Is this as good as it gets? An appraisal of ICANN's uniform domain name dispute resolution...

IMAGE TABLE 1IMAGE TABLE 166

1. INTRODUCTION

In October 1999, the Internet Corporation for Assigned Names and Numbers (ICANN) implemented

the Uniform Domain Name Dispute Resolution Policy (UDRP or Policy), establishing a uniform system for trademark owners to challenge "cybersquatting," which occurs when Internet domain names are abusively registered.1 Approximately three years after its implementation, it is important to examine the UDRP to determine whether it is achieving its intended result of providing a quick, inexpensive method for resolving domain name disputes fairly and accurately. This Note arrives at three main conclusions regarding the UDRP's success in realizing its goals. First, the UDRP deserves the acclaim it has received for providing a cheap, quick method of dispute resolution; in this regard, the UDRP offers a significant improvement over litigation. Second, the UDRP's procedural guidelines are ineffective for guarding against misapplication of the Policy. One consequence of the defects in the UDRP's procedures is that arbitrators lack guidance concerning which interpretations of the Policy are best. In the absence of such guidance, some dispute resolution providers appear to be applying the UDRP in a biased manner to gain market advantage over their peers. Third, the UDRP's incentive structure is flawed because it facilitates cybersquatting and "reverse domain name hijacking."2 As IMAGE FORMULA 170

a result, the phenomenon of cybersquatting continues unabated three years after the UDRP was implemented, and innocent domain name registrants are badgered by overly aggressive trademark owners.

In order to correct these problems, ICANN should first address the UDRP's procedural defects by creating a UDRP Review Board. The Board's purpose would be to identify or issue decisions that would act as binding precedents on all arbitrators. Over time, this process would flesh out the meaning of the Policy, to help ensure its consistent, fair, and accurate application. Second, ICANN should balance the incentives created through the implementation and use of the UDRP so that it encourages neither cybersquatting nor reverse domain name hijacking. Domain name registration fees should be modified to remove the economic incentives of cybersquatting. Additionally, the Rules for UDRP3 should be altered to include a penalty for reverse domain name hijacking to deter overly aggressive complainants from misusing the UDRP process.

Part II of this Note begins with a problem orientation that discusses the history of domain names and the domain name registration process, explores the rise of cybersquatting, and considers the circumstances that led to the enactment of the UDRP. Part III describes the goals of the UDRP, examines its substantive and procedural aspects, and recommends solutions for its procedural flaws. Part IV discusses problems with the incentive structure created by the implementation and use of the UDRP and proposes changes to address those problems. Part V discusses barriers to changing the UDRP.

II. PROBLEM ORIENTATION

A. Domain Name Fundamentals

Domain names are easy-to-remember combinations of letters and numbers (e.g., www.collectivemed.com) that facilitate the receipt and transmission of data over the Internet. Domain names were developed in the mid-1980s to make referring to IP addresses IMAGE FORMULA 175

(numerical strings like "98.37.241.30" that uniquely identify Internet computers and enable Internet communication) simpler and easier for human computer users.4 Domain names, IP addresses, and the communications protocols that manage how data is sent over the Internet comprise the domain name system (DNS). The DNS is constructed as a hierarchy of top-level domains (TLDs) and sublevel domains.5 A "root" server sits at the top of the hierarchy, maintaining a directory of all the servers that hold top-level domain directories.6 These servers, in turn, catalogue domain names registered under one of the more than 200 national, or country-code, TLDs (ccTLDs),7 or one of the generic top-level domains (gTLDs) including <.com>, <.net>, <.org>, <.edu>, <.gov>, <.mil>, <.name>, <.info>, <.biz>, <.pro>, <.coop>, <.museum>, or <.aero>.8 The servers that hold TLD directories link to other servers that maintain sub-level domain directories.9 When an Internet user inputs a domain name into his or her computer, a domain name server breaks the address into its top-level and sub-level constituent parts, queries server registries for the IP address associated with the domain name in question, and returns it to the user's computer, enabling it to retrieve or transmit data accurately. 10

Like IP addresses, domain names must be unique; otherwise, computers would be unable to accurately send and receive data over the Internet. This design limitation seems trivial, because a large number of variations exist for any given term an Internet user might wish to use as a domain name. However, there is no question that United Airlines finds <www.united.com> preferable to <www.ual.com>, so it is not surprising that many individuals are IMAGE FORMULA 177

reluctant to settle for such variations. Therefore, the domain name registration process plays an important role in the domain name system by establishing the rules under which Internet users compete for prized domain names.

B. Domain Name Registration

Since the mid-1980s, most domain names have been registered on a first-come, first-served basis. This laissez-faire approach remains in effect today for the three most popular TLDs: <.com>, <.net>, and <.org>.11 Registering a domain name in an unrestricted domain (e.g., <.com>) involves the following process: using the online "WHOIS" directory provided on domain name registrars' websites, would-be registrants check the availability of a particular domain name, select an unregistered name, and complete an online form by providing contact and payment information. Within several days the registrar establishes the new domain name by affiliating the name with an IP address and adding the domain name to the proper TLD directory servers.12 The notable aspect of this process is that anyone is free to register any domain name. The only prerequisite is being first in time to register.

It is worth noting that some TLDs and ccTLDs are restricted for use by certain user groups, such as <.mil>, <.edu>, and <.aero>. The restrictions are enforced through the addition of preregistration screening procedures that ensure that a registrant is qualified to register a domain name before a registration request is processed. The use of restrictive domain name registration procedures is growing in popularity in response to the rise in Internet trademark disputes.13 Nevertheless, it is unlikely that the introduction of restrictive domain name registration procedures will markedly reduce the number of Internet domain name registration disputes, IMAGE FORMULA 181

because the availability of new TLDs has not prompted companies to abandon their old <.com> domains. According to Wired, although more than one million <.info> and <.biz> addresses have been registered, most of the domains are being used as "pointers" to forward Internet browsers to websites in the <.com> TLD.14 By choosing not to shift their websites from the <.com> TLD, Internet users help ensure that TLDs with unrestricted registration procedures will not disappear anytime soon. Moreover, ICANN continues to roll out new unrestricted TLDs, which are just as prone to domain name disputes as the <.com>, <.net> and <.org> TLDs.

C. The Rise of Cybersquatting

While the unrestricted domain name registration process worked fine for academics and government personnel in the 1970s and 1980s, it quickly generated controversy after 1992, when commercial entities were allowed to begin using the <.com>, <.org>, and <.net> Internet domains.15 Soon afterwards, it became apparent that trademark law and the domain name registration process were in conflict; multiple non-competing users of trademarks who had previously shared the same mark to sell different goods and services suddenly found themselves competing to use a single <www.trademark.TLD> domain name.16 With only three TLDs available for use by the general public, certain domain names suddenly became scarce. Many trademark owners were slow to appreciate the importance of the Internet for their businesses and discovered too late that someone else had registered some or all of IMAGE FORMULA 185

the domain names that were identical or similar to their trademarks.17 Although a virtually endless supply of domain names reflecting variations on trademarked terms were available to these latecomers, such alternatives were unsatisfactory to many trademark owners, particularly those who had invested considerable sums of advertising dollars in their trademarks to make them famous. 18 For businesses with established trademarks like Vail Resorts, the prospect of settling for a second-best domain name like <www.snow.com> in place of <www.vail.com> was a bitter pill to swallow.19 Therefore, trademark owners began asserting that they were entitled to obtain the domain names that reflected their trademarks and that anyone who registered the domains before them was simply an unauthorized "cybersquatter."20

Some domain name registrants arguably deserved to be called cybersquatters because they used domain names to engage in trademark infringement and dilution, activities that are clearly barred by the Lanham Act.21 Other domain name registrants arguably did not deserve to be labeled cybersquatters, because their uses of domain names were not illegal according to then-existing law.22 IMAGE FORMULA 187

These uses included cyber speculation, which entailed registering well-known trademarks as websites to warehouse and resell them; cyber griping, which described a broad category of websites used to comment on, criticize, or applaud trademark owners' goods and services; and registration, for legitimate private uses, of domain names that happened to coincide with trademarked terms.23 With respect to these categories of domain name registrants, the allegation that one was a cybersquatter represented little more than a policy argument that the registrant was benefiting unfairly from the strength of the trademark owner's mark and should therefore surrender it.

D. Ineffective Legal Remedies

Trademark owners tried several different legal strategies for wresting control of contested domain names from their registrants, only to discover that current law largely failed to support their claims.24 Virtually all lawsuits against alleged cybersquatters were brought directly or indirectly as trademark infringement suits under the Lanham Act.25 To succeed on an infringement claim, the trademark owner was required to show rights in the trademark and establish that unauthorized use of the trademark by the cybersquatter was causing consumer confusion, deception, or mistake.26 To determine infringement, courts usually applied a "likelihood of confusion" test like the one developed in Polaroid Corp. v. Polarad Electronics Corp.,27 and refined by the Ninth Circuit Court of Appeals in AMF Inc. v. Sleekcraft Boats.28 Such tests evaluate: "(1) the strength of the mark; (2) proximity of the goods; (3) similarity of the marks; (4) evidence of actual confusion; (5) marketing channels used; (6) type of goods and the degree of care likely to be exercised IMAGE FORMULA 191

by the purchaser; (7) defendant's intent in selecting the mark; and (8) likelihood of expansion of the product lines."29 The application of tests such as this one produced inconsistent results for trademark owners, because many domain name registrants were not engaging in trademark infringement as they were not offering products or services similar to the trademark owners'.30 Furthermore, the Lanham Act features a "use in commerce" requirement to maintain an infringement claim, and many domain name registrants were able to convince courts that their uses of the domain names in question did not constitute such a commercial use.31

Domain name registrants relied on these loopholes in the current law to persuasively argue that there was nothing wrong with registering domain names that reflected trademarked terms if the domain names were not used to infringe or dilute the marks. After all, they argued, they were nothing more than good-faith opportunists participating in the high-tech equivalent of the 19th century land rushes.32 Because it was well established that trademark owners did not possess a monopoly right to fence off any use of a mark, or to block every possible combination of letters bearing any similarity to a protected mark,33 courts tended to reluctantly agree with this assessment.34

Sensing that the courts would not provide the relief they desired, trademark owners attempted to recover domain names from cybersquatters using other devices such as the dispute resolution process offered by Network Solutions, Inc. (NSI). NSI's dispute resolution process commenced when a trademark owner presented IMAGE FORMULA 193

NSI with a certified copy of its trademark registration documents and proof of written notice of the existence of the trademark to the domain name registrant.35 If the domain name was identical to the trademark, and the domain had been registered after the trademark was granted, NSI would place the domain name on hold until the parties settled their dispute or a court issued a final ruling in the case.36 Despite the appearance of providing an alternative to litigation, it soon became apparent that NSI's process afforded trademark owners the same protection as the Lanham Act, because the entire process depended on existing trademark law for its rules of decision.37

With the passage of the Federal Trademark Dilution Act (FTDA)38 in 1995, trademark owners obtained a new basis for bringing suit against alleged cybersquatters. The FTDA created a cause of action for lessening the capacity of a famous trademark to distinguish specific goods or services, regardless of the presence or absence of competition between the parties or the likelihood of confusion, mistake, or deception.39 Courts interpreted the FTDA's "commercial use" requirement broadly enough to include the registration of domain names with an intent to resell them, and trademark owners began prevailing against cybersquatters more consistently.40 Courts also stretched the scope of the FTDA by recognizing a new type of trademark dilution called "dilution by elimination."41 Because domain name registration and use is generally neither dilution by blurring nor by tarnishment, courts accepted the argument that domain name registrants were diluting trademarks by totally precluding the trademark's use as a domain IMAGE FORMULA 195

name and in turn eliminating the trademark holder's ability to effectively distinguish its goods and services on the Internet.42 Despite the adoption by many courts of this strained reading of the FTDA, obtaining relief under federal law from alleged cybersquatters remained an expensive gamble for trademark owners because alleged cybersquatters were often able to convince courts that they were not breaking any laws by registering domain names reflecting trademarked terms.43

E. From Allegation to Cause of Action

With the law governing domain name registrations growing increasingly unclear, the debate over cybersquatting shifted from courts of law to the court of public opinion and Congress.44 After rancorous debate between trademark interests and domain name rights types, new rules finally emerged in 1999 that attempted to clarify which domain name registration practices were acceptable and which were not.45 In rapid succession in late 1999, the newly created ICANN adopted the Uniform Domain Name Dispute Resolution Policy,46 and Congress passed the Anticybersquatting Consumer Protection Act (ACPA).47 With the implementation of the UDRP and passage of the ACPA, the term "cybersquatting" was transformed from a mere allegation into a legal term of art. According to the UDRP and ACPA, the essence of cybersquatting is the registration of domain names that reflect trademarked terms in bad faith and without a legitimate right.48 Beyond this basic IMAGE FORMULA 199

definition, the actual legal contours of cybersquatting remain unclear. In fact, the UDRP does not even use the word "cybersquatting." To trademark owners, the term "cybersquatter" conjures up images of technically savvy individuals like Sahar Sharid, who was described in one UDRP proceeding as "a young computer programmer whose interests include think[ing] of Internet domain names that he [can] buy for the standard $70 fee and sell at a huge profit to large corporations."49 However, what many trademark owners describe as "ruthless scalping in an under-regulated medium"50 is just good business, according to Internet entrepreneurs like John Zuccarini, who has registered more than three thousand web sites and earns between $800,000 and $1 million per year from their resale and use.51 More than two years after the enactment of the UDRP, fundamental disagreements regarding which types of domain name registration practices are acceptable and which are not remain largely unresolved, partly because a key element in a cybersquatting claim-bad faith-is notoriously difficult to define and prove, and partly because the UDRP's definition of cybersquatting is subject to widely varying interpretations. Because the debate over the moral and legal probity of cybersquatting remains unresolved, with no agreement in sight, this Note remains neutral on the issue and analyzes the effectiveness of the UDRP according to its stated goals and objectives.

III. THE UDRP

A. UDRP Goals

The UDRP is a bold experiment, arguably too bold. It expands trademark rights in cyberspace. It tests the outer limits of the feasibility of Internet self-governance. It establishes uniform global rules for domain name dispute resolution. It promises to IMAGE FORMULA 204

reduce the cost of resolving domain name disputes and to make the dispute resolution process quicker than litigation.52 In short, the UDRP's goal is to do everything national court systems do to resolve domain name disputes, but to do it on a global scale, for less money, in less time-all to prove that an alternative dispute resolution system designed by Internet users for Internet users is superior to relying on legislatures in various countries to impose solutions for Internet problems from the outside. To say the least, the standard for success set by the UDRP's creators is a high one, and like many other Internet experiments that were rolled out with fanfare and early acclaim, the UDRP has been only a partial success.

B. Adoption and Acceptance of the UDRP

ICANN ensured that the global adoption of the UDRP got off to a good start by relying on the assistance of the U.S. government, which legitimized ICANN (and thus, the UDRP) by promising to transfer authority over the domain name root server to it.53 ICANN used this authority to compel acceptance of the UDRP by leveraging its control over access to the domain name root server, forcing domain name registrars to adopt the UDRP as a prerequisite for accreditation.54 As a result, all registrars agreed to contractually bind registrants to submit to the UDRP process as a precondition to registering a domain name.55

Since its adoption, the UDRP has become the method of choice for resolving domain name disputes. As of February 8, 2002, approximately two years and four months after implementation, 4,288 proceedings have been brought under the UDRP, versus only eighty-three cases under the ACPA, the federal law similar to the UDRP that is also designed to combat cybersquatting.56 The IMAGE FORMULA 208

popularity of the UDRP is mostly attributable to its speed and costeffectiveness: resolution of a domain name dispute using the UDRP costs from $750 to $1,500 for a single panelist to resolve a dispute over a single domain name, and most disputes are resolved in about forty days.57 This is far quicker and cheaper than litigation. Additionally, the UDRP is a low-risk endeavor for the parties, who may bring suit after a decision is issued and obtain de novo review in a court of competent jurisdiction (respondents may do the same).58 The UDRP deserves acclaim for its ability to provide speed and costeffectiveness to the domain name dispute resolution process. However, despite its popularity, the UDRP needs significant substantive and procedural modifications before it may be considered a model applicable to other areas of law.

C. Substantive Aspects of the UDRP

Paragraphs 4(a)-(c) of the UDRP define the conditions that justify the cancellation or transfer of a domain name from its registrant to a complainant.59 The rules represent the culmination of a several-year-long consultative process designed to create a consensus between trademark owners and free speech, fair use, and domain name rights groups regarding acceptable and unacceptable domain name registration practices.60 Paragraph 4(a) states that any person may challenge a domain name registration using the UDRP by claiming all of the following three elements:

(i) [the registrant's] domain name is identical or confusingly similar to a trademark or service mark in which the complainant has rights; and

(ii) [the registrant] has no rights or legitimate interests in respect of the domain name; and IMAGE FORMULA 214

(iii) [the registrant's] domain name has been registered and is being used in bad faith.61

In theory, these requirements (the essence of a cybersquatting complaint) appear straightforward-simply prove all three elements, and the domain is yours.62 In practice, application of the UDRP's substantive provisions defining cybersquatting is not nearly so easy, because Paragraph 4(a)(i) is undefined, and the explanatory paragraphs for 4(a)(ii) and 4(a)(iii) provide arbitrators significant freedom in deciding whether the domain name registrant acted in good or bad faith. As the following sections describe in greater detail, these features make the UDRP a broad, flexible policy that invites widely varying interpretations as to what types of behavior justify the cancellation and transfer of a domain name.

1. Paragraph 4(a)(i)

The UDRP's drafters apparently believed that deciding whether a domain name is "identical or confusingly similar" to a trademark would be an easy task, because they chose not to provide a definition for this requirement.63 As a result, UDRP panelists are still determining what constitutes "identical" and "confusingly similar." As the following two paragraphs describe, no consensus has yet been reached.

Whether a domain name is deemed "identical" to a registered trademark depends on several factors. Obviously, the first question is whether the second-level domain name (e.g., "ford") is the same as the trademark. The more difficult question relates to the meaning of the TLD itself (e.g., <.com>). Some arbitrators assert that TLDs are irrelevant in trademark infringement analyses, while others assert IMAGE FORMULA 219

that the domain registrant's choice of TLD is particularly significant. For instance, one arbitrator reasoned, "The Panel sees no reason to require domain name registrants to utilize circumlocutions like <trademarksucks.com> to designate a website for criticism or consumer commentary."64 He went on to explain that this is true particularly where "the domain name registrant has not usurped the <.com> domain but has utilized only the <.net> domain."65 For these panelists, finding that a second-level domain is identical or confusingly similar to a trademark is only the first step in a two-step analysis. Panelists also consider the nature of the trademark owner's activities to determine whether they fall within the intended scope of each TLD. Thus, as the panelist in the Bridgestone Firestone case reasoned, the use of <www.bridgestone-firestone.net> was legitimate because "the domain name registrant has not usurped the <.com> domain but has utilized only the <.net> domain."66 Other panelists flatly disagree and ignore the TLD completely.67 The latter position ignores the architecture of the Internet, disregards the longstanding recognition that a trademark does not provide a monopoly over any and all uses of a term, and is almost certainly wrong. Nevertheless, anecdotal evidence gathered from reading numerous dispute resolution provider decisions leads this author to believe that this is the majority position among UDRP panelists.

The significance of TLDs in Paragraph 4(a)(i) analyses promises to become increasingly contentious as ICANN rolls out new TLDs, making it easier to distinguish different uses of domain names. ICANN is currently rolling out seven new top-level domains, including <.aero>, <.biz>, <.coop>, <.info>, <.museum>, <.name>, and <.pro>.68 These domains will facilitate the process of IMAGE FORMULA 221

meaningfully organizing the content of the Web by TLD, thereby strengthening the argument that different people can share secondlevel domains by using different TLDs to indicate the purpose each domain serves. For example, Wal-Mart could use <www.walmart.com> as its Web portal for online sales, while a union representing Wal-Mart employees could use <www.wal-mart.info> to publicize its stance on contract negotiations. Furthermore, ICANN is implementing new registration procedures for the <.aero>, <.biz>, <.coop>, <.museum>, <.name>, and <.pro> TLDs to help ensure that each TLD is used only for its intended purpose: helping to ensure that meaningful differentiation between TLDs becomes a reality.69 With all of these changes underway, the only certainty is that the debate over the significance of TLDs will likely remain unresolved for some time; as a result, so will the meaning of Paragraph 4(a)(i).

The other major unresolved issue for Paragraph 4(a)(i) concerns the meaning of the term "confusingly similar." The debate concerns whether websites that are supposedly dedicated to criticism of the trademark holder may use the domain name <trademarksucks.com>, or whether such a domain name is confusingly similar to the trademark. There exists a clear split among arbitrators on this issue. On one side arbitrators reason, "I do not see how a domain name including 'sucks' ever can be confusingly similar to a trademark to which 'sucks' is appended."70 This reasoning is followed by these arbitrators' belief that "[t]he use of the [sucks.com] suffix attached to a company name has become a standard formula for Internet sites protesting the business practices of a company."71 Other arbitrators reach the opposite conclusion. For instance, in a case involving the domain name <www.guinness-- beer-really-sucks.com>, the panelist found that the domain name was confusingly similar to the trademark "Guinness," by reasoning that IMAGE FORMULA 223

any domain name that includes a trademarked term is confusingly similar to the trademark itself. 72

Because the UDRP does not resolve these issues, it remains up to arbitrators to consider the circumstances in each case and then decide whether Paragraph 4(a)(i) is satisfied. In easy cases, where egregious behavior is involved, the existence of clear bad faith drowns out the significance of Paragraph 4(a)(i), making it easy for UDRP panelists to make what seems to be a reasonable decision. In hard cases, where each party presents seemingly legitimate claims to the domain name, it is difficult to predict how a particular arbitrator will decide whether the domain name is "identical or confusingly similar" to the mark in question.

2. Paragraph 4(a)(ii)

Paragraph 4(c) clarifies interpretation of Paragraph 4(a)(ii). Paragraph 4(c) states:

Any of the following circumstances, in particular but without limitation, if found by the Panel to be proved based on its evaluation of all evidence presented, shall demonstrate your rights or legitimate interests to the domain name...:

(i) before any notice to you of the dispute, your use of, or demonstrable preparations to use, the domain name or a name corresponding to the domain name in connection with a bona fide offering of goods or services; or

(ii) you, (as an individual, business, or other organization) have been commonly known by the domain name, even if you have acquired no trademark or service mark rights; or IMAGE FORMULA 232

(iii) you are making a legitimate noncommercial or fair use of the domain name, without intent for commercial gain to misleadingly divert consumers or to tarnish the trademark or service mark at issue.

These provisions carve out good faith "safe harbors" to provide arbitrators with clear examples of the types of behavior that may rebut a charge of cybersquatting by demonstrating the registrant's legitimate use of or interest in a domain name. If the domain name registrant can demonstrate a bona fide plan to use the domain name for a legitimate purpose, a relationship between the domain name and the registrant's name, or a bona fide noncommercial use for the domain name, the registrant should prevail in the dispute. In practice, determining good faith is difficult because every domain name registrant attempts to cloak his or her behavior in the language of Paragraph 4(c). As domain name registrants grow increasingly savvy, it becomes less and less clear which registrants are truly acting in good faith and which ones are not.

3. Paragraph 4(a)(iii)

Paragraph 4(b) provides guidelines regarding what Paragraph 4(a)(iii) terms "bad faith." Paragraph 4(b) states:

[T]he following circumstances, in particular but without limitation, if found by the Panel to be present, shall be evidence of the registration and use of a domain name in bad faith:

(i) circumstances indicating that you have registered or you have acquired the domain name primarily for the purpose of selling, renting, or otherwise transferring the domain name registration to the complainant who is IMAGE FORMULA 241

the owner of the trademark or service mark or to a competitor of that complainant, for valuable consideration in excess of your documented out-of-pocket costs directly related to the domain name; or

(ii) you have registered the domain name in order to prevent the owner of the trademark or service mark from reflecting the mark in a corresponding domain name, provided that you have engaged in a pattern of such conduct; or

(iii) you have registered the domain name primarily for the purpose of disrupting the business of a competitor; or

(iv) by using the domain name, you have intentionally attempted to attract, for commercial gain, Internet users to your web site or other on-line location, by creating a likelihood of confusion with the complainant's mark as to the source, sponsorship, affiliation, or endorsement of your web site or location or of a product or service on your web site or location. 75

These provisions establish that registering a domain name with an intent to resell it, to prevent a trademark owner from reflecting its trademark in a domain, or to disrupt a competitor's business, or using a domain name to confuse the complainant's customers are the basic behaviors that constitute bad faith. However, bad faith is defined "in particular but without limitation."76 This raises the question whether other factors alone are sufficient to prove bad faith or whether some or all of the elements listed with particularity are required. At least one panelist has interpreted the IMAGE FORMULA 250

UDRP to suggest that none of the enumerated factors need be present,77 but the majority of arbitrators reject this interpretation, recognizing that bad-faith registration and use are required to satisfy Paragraph 4(a)(iii), and adhere religiously to the factors listed in Paragraph 4(c) in their bad faith analyses.78

On the whole, the UDRP's substantive provisions reflect the consensus-oriented approach that was used to draft them. The result of attempting to achieve consensus between groups with strongly competing viewpoints 79 is an extremely broad, flexible policy that leaves most of the work of deciding what constitutes cybersquatting up to the arbitrators who apply the policy. Because arbitrators are relatively unconstrained in their interpretation and analysis of cases, the definition of cybersquatting has turned out to be whatever the arbitrator in a particular case says it is. While it is desirable for arbitrators to perform a fact-specific analysis in each case to determine whether the domain name registrant abusively registered a domain name, this vagueness greatly exacerbates the difficulties created by the UDRP's procedural framework, which fails to provide a mechanism through which arbitrators can refine the meaning of the UDRP over time. As a result, it is difficult to ensure that the UDRP is applied consistently, fairly, and accurately for all of the parties who invoke it.

D. Procedural Aspects of the UDRP

If the UDRP were backed up by a robust procedural system like that of the American legal system, complainants and respondents could rest assured that the UDRP's flexible definition of IMAGE FORMULA 254

cybersquatting would gain in definiteness as precedents were established to govern the UDRP's interpretation. Unfortunately, this process cannot occur for the UDRP because its austere procedural framework fails to provide a mechanism for establishing which UDRP decisions should be treated as precedents.

In order to understand the UDRP's procedural flaws, it is necessary to appreciate the UDRP's current procedural system. Trademark holders initiate the dispute resolution process by creating a complaint that: alleges the elements defined in Paragraph 4(a)(i)(iii) of the UDRP; lists the domain names in dispute and the domain name registrar who registered them; identifies the trademarks in dispute; lists the contact information of the domain name registrant; selects a one- or three-person panel; states that the domain name registrant has received notice of the dispute; and affirms the complainant will submit to the jurisdiction of the courts in at least one specified mutual jurisdiction. Once the domain name registrant receives the complaint, he or she is bound to submit to the arbitration process per the domain name registration contract that exists between the registrant and the registrar. Within twenty days the registrant must create a response to the complaint according to Paragraph 4(c) of the UDRP, describing his or her rights and legitimate interests in the domain name, or face a likely default judgment and transfer of the name.81 Upon receiving the response, the dispute resolution provider appoints either one or three "impartial and independent" dispute resolution panelists depending upon the election made by the complainant or respondent.82 If the complainant and respondent each provide rosters from which two panelists-- one from each side-are selected, and the third panelist is chosen by the provider from a list of five candidates approved by the parties.83 In exceptional cases, the panel may conduct in-person hearings or request further statements from the complainant and IMAGE FORMULA 256

respondent.84 Otherwise, the panel reviews the complaint and response and issues a decision in writing within fourteen days.85 The panel's decision is published to the parties within three calendar days after the decision is made, and the decision takes effect ten days later unless court proceedings are initiated.86

As a general rule, the UDRP lacks in-person hearings, the opportunity to file supplemental documents, and, most importantly, the opportunity to appeal a decision within the UDRP framework. These limitations enable the UDRP to achieve speed and low cost by sacrificing the procedural mechanisms that court systems rely on to achieve consistency and predictability of results. Recognizing that its emphasis on speed and low cost might erode the UDRP's ability to provide substantive justice, ICANN explicitly provided that complainants or respondents may file suit in a court of competent jurisdiction to obtain de novo review of a domain name dispute before, during, or after a decision is issued by a UDRP panel.87 This safeguard helps ensure that the parties in any particular dispute can obtain justice if they do not receive it from the UDRP. However, from a policy standpoint, this option should be viewed as a last resort. If the UDRP is to be truly successful, it must provide speed, low cost, and fairness on its own. The option of resorting to the courts fails to move the UDRP toward this goal, because lawsuits must be brought under the Anticybersquatting Consumer Protection Act of 19998 or the trademark laws of another country, rather than under the provisions of the UDRP. Consequently, courts never have occasion to interpret the UDRP. Even if they did, it is unlikely that arbitrators would be bound by such decisions anyway. The UDRP Rules provide: "A Panel shall decide a complaint on the basis of the statements and documents submitted and in accordance with the Policy, these Rules and any rules and principles of law that it deems applicable."89 This rule reflects the fact that the UDRP is a global dispute resolution policy not governed by the laws of any particular IMAGE FORMULA 258

jurisdiction. Therefore, while a panelist may consider established trademark law principles from any legal system in reaching a decision, such principles are secondary authority and are not binding.

E. Manifestations of Procedural Weaknesses in the UDRP

The interplay of the UDRP's vague wording and thin procedures makes it difficult for UDRP panelists to decide cases properly. As a result, panelists tend to make three common mistakes, each of which is discussed below.

1. Under-Reliance on Other UDRP Decisions and Principles of Law

Some arbitrators ignore the clutter produced by the thousands of (often conflicting) UDRP decisions that are issued by their fellow arbitrators each year and the principles of trademark law in favor of deciding cases in isolation. This approach is appealing for its simplicity, but it suffers from major drawbacks. First, it is inefficient because the arbitrator has to "reinvent the wheel" in each case. Second, it encourages superficial legal analysis, because panelists are operating under time constraints and do not have the leisure of taking as much time as might be necessary to reason through every issue "from the ground up." Third, it increases the number of conflicting interpretations of the UDRP that exist, confusing other arbitrators and the public concerning the Policy's real meaning. A simple example of the "go it alone" approach helps illustrate these drawbacks. One panelist breezed over his entire analysis of Paragraph 4(a)(i) with the following statement: "Zegna is a very strong trademark protected amongst other for [sic] products in the area of fashion and the panel is of the opinion that there is a confusing similarity between this mark and the domain name 'zegnafashion.com."'90 This analysis may be correct, but it leaves many questions unanswered. Why is "zegnafashion" confusingly IMAGE FORMULA 264

similar to "zegna"? What legal principle is this decision based on? How can the domain name registrant determine whether the arbitrator is right? It is impossible to tell. No matter how clear-cut the case or how easy the issue, arbitrators should not go it alone unless the issue is one of first impression, where guidance is unavailable from other UDRP decisions or from applicable principles of trademark law. Going it alone unacceptably decreases arbitrators' ability as a group to produce coherent, well-reasoned opinions, and it makes it impossible for complainants and respondents to understand the bases on which rulings are made.

2. Over-Reliance on Other UDRP Decisions

Under current UDRP procedures, there is no such thing as controlling authority; external courts do not interpret the UDRP's provisions, and there is no internal appeals process for affirming or reversing panelists' decisions.91 Yet panelists must grasp at something to guide their analyses. As a result, some panelists make the mistake of attempting to unilaterally decide which of the thousands of UDRP decisions issued each year deserve to be treated as governing authority in order to use them as justification for deciding cases. While looking to other panel decisions as persuasive authority is justified and laudable, these panelists give too much deference to prior panel decisions (many of which are wrongly decided) by claiming that they have stare decisis effect. For instance, the panelist in one case wrote:

Complainant, through registration and use has established that it has rights in the KINKO'S mark. Furthermore, Respondent's <kinkoids.com> domain name is confusingly similar to Complainant's mark because the disputed domain name merely adds the letters "ids" to Complainant's famous mark. It has been established that the addition of letters to a IMAGE FORMULA 269famous mark does not overcome a claim of confusing similarity. 92

The assertion that "it has been established" cannot be right, since the UDRP is a contract that binds only the domain name registrant. It is intellectually dishonest to claim that a prior panel decision has the force of binding precedent when it does not, particularly when there is no review process in place to confirm that such prior decisions are correct. Thus, in contrast to the case described above, the problem here is that arbitrators are leading parties to disputes to believe that an established rule of decision governed the outcome of the case, when no such rule exists.

3. Under-Reliance on the Policy as a Whole

A third error some arbitrators make is treating the UDRP's loose substantive and procedural provisions as a license to "do justice" in each case, regardless of whether or not the UDRP provides rules of decision governing the outcome. Dispute resolution panelists understandingly seek, in each UDRP proceeding, to reach a conclusion that does justice and reaches the right result. However, the UDRP procedures do not grant panelists the power of a court acting in equity. In fact, due to the sensitivity of replacing national laws with one uniform global law, ICANN narrowly restricted the UDRP's applicability to bad-faith cybersquatting and restricted the remedies it provides to canceling or transferring a domain name registration.93 These restrictions were intended to make it clear that the UDRP was not intended to displace other legal regimes that were already in effect.94 Rather, the UDRP was only supposed to be applied to "easy cases" that did not justify the high cost of litigation and were therefore not being brought under IMAGE FORMULA 274

established laws anyway.95 Many UDRP panelists appear to have forgotten their limited mandate. Indeed, instead of demonstrating restraint, they seem willing to ignore the express language of the UDRP in order to reach the outcome they think is best in each case.96 Perhaps the best examples of panelists overreaching the scope of the UDRP are the "nonuse as use" and "preclusion" doctrines that have been applied in UDRP decisions.

The "nonuse as use" doctrine was created in Telstra Corp. Ltd. v. Nuclear Marshmallows.97 The case turned on the question of the meaning of the word "use" in UDRP Paragraph 4(a)(iii), which requires that a domain name be registered and used in bad faith.98 The language of the UDRP is unambiguous on this point. Furthermore, during the comment period before the UDRP was formally adopted, an ICANN advisory panel specifically rejected arguments that the definition of bad faith should be expanded to include cases of either registration or use where both elements are not present:

These comments point out that cybersquatters often register names in bulk, but do not use them, yet without use the streamlined dispute-resolution procedure is not available. While that argument appears to have merit on initial impression, it would involve a change in the policy adopted by the Board. The WIPO report, the DNSO recommendation, and the registrars-group recommendation all required both registration and use in bad faith before the streamlined procedure would be invoked. Staff IMAGE FORMULA 277

recommends that this requirement not be changed without study and recommendation by the DNSO.99

The Telstra case involved the precise scenario described in the quote-an individual purchased the domain name <www.telstra.org> but did not create a website at the address over the two-year period that the domain was registered in his name.100 Because the registrant was not using the domain name for any purpose, the complainant risked losing in arbitration because it could not establish bad-faith use.101 Even though the panelist who decided the case quoted the advisory panel passage cited above and displayed a clear understanding of Paragraph 4(a)(iii)'s requirement of registration and use, he nonetheless decided:

[T]he relevant issue is not whether the Respondent is undertaking a positive action in bad faith in relation to the domain name, but instead whether, in all the circumstances of the case, it can be said that the Respondent is acting in bad faith .... [I]t is possible, in certain circumstances, for inactivity by the Respondent to amount to the domain name being used in bad faith.102

This analysis completely voids the intended effect of the UDRP provision requiring registration and use by suggesting that inactivity constitutes "use." While the arbitrator's analysis may produce a fair result that promotes the underlying policy upon which the UDRP is based, it is contrary to the UDRP's plain language and the wording of the UDRP advisory committee. Because the advisory committee explained that the registration and use requirement represents a compromise between competing interests, it is inappropriate for IMAGE FORMULA 282

panelists to unilaterally overrule the UDRP's wording. Indeed, nothing in the UDRP Rules states that panelists have the power to modify the UDRP's terms. Because they lack this power, it is unfair for individual panelists to rewrite contractual terms that were arrived at through a public process that ultimately reached a rough consensus regarding the proper scope of the UDRP.

Panelists have also demonstrated their willingness to ignore the UDRP's language with the "preclusion" doctrine.103 The preclusion doctrine appears to have been first adopted in J Crew International, Inc. v. crew.com,104 where the panelists determined that by choosing to register a domain name identical or confusingly similar to a trademark before the trademark owner, the registrant had displayed bad faith because his doing so prevented the complainant from having the name.105 The majority of the crew.com panelists found:

Respondent's registration prevents Complainant from using the <crew.com> or <j.crew.com> domain names corresponding to Complainant's registered trademarks. We recognize that Complainant has registered the domain name <jcrew.com>. However, the ability of the Complainant to obtain alternate domain names should not make this provision inapplicable. Otherwise, the provision would always be inapplicable for it would nearly always be possible for the Complainant to obtain an alternative domain name or even to register the same name as its trademark in another gTLD or ccTLD. 106 IMAGE FORMULA 285

Since domain name registrations are by definition exclusive, this analysis is not particularly insightful. However, the effect of the panelists' decision is far-reaching because it displays a willingness to find that any registration that has the effect of preventing a trademark owner from reflecting its mark in a corresponding domain name constitutes bad faith. The UDRP does state that it constitutes bad faith if the domain name registrant "registered the domain name in order to prevent the owner of the trademark or service mark from reflecting the mark in a corresponding domain name, provided that you have engaged in a pattern of such conduct."107 However, the language "engaging in a pattern of conduct" and "in order to prevent a trademark owner from registering a domain name" clearly suggests that the complainant must demonstrate that the registrant intended to block the trademark owner. Once the intent element is added, the rule makes sense; if the registrant registered every conceivable variation of <jcrew.com> to block the J. Crew Corporation from registering them, it would have a strong claim that the registrant was acting in bad faith. In contrast, without the intent element, every trademark owner could establish bad faith simply by alleging that it wanted to register a particular domain name but was barred from doing so because someone else had already (innocently) done so. This interpretation of the UDRP favors trademark owners too much.

The tendency of panelists to resort to rationales like the nonuse or preclusion doctrines is disturbing because it indicates confusion about the scope of the UDRP. A better approach would be to exercise restraint by recognizing that the UDRP is not equipped to deal with every complaint that is filed. For example, in Ermenegildo Zegna Corp. v. Estco Enterprises Ltd.,108 the panelist found that the domain name in question was confusingly similar to the complainant's trademark and that the registrant had no legitimate right or interest in the domain <www.zegnafashion.com>.109 Nevertheless, because the complainant could not show bad-faith registration or use, the panelist dismissed the complaint, adding, "If, once the domain name are [sic] used by the Respondent, it appears IMAGE FORMULA 287

that likelihood of confusion arises with the trademarks of the Complainants, the Complainants could of course start legal action on the basis of trademark infringement."110 In another admirable demonstration of restraint, a different UDRP panelist wrote:

"Bad faith" under the UDRP is a term of art. It does not reach every use of a domain name that might constitute bad faith in the ordinary sense of [the] term . ... The Respondent may be acting unfairly. He may be engaged in unwarranted disparagement. He may be acting childishly. He may be retaliating for having lost earlier Cybersquatting cases. But this does not necessarily mean that he may be forced to transfer the accused domain name to the complainant under the UDRP, considering the purpose.111

Other panelists should follow this example. If the UDRP actually determines the outcome of the case, arbitrators should let the UDRP's language dictate the result. Where the UDRP does not determine the outcome, arbitrators should point out the UDRP's limited scope and dismiss the complaint.

F. Statistical Research on UDRP Decisions

Effects of the UDRP's procedural weaknesses on the fairness of the UDRP process were revealed by two recent statistical analyses of UDRP outcomes.112 The studies examine why complainants prevailed in approximately 80% of the 4,288 cases resolved to IMAGE FORMULA 293

date.113 Both studies conclude that forum shopping and complainant bias help explain the complainants' success rate. 114

The first study, Rough Justice: An Analysis of ICANN's Uniform Dispute Resolution Policy, was conducted by Professor Mueller and published in late 2000.115 It found that the dispute resolution providers that exhibit the strongest complainant bias also attract the largest number of complainants.116 Professor Mueller suggested that the reason for the bias is the way in which different dispute resolution providers' panelists interpret the UDRP:

The reason WIPO and NAF have higher name transfer rates is that many of their panelists ... first form conclusions about whether a registrant is a cybersquatter of some sort. If they believe that the registrant is a bad actor, they tend to stretch the UDRP definitions to cover the a [sic] particular facts of the case, in many cases coming up with highly imaginative definitions of ... policy criteria. In some of the worst decisions, the panelists appear to ignore both the language and the intent of the policy and set themselves up as judges of who has a more meritorious claim on a name. The eResolutions panelists, on the other hand, (in a statistical sense) tend to adopt a more literal reading of the policy .... NAF and WIPO panels have a demonstrable tendency to expand the strict language of the policy when they conclude that a registrant is a bad actor."117

Professor Mueller's findings were recently confirmed by Professor Michael Geist in a report titled Fair.com?: An Examination of the Allegations of Systemic Unfairness in the ICANN IMAGE FORMULA 297

UDRP,118 that analyzes UDRP data collected through July 7, 2001. Using additional data, Geist's analysis produced results nearly identical to Dr. Mueller's: complainants win 82.2% of the time with WIPO, and 82.9% of the time with NAF, but only 63.4% of the time using eResolution.119 WIPO still commands 58% of the UDRP caseload, despite the highest fees of the three dispute resolution providers, compared to 34% for NAF and 7% for eResolution.120 Interestingly, Professor Geist's analysis suggests that complainant bias is due to the manner in which arbitration panels are chosen. He found that complainant bias is worst (complainants win nearly 83% of the time) when dispute resolution providers control who decides a case, as they do for single-panelist decisions.121 In contrast, threemember panels, composed of panelists identified by complainants and respondents, favor complainants only 60% of the time.122 Professor Geist suggests that the reason single-member panels exhibit complainant bias is that dispute resolution providers tend to steer a majority of the 90% of cases decided by only one panelist toward complainant-friendly panelists. 123

The studies by Professors Mueller and Geist are an indictment of the UDRP's procedural provisions for their failure to prevent or correct complainant bias. The studies also seem to imply that dispute resolution providers and panelists are intentionally deciding cases for complainants. While this is possible, neither Professor Mueller nor Professor Geist offers direct evidence showing that panelists are intentionally skewing results. It seems more likely that complainant bias is the result of two interrelated factors previously discussed: the UDRP's language is inadequately employed in determining the outcome of cases, allowing subjectivity to creep in, while the UDRP's procedures fail to prevent mistakes or to correct them when they occur. IMAGE FORMULA 299

G. Possible Remedies for UDRP Weaknesses

The procedural weaknesses of the UDRP leave the entire dispute resolution process in bad shape by failing to provide a mechanism for settling open issues panelists face every day. As a result, panelists find themselves stymied in their attempts to apply the UDRP fairly and forced to resort to personal predilection or overreliance on other panelists' decisions. Bias, mistakes, and poorly reasoned decisions characterize the interpretation of the UDRP. This cannot be as good as it gets. The UDRP can be applied fairly and efficiently to resolve domain name disputes in a manner that lives up to the purposes it was designed to achieve. It is not too late for the UDRP to become a model for dispute resolution in other areas of the Internet. However, to achieve its goals, the UDRP must be modified.

Professors Mueller and Geist have already suggested several procedural changes to reduce complainant bias. Professor Mueller's suggestions include random selection of resolution service providers, the creation of an appeals process, or domain name registrar selection of resolution service providers.124 Professor Geist recommends mandatory three-member panels, caseload minimums and maximums, new quality control mechanisms to review panelists' performance, and greater transparency of data regarding the UDRP.125 All of these suggestions have merit, especially the use of mandatory three-member panels, but they are inadequate in that they do not provide UDRP panelists with better guidance in actually resolving disputes.

In addition to some or all of the procedural changes recommended by Professors Mueller and Geist, ICANN should make changes to the UDRP's procedural framework that are specifically designed to assist panelists in their work. One way to accomplish this goal (without adding unduly burdensome layers of procedure that an appeals process would create) is to create a UDRP Review Board composed of a half-dozen or more UDRP panelists to identify UDRP decisions that deserve to be treated as precedent. The goal would be to create the infrastructure for the concrete IMAGE FORMULA 302

determination of the meaning of confusing sections of the UDRP through the identification of model interpretations of the Policy that other panelists should emulate.

One model ICANN could adopt is the U.S. immigration system's Board of Immigration Appeals. The Board periodically selects cases to be published, which constitute precedent that binds the Board, the Immigration Courts, and the INS.126 Decisions selected for publication resolve issues of first impression, clarify existing rules of law, reaffirm existing rules of law that have been called into question, resolve conflicts of authority, or discuss issues of significant public interest. 127

By adopting this model, ICANN would accomplish several important things. First, by making the UDRP Review Board's decisions binding precedent, ICANN would provide an authoritative source for arbitrators to turn for guidance. Currently, arbitrators have no reliable source for guidance in interpreting the UDRP, so it is no surprise that they make little or no effort to avail themselves of the large amount of commentary and criticism regarding UDRP decisions that exists.128 Second, a UDRP Review Board could effectively distill the meaning of the Policy without creating a cumbersome appeals process by selecting UDRP decisions that were correctly decided as precedential on its own initiative. The Board could also accept a limited number of appeals on a discretionary basis. Parties to disputes would retain the right to bring suit in a court of competent jurisdiction, so the unavailability of appellate review in each case would not jeopardize parties' rights. Third, creating and implementing the Review Board would involve little cost or structural reform; ICANN could appoint the Board members and easily ensure that the Review Board's decisions were followed by simply decertifying arbitrators who repeatedly ignored precedents (of course, the possibility that arbitrators would ignore the Board's IMAGE FORMULA 304

decisions is extremely unlikely, as arbitrators appear desperate for more guidance in applying the UDRP). Perhaps best of all, the Review Board could nail down the meaning of the UDRP's provisions without sacrificing the speed and low cost that make it so popular.

The effect of creating a UDRP Review Board would be to dramatically improve UDRP decisions so that they are more predictable, they reduce complainant bias, and they strengthen the UDRP's legitimacy. It is even possible that the number of UDRP complaints that are filed might drop, because it would be easier to discern what constitutes acceptable domain name registration behavior and what does not. Most importantly, these changes could help ICANN prove that the UDRP is a success that is worth keeping and worth replicating in other areas where litigation fails to serve stakeholders' needs.

IV. RECONSIDERING THE UDRP's INCENTIVE STRUCTURE

From a policy standpoint, the UDRP should operate in an incentive-neutral manner that makes it easy for complainants and respondents to resolve disputes fairly and impartially. It should also be structured to resolve disputes in a manner that prevents misuse of the UDRP by trademark interests or domain name rights supporters while furthering the policy goals it was created to achieve. Therefore, if it is working properly, the UDRP should provide disincentives for engaging in cybersquatting in order to further the goal of reducing or eliminating abusive domain name registration, and it should provide disincentives for bringing frivolous complaints against innocent domain name registrants to prevent the UDRP from becoming a tool for trademark owners to wrongfully seize domain names. Anecdotal evidence suggests that the UDRP currently fails to live up to this standard.

A. Incentives for Cybersquatting

Considering that approximately eighty percent of the UDRP decisions handed down over the past two years favored complainants, it would seem that cybersquatting must be on the decline.129 Surprisingly, the opposite seems to be true: in April, 2002, more than 430 cases were opened under the UDRP-a monthly pace that would put 2002 on track to generate more than twice the number of cases brought under the UDRP in 2000 and 2001,130 This surprising trend raises the question of why domain name disputes are increasing and whether anything can be done to reverse the trend. Several factors may contribute to the UDRP's heavy caseload, including the rollout of new TLDs and ongoing uncertainty about what actually constitutes cybersquatting. However, cybersquatting is essentially a profit-driven phenomenon, and its ongoing profitability lies at the heart of its continued vitality. Therefore, it is worth considering why cybersquatting remains profitable and whether the implementation and use of the UDRP has anything to do with cybersquatting's continued profitability.

1. Cybersquatting Incentives and the ACPA

The best benchmark against which to compare the UDRP's incentive structure is the Anticybersquatting Consumer Protection Act,131 which was enacted at approximately the same time as the UDRP with the same goal: to resolve domain name registration disputes. Absent the UDRP, cybersquatters might not remain in business, because trademark owners would litigate far more domain name disputes. Under the ACPA, alleged cybersquatters would be forced to retain an attorney to defend themselves or risk the unsavory prospect of a default judgment and money damages being awarded against them (compared to the relatively simple UDRP, navigating the intricacies of the Federal Rules of Civil Procedure and the Lanham Act is no easy task, and an alleged cybersquatter would be foolish to opt for a pro se defense). Assuming many alleged cybersquatters would choose to retain legal counsel, they would incur significant defense fees. Furthermore, alleged cybersquatters who lost at trial would face potentially devastating money judgments IMAGE FORMULA 313

comprised of statutory damages of $1,000 to $100,000 per domain name, court costs, and its opponents' attorney's fees, all in addition to losing the disputed domain name itself.132 These penalties are too great to ignore, even if some trademark owners would forego litigation due to the time and expense involved, thereby reducing the chances of a ruinous judgment.

2. Cybersquatting Incentives and the UDRP

Under the UDRP, which is now the domain name dispute resolution vehicle of choice for approximately 98% of complainants,133 the alleged cybersquatter faces a much different risk scenario. In fact, the UDRP makes cybersquatting nearly risk-free, because the UDRP is free to the respondent unless it opts for a threemember arbitration panel.134 Therefore, the alleged cybersquatter faced with a UDRP complaint has plenty of options: he may default and lose only his thirty-dollar registration fee; he may represent himself in the UDRP proceeding for free (this is a viable option because the UDRP is much simpler than litigation under the Lanham Act in federal court); or he can hire an attorney to defend the domain name. No matter what choice he makes, the alleged cybersquatter's liability is limited to the cost of his legal services, his thirty-dollar registration fee, and one-half of the cost of a three-member arbitration panel (assuming he selects this option). In other words, because the UDRP does not award any other type of relief besides the transfer or cancellation of the domain name, the alleged IMAGE FORMULA 317

cybersquatter's risks under the UDRP are limited and easily quantifiable. 135

3. Comparing Incentives Under the UDRP and ACPA

The three primary UDRP dispute resolution providers charge between $750 and $1,500 for a single panelist to resolve a dispute regarding a single domain name.136 A rational complainant who is aware of these costs will likely be willing to pay up to $1,500 to an alleged cybersquatter to obtain a domain name before entering a formal complaint under the UDRP or ACPA to save time and money.137 As a result, the alleged cybersquatter stands to make a profit of up to $1,470 per domain name (assuming that it costs thirty dollars on average to register a domain name and that the registrant warehouses his domains or "points" them all at a single web page to minimize his other costs). This is easy money by any standard. Even if the alleged cybersquatter lost nine out of ten domain names in UDRP proceedings and sold the tenth domain name for a measly $750, the cybersquatter could still be ahead as much as $450 (assuming the cybersquatter has no other costs besides domain name registration). In contrast, under the ACPA, even one adverse judgment per year could easily erase an entire year of profits and send the cybersquatter into bankruptcy. This rough profit/loss analysis suggests that cybersquatting continues to plague trademark owners partly because the UDRP reduces the costs incurred by cybersquatters enough to make their speculative activities profitable on an ongoing basis. IMAGE FORMULA 321

4. Relief Requires Eliminating Cybersquatters' Profits

ICANN should modify the UDRP to eliminate the economic shield it provides to cybersquatters. One approach ICANN should consider is increasing the cost of abusive domain name registration by requiring each domain name registrant to post a $350 refundable deposit that would be forfeited if the domain name were successfully challenged during the first year of registration (see Appendices A and B for a rough analysis of the economic impact on cybersquatters of imposing such a fee). Considering that the average UDRP complaint is filed one year and three months after it was registered,138 a one-year deadline would balance the complainant's need for adequate time to discover and challenge abusive registrations with the respondent's desire to lay out no more money than necessary to obtain a domain name. After one year, the deposit could be used to defray ongoing domain name registration costs or could be refunded to the registrant. Raising fees in this manner would admittedly prevent some Internet users from establishing their own web sites, but this drawback is worth accepting to prevent abuse of the domain name registration system.

B. Deterring Reverse Domain Name Hijacking

An equally serious problem with the incentive structure created by the implementation and use of the UDRP is that it currently imposes no penalty on complainants for abusing the UDRP process. While Paragraph 15(e) of the UDRP Rules does include a provision that allows UDRP panelists to declare that a complaint was brought in an attempt to engage in reverse domain name hijacking, the Rules fail to provide a sanction for this type of behavior. 139 This is somewhat surprising, because the risk of reverse domain name hijacking was recognized before the UDRP was implemented,140 and IMAGE FORMULA 326

it is common for legal systems to impose sanctions for abusive legal action.141

Since anecdotal evidence suggests that reverse domain name hijacking is a problem under the UDRP, 142 the UDRP should be modified to protect innocent domain name registrants. One of the simplest ways to deter reverse domain name hijacking would be to include a provision in Rule 15(e) that would bar complainants found guilty of reverse domain name hijacking from invoking the UDRP for at least one year and terminate any other UDRP proceedings that the complainant currently had underway.

This policy would be easy for ICANN to implement, and it would create a strong incentive for trademark owners to consider whether they were bringing complaints in bad faith before filing them. Additionally, this rule could have a significant impact on the frequency with which reverse domain name hijacking occurs, because large trademark owners, who are most likely to engage in reverse domain name hijacking, appear to rely heavily on the UDRP for protection of their trademarks. For example, a search of UDRP proceedings to date returns 162 entries for "Nike." 143 If other large corporations rely on the UDRP to a similar extent, being banned from using it to resolve domain name disputes would be a significant penalty. On the other hand, this rule is not too draconian, because complainants have alternative (albeit slower and more costly) means for protecting their trademarks, such as the ACPA.

V. CONCLUSION

This Note argues that the UDRP's procedural flaws must be repaired in order for the UDRP to achieve its goal of providing speed, low cost, and fairness of results. It suggests that ICANN can fix this weakness by establishing a UDRP Review Board to identify IMAGE FORMULA 330

important UDRP decisions as binding precedents that all arbitrators would be required to follow. It also proposes revising the fee structure for domain name registration to reduce or eliminate cybersquatting's profitability and modifying the UDRP Rules to impose penalties for complainants who attempt to engage in reverse domain name hijacking. Enactment of these proposals would assist the UDRP in reducing or eliminating the practice of cybersquatting while simultaneously protecting innocent domain name registrants from overly aggressive complainants.

One issue this Note has not addressed is whether the changes it proposes are feasible. Theoretically, the answer is "yes," because ICANN possesses the authority to change both the UDRP and domain name registration procedures. However, ICANN's fractured board of directors is unlikely to mandate changes to the UDRP unless the public demands them. Therefore, it is up to external stakeholders to lobby for improvements to the UDRP that will help it realize its goal of offering a global dispute resolution process that is cheap, quick, and fair. Whether the UDRP is actually repaired will be a good test of the long-term feasibility of Internet selfgovernance. If the global community of Internet users pressures ICANN to ensure that changes are made, it will be a testament to the fact that a market-oriented legal regime operating largely beyond the oversight of national judicial systems can work effectively.

IMAGE TABLE 334

APPENDIX A:

IMAGE TABLE 335

APPENDIX B:

AUTHOR_AFFILIATION

Orion Armon*

AUTHOR_AFFILIATION

* Candidate for Juris Doctor, The University of Texas School of Law, 2003; B.A., summa cum laude, The University of Colorado at Boulder, 1997. I wish to express my gratitude to Professor R. Anthony Reese for his insightful comments on drafts of this Note.