion of the other retailers.
In fact, it seems like the opposite is occurring—many consumers use shopbots rarely or never. Is it because the price savings consumers may experience do not warrant changing their shopping patterns to check with the shopbots? Because consumers compare more factors than price? Any smart online shopper knows that you have to evaluate merchant reliability as part of the purchasing decision. So perhaps the price dispersion cited in the article reflects inefficiencies in the information flows or price premiums for merchant quality. Or perhaps the low-price leaders are pricing below-cost as one last marketing hurrah before they flame out in a blaze of glory.