John Selby, Lecturer, Division of Law, Macquarie University
This paper examines the process implemented by ICANN for resolving domain name disputes, the UDRP, has worked in practice. It quantitatively and qualitatively assesses whether the competitive goals set out as reasons for the UDRP are being achieved. This paper is an extract of a more detailed analysis.
In 1999, the Internet Corporation for Assigned Numbers and Names (“ICANN”) implemented a Uniform Dispute Resolution Policy (“UDRP”) designed to resolve the pressing issue of cyber-squatting. Prior to the development of the UDRP, the World Intellectual Property Organisation (“WIPO”) was asked to prepare a report on the best procedural method to use when resolving such disputes. WIPO’s report suggested that the features of the Domain Name System (“DNS”) should be reflected in any such dispute resolution process. This meant that the process had to be quick, cheap and able to handle a high volume of complaints. On that basis, the UDRP was drafted to include a quasi-arbitral form of dispute resolution with multiple private sector service providers (“Providers”).
As the bodies that would provide such dispute resolution services would operate in the private sector, WIPO recommended that they should compete against each other. ICANN’s policy organ, the Domain Name Supporting Organisation (“DNSO”), recommended to ICANN that WIPO’s position be adopted.1 HREF1
Therefore, the UDRP established the means by which Providers compete. Through the UDRP, ICANN required that Providers appoint skilled and impartial individuals (“Panelists”) to resolve domain name disputes. Trademark owners (“Complainants”) that seek to reclaim a domain name from a third party that has already registered that domain name on the DNS (“Respondent”) must select a Provider with whom to lodge their complaint. The Complainant must pay the Provider’s entire fee unless the Respondent seeks to have three Panelists determine the dispute. The Panel will resolve the dispute by applying the UDRP and if the Respondent is found to have been cybersquatting, the Panel will order the appropriate registrar to transfer contractual control of the domain name to the Complainant.
Public comments received in relation to the First ICANN Staff Report would suggest that at least some participants in the process did not believe that the Uniform Dispute Resolution Policy would lead to the creation of a market that distinguished Providers on price alone. Instead, commentators claimed that competition between providers should operate at two levels, competition on price and competition on service offerings.2
Unfortunately, the commentators did not specify the means by which they foresaw Providers competing over service offerings. Presumably, a Provider could distinguish their service offering on either or both quantitative or qualitative bases.3 HREF2
The article (an extract of a more detailed analysis) will compare three quantitative and two qualitative means by which a Provider could differentiate their service offerings and therefore compete against the other Providers. Quantitative means of Provider differentiation include: the number of prior complaints resolved by a Provider; the speed by which Panels appointed by that Provider reach decisions and the overall success rate of Complainants that have previously used that Provider. Qualitative means of Provider differentiation include: the perceived reputation of each Provider and the extent to which the Provider utilises a core group of experienced Panelists to resolve the majority of its disputes.
Complainants have not demanded the services of each of the four Providers at an equal rate. A single Provider, WIPO, has resolved over 60% of all domain name disputes made between 1999 and 2002. At the other extreme, the CPR has been accepting complaints for only six months less than WIPO, yet it has only received 55 complaints (0.74% of all domain name disputes).4 HREF3 Given the broad range of factual issues and different defences that can arise during a domain name dispute, a Complainant may feel more comfortable that a random WIPO Panelist would be more likely to have practical experience through drafting decisions on that topic than a CPR Panelist.
This attitude can create a catch-22 situation for Providers with low market share. If potential Complainants are reluctant to make complaints to a Provider whose Panelists have not had the experience of resolving a significant number of prior complaints, that Provider may never achieve the critical mass of market share that it requires to attract many complaints.
The average time taken to resolve a domain name dispute under the UDRP is 43 days.5 HREF4 As the table below shows, the Providers that resolve disputes faster than average do have larger market shares. The NAF resolves disputes in an average of 37 days and has 35.90% market share. WIPO resolves disputes in an average of 45 days and has almost 60% market share.6 Information is not available on the average time taken by CPR or ADNDRC Panels to resolve disputes. Prior to its bankruptcy, eResolution averaged 55 days to resolve a dispute.
Comparison of Provider Market Share against Average time to resolve disputes
| Provider |
Average Time to resolve a dispute (days) |
Market Share |
| WIPO |
45 |
59.06% |
| NAF |
37 |
35.90% |
| eResolution |
55 |
<8% (at time of bankruptcy) |
| ADNDRC / CPR |
unknown |
<1% |
From this information, although the time differences are not that great between the slowest and fastest Provider’s Panels, it would appear that there is a slight correlation between the time taken to resolve a dispute and the market share of each Provider. This does not mean necessarily that a Provider with a history of slower than average decision making Panels would actually discourage potential Complainants from choosing to use its service.
For commercial reasons, some Complainants are desperately keen to recover their domain names. On the other hand, the average time elapsed between the registration of a domain name and the challenging of that registration under the UDRP was 477 days in 2000.7 Thus, an 18 day difference between the average times required to resolve a dispute appears to be particularly insignificant compared to the almost 16 month period prior to the complaint being lodged. More research into actual Complainant preferences is necessary to determine whether this issue actually influences Provider selection.
Therefore, it is unclear whether Provider market share is significantly affected by the average time taken by that Provider’s Panelists to resolve a dispute. It may be advisable for a Provider seeking to increase its market share of disputes to encourage its Panels to resolve disputes promptly (whilst still maintaining the quality of the decisions made) as this could lead to higher Complainant and Respondent goodwill.
In “Fundamentally Fair.com? An Update on Bias Allegations and the ICANN UDRP”, Professor Michael Geist stated that across all single-Panelist non-default cases from 1999 till 18 February 2002, Complainants won 70% of decisions made by WIPO Panelists, 69% of decisions made by NAF Panelists and 50% of decisions made by eResolution Panelists8 HREF5. Given that WIPO and NAF have resolved over 90% of all domain name disputes and eResolution is in bankruptcy, this reveals a clear correlation between Provider market share and the single-Panel Complainant success rate.
The Complainant success rate in three member Panel non-default disputes did not correlate with Provider market share. 48% of such Panels appointed under WIPO, 42% of NAF Panels and 47% of eResolution Panels, respectively, found in favour of the Complainant.9
The percentage of potential Complainants that actually consider the success rate of previous complainants under each Provider before deciding which Provider to select is unknown. This is an issue that warrants further study. In single-Panel disputes, Panelists are appointed by their Provider rather than selected by the parties to resolve the dispute. In such situations, a Provider could not directly control its Complainant success rate. It is interesting that when a Provider has no control over its Complainant success rate, such as in three-person Panel disputes where the parties select the Panelists, Provider market share does not correlate with Complainant success.
It is possible that a Provider that sought to encourage potential Complainants to use its dispute resolution service (ie increase its market share) could indirectly influence its Complainant success rate by directing a greater number of complaints to sole-Panelists who had previously resolved a high rate of disputes in favour of Complainants. The way in which Providers allocate disputes amongst Panelists will be discussed more below.
Whether publicising a Provider’s Complainant success rate would cause forum-shopping by Complainants (as Geist suggests10) is unknown. There is a clear correlation between Complainant success in single-Panel disputes and Provider market share. Whether there is a causal link between those two statistics is a matter for further investigation. If such a causal link were found, then the historical success rate of Complainants in single-Panelist disputes would be a clear means for a Provider to distinguish itself in a competitive market.
The perceived stature, experience and history of a Provider could also affect Complainant demand for its dispute resolution service. As discussed above, a new Provider would face the catch-22 situation of potential Complainants being unwilling to try its untested service, which would preclude the development of a substantial body of resolved complaints through which Complainants could develop the trust necessary to use the service.
There would appear to have been definite first-mover advantages for those Providers that were initially approved by ICANN in 1999 / 2000. The pent-up demand amongst Complainants that led to the development of the UDRP ensured that the earliest-appointed Providers were able to develop significant technical experience and a body of resolved complaints that could be used to distinguish their services from those provided by newer, less experienced, Providers.11
John Berryhill12 HREF6 has written an interesting, if slightly tongue-in-cheek, analysis of the implicit incentives for Complainants (and their lawyers) within the UDRP. He states that "Clients do not pay their attorneys to obtain "justice". Clients pay attorneys to win. The last thing the [attorney] wants in an arbitrator is someone who is going to think. The [attorney] wants someone who is going to listen to be persuaded to find in the client's favour.”13 A practical attorney would not, "... pick an arbitration provider wherein at least one of the members is known to have some kind of association with material critical of the interests of trademark holders with regard to the UDRP." Especially not, "... to save $250".14
As Complainants choose which Provider will resolve their dispute, Providers are implicitly encouraged to offer a service which is attractive to potential Complainants. Berryhill cites the [un-named] Provider advertising on the index page of their website that a "cybersquatter has been evicted from a domain name". Berryhill quotes another Provider that "boasts of how they are the first to "go after cybersquatters" with their online filing system".15
Professor Michael Geist has claimed that Providers do attempt to distinguish their service offerings through marketing techniques. He received ten press releases from the NAF in a four month period in 2001, all of which promoted decisions in which Complainants won. These featured headlines such as “Arbitrator Delivers Internet Order for Fingerhut” and “May the Registrant of magiceightball.com Keep the Domain… Not Likely”16 HREF7. The reputation and status of eResolution, the only Provider to have fallen into bankruptcy since the UDRP commenced, may shed light on the nature of the Provider market. Prior to its closing, eResolution had appointed a large number of university professors from the USA and Canada as Panelists, with few Panelists located in Europe or Asia.17
Unfortunately, eResolution's original website is no longer operational.18 However, the WayBack Machine (http://www.archive.org) has retained copies of a number of different versions of eResolution's website19 HREF7.The earliest copy is dated April 7, 2000 which is four months after eResolution was appointed by ICANN as a Provider20 HREF8. At that time, eResolution had appointed 49 Panelists, twenty-five of whom were academics and twenty-three practicing lawyers21 HREF9 Seventy-three percent of eResolution's Panelists were US-based, with the remaining thirteen Panelists spread across ten other countries. The Complainant market regarded eResolution as being more Respondent-friendly than the other Providers. This was attributed to eResolution's higher concentration of academic Panelists compared to the other Providers22 HREF10 and the criticism of the UDRP by some of its founders23 HREF11 HREF12.
It would appear that, prior to its bankruptcy, eResolution had made significant efforts to increase the number of practitioner-Panelists on its roster. On the most recent copy of eResolution's website stored by the Wayback Machine24 HREF13 (dated September 26, 2001), eResolution had increased its roster of Panelists from 49 to 100. Virtually all of these additional Panelists were practitioners, not academics. In fact, 76% of eResolution's Panelists were practitioners at that date. eResolution had also widened its Panelist base to include members from 18 countries. Non-US based Panelists increased their representation from 27% to 48% in only sixteen months25 HREF14
E-Resolution’s changing Panelist composition
| Date |
Academic Panelists |
% Academic Panelists |
Practitioner Panelists |
% Practitioner Panelists |
Total Panelists |
| 7/4/2000 |
25 |
51% |
23 |
47% |
49 |
| 26/9/2001 |
24 |
24% |
76 |
76% |
100 |
Unfortunately, the changes made by eResolution to its Panelist base were not sufficient to improve its standing within the Complainant market. eResolution was only able to attract 288 complaints before its bankruptcy in November 200126 HREF15. The criticisms of ICANN and the UDRP that were made by academics involved in the management of eResolution and statistical analysis which revealed that it had the lowest rate of Complainant-success of the four Providers (at a mere 61%) appear to have influenced potential Complainants (and their advising legal counsel).27 Those Complainants voted with their feet, took their complaints to other Providers and starved eResolution into bankruptcy less than two years after it opened.
As the discussion above has highlighted, the reputation of a Provider would appear to be a significant factor in determining the market share of that Provider. It would therefore appear to be a significant factor that drives competition between Providers.
At least one Provider, NAF, has claimed that it randomly appoints Panelists to determine complaints.28 It is therefore worthwhile to examine whether Providers tend to favour the appointment of some Panelists and ignore others, or share their overall complaint workload relatively evenly.
Panelists that have been appointed more often when working with one Provider would appear to remain popular when they are appointed by a subsequent Provider. Of the forty-three Panelists appearing on the rosters of both the NAF and WIPO, six names appear on the list of the top-10 most commonly appointed dual-listed Panelists by each of those Providers. In fact, the most commonly appointed dual-listed Panelist for each of the NAF and WIPO was the same person, Alan Limbury.29 Those top-10 most commonly appointed dual-listed Panelists have between them been appointed to 85% of the NAF complaints and 73% of the WIPO complaints resolved by Panels containing dual-listed Panelists.30
Conversely, a dual-listed Panelist that has not been appointed to resolve many disputes by WIPO is unlikely to be appointed to resolve disputes by the NAF. Of the 43 dual-listed WIPO – NAF Panelists, 26 have not been appointed by the NAF to resolve any disputes.31 One dual-listed Panelist has never been appointed by either WIPO or NAF to resolve a domain name dispute.32 Those same 26 dual-listed Panelists who had never been appointed by the NAF to resolve a dispute were appointed by WIPO to make 24% of the 862 panel decisions made by dual-listed WIPO Panelists.33 This equates to one fewer WIPO panel appointments than have been resolved by the two most commonly appointed dual-listed WIPO and NAF Panelists.34
eResolution appears to have distributed its decisions more widely amongst its Panelists than some of the other Providers. No single eResolution Panelist had decided more than 2.8% of the 288 complaints resolved by that Provider before its bankruptcy.35 The six most commonly appointed eResolution Panelists had decided 16.3% of all complaints received by that Provider and thirty eResolution Panelists made 61% of all decisions. This contrasts with six NAF Panelists who had decided 53% of all NAF decisions and the six WIPO Panelists who had decided 17% of all WIPO decisions (as at 7 July, 2001).36
Many former eResolution Panelists can now be found on the lists of other Providers. Of the fifty-eight Panelists that currently appear on more than one Provider's roster, fourteen were also Panelists with eResolution37 HREF17. Mr Chiasson Q.C. of Canada is an eResolution alumnus, which means that he has served as a Panelist with every UDRP Provider. Of the six Panelists that are currently on three Provider rosters, half were also eResolution Panelists.38
This essay has examined whether ICANN’s use of competing Providers to resolve domain name disputes under the UDRP has actually achieved the goals that were enunciated when the UDRP was being developed. It has revealed that whilst the UDRP has been a low cost, high speed and high volume method of resolving domain name disputes, Providers may compete with each other on a number of quantitative and qualitative non-price bases.
Comments made by the public during the development of the UDRP revealed an awareness of the fact that non-price competition between Providers could arise. Given that Provider market share is currently inversely related to price, such comments would appear to have been prescient.
The experience gained from resolving a significant number of disputes would appear to be one way in which Providers can distinguish their service offerings. This will make it difficult for newer, less experienced Providers to compete effectively. There does not appear to be a significant difference between the amount of time it takes for each Provider’s Panelists to resolve a domain name dispute.
As all current Providers offer a service significantly faster than the substitutable forms of dispute resolution, especially litigation, it may be difficult for Providers to differentiate their offering on the basis of the speed with which disputes filed with them are resolved. Anecdotal evidence suggests that a Provider could raise its market share if it could encourage Registrars to implement its decisions faster than is currently the case.
The rate at which Complainants succeed when using a particular Provider would appear to be a significant factor in determining Provider market share. However, as Panelists make decisions rather than Providers, it would be of concern if Providers were to attempt to compete with each other on the basis of their respective Complainant success rates. Panelists are required to decide their disputes with impartiality and integrity. If it was revealed that Providers were attempting to compete by influencing Panelist decisions, or deliberately only appointing Panelists due to their history of Complainant success, such competition could bring the integrity of the UDRP into question.
Whilst the majority of domain name disputes arise from only a few countries (particularly the USA), Providers appear to have competed on the basis of the geographical breadth and depth of their rosters. The extensive cross-fertilisation of Provider rosters which has occurred due to ICANN’s prohibition on Providers requiring Panelists to only work for them has made it harder for such competition to eventuate.
Promoting the background and experience of their Panelists would appear to be a more effective method for Providers to compete. However, if this message does not reach its target audience (Complainants) effectively, significant gains in market share may not eventuate. The demise of eResolution is evidence of this.
The newest Provider, ADNDRC, has appointed a significantly different Panelist roster than its competitors. ADNDRC appears to be expecting a significant rise in the number of disputes requiring specialist Chinese, Hong Kong or Korean trade mark expertise. If such an increase does not eventuate, ADNDRC may be forced to change its strategy to compete in a more traditional manner with the other Providers, or face bankruptcy as eResolution did.
A Provider’s perceived reputation in the Complainant market appears to impact significantly on its ability to compete. Anecdotal evidence suggests that WIPO’s strong international reputation is one reason for its success. Criticisms of the UDRP system by members of eResolution appear to have contributed to the decline in its reputation amongst Complainants and therefore to a decline in its market share. For the purposes of Provider competition, the validity or inaccuracy of such criticisms is unfortunately irrelevant. The way in which Complainants react to such criticism is of paramount importance.
1 ICANN's domain name policy organ, the Domain Name Supporting Organisation (DNSO) agreed with WIPO's Report which stated (at paragraph 225), '[i]t is suggested that the administration and decision-maker's fees should be set freely by the dispute-resolution service providers that may be mandated to administer the procedure. Allowing the institutions to set their own rates should stimulate competition and ultimately benefit the public.' See World Intellectual Property Organisation, The Management of Internet Names and Addresses: Intellectual Property Issues – Final Report of the WIPO Internet Domain Name Process (1999) <http://wipo2.wipo.int/process1/report/finalreport.html> at 27 November 2002.
2 National Basketball Association and MCI Worldcom, above n48.
3 Some of these quantitative and qualitative methods that Providers could use to distinguish their service offerings were briefly suggested in: International Trademark Association, The UDRP by all Accounts Works Effectively: A Rebuttal to Analysis and Conclusions of Professor Michael Geist in “Fair.com?” and “Fundamentally Fair.com?” (2002) [3, 7-8] <http://www.inta.org/downloads/tap_udrp_2paper2002.pdf> at 12 December 2002.
4 ICANN, UDRP decisions listed by proceeding number, <http://www.icann.org/udrp/proceedings-list-number.htm> at 12 December 2002.
5 Milton Mueller, Rough Justice: An Analysis of ICANN’s Uniform Dispute Resolution Policy (2000) [17-18]<http://dcc.syr.edu/miscarticles/roughjustice.pdf> at 1 August 2002.
6 Ibid.
7 Mueller, above n5 at 9.
8 Michael Geist, Fundamentally Fair.com? An Update on Bias Allegations and the ICANN UDRP (2001) [8] <http://aix1.uottawa.ca/~geist/fairupdate.pdf> at 3 August 2002.
9 Ibid.
10 Ibid.
11 Christopher Lee, The Development of Arbitration in the Resolution of Internet Domain Name Disputes (2000) 7 Richmond Journal of Law and Technology 2 at 6.
12 John Berryhill, ‘The UDRP Provides Disputable Resolution Incentives’ (2000) <http://www.icannwatch.org/archive/udrp_and_incentives.htm> at 12 December 2002.
13 Ibid at 1.
14 Ibid at 2.
15 Ibid at 2-3.
16 Geist, Fair.com? An examination of the Allegations of Systemic Unfairness in the ICANN UDRP (2001) [17] <http://aix1.uottawa.ca/~geist/geistudrp.pdf> at 3 August 2002, at 4-5.
17 Lee, above n11 at 6; John G. White, 'Arbitration: ICANN's Uniform Domain Name Dispute Resolution Policy in Action' (2001) 16 Berkeley Technology Law Journal 229 at 232-3.
18 <http://www.disputes.org/eresolution/> now only lists the decisions made by eResolution prior to its bankruptcy – complying with an requirement to publish all UDRP decisions.
19 See <http://web.archive.org/web/*/http://www.disputes.org> for copies of eResolution’s website.
20 ICANN, Approved Providers for Uniform Dispute Resolution Policy (2002) <http://www.icann.org/dndr/udrp/approved-providers.htm> at 12 December 2002.
21 See <http://web.archive.org/web/20000407083258/http://www.disputes.org/>.
22 Michael Geist, Fundamentally Fair.com? An Update on Bias Allegations and the ICANN UDRP (2001) [8] <http://aix1.uottawa.ca/~geist/fairupdate.pdf> at 3 August 2002, at 3.
23 See generally Froomkin, Michael, A Catalog of Critical Process Failures; Progress on Substance; More Work Needed (1999) ICANN <http://www.icann.org/comments-mail/comment-udrp/current/msg00101.html> at 27 November 2002 at 7-8 and Michael Froomkin and David Post, Froomkin and Post Send Letter to ICANN Board (2000) <http://www.icannwatch.org/archive/post_froomkin_udrp_letter.htm> at 12 December 2002.
24 Available at <http://web.archive.org/web/20010926024913/http://www.disputes.org/>.
25 Available at <http://web.archive.org/web/20011020004251/www.eresolution.ca/services/dnd/arbitrators.htm>.
26 ICANN has stored copies of eResolution Panel decisions at: <http://www.icann.org/udrp/Detail155.htm>.
27 See generally Froomkin, above n23; Froomkin and Post, above n23 and Mueller, above n5 at 11.
28 Geist, above n16 at 8.
29 Searches conducted using ICANN’s text-searchable database of UDRP decisions <http://www.icann.org/udrp/udrpdec.htm>.
30 Ibid.
31 Ibid.
32 Ibid – Mr Daniel Peòa of Colombia.
33 Ibid.
34 Ibid.
35 Ibid.
36 Geist, above n16, at 8.
37 Wayback Archive of eResolution’s website as at 21 September 2001, <http://web.archive.org/web/20011020004251/www.eresolution.ca/services/dnd/arbitrators.htm>.
38 ibid.