|
Auctions will play a critical role in ICANN’s new generic Top-Level Domain (gTLD) process, the only question is whether they reinforce ICANN’s position as a steward of a global public resource, or undermine it. Auctions are one of two widely used mechanisms used in the allocation of scarce global resources, the other being a comparative analysis (aka beauty contest). ICANN has traditionally relied upon comparative evaluations in resolving TLD contention sets. While auctions clearly constitute an important piece of the new gTLD allocation process, the current proposal in the Applicant Guidebook is ripe to produce serious unintended consequences. This paper seeks to highlight those potential unintended consequences and propose a more equitable model consistent with ICANN’s role as a steward of public global resource.
HISTORICAL PERSPECTIVE OF COMPARATIVE ANALYSIS IN TLD ALLOCATION
A. .BIZ Allocation
One of the first uses of the comparative evaluation process by ICANN was in connection with the 2000 proof of concept TLD process, when five applicants were seeking the .BIZ string.1 Unfortunately, the original independent evaluator’s analysis did not make specific recommendations regarding string contention resolution. In fact when analyzing the entire group of general purpose TLDs the specific recommendation was “the team believes that the Board could responsibly select a limited number of applications from this group.”2 Lacking specific recommendations from the Independent Evaluators the ICANN Board found itself in the unenviable situation of having to pick winners and losers through their now infamous “in the basket/out of the basket” analysis.3
B. .ORG Allocation
The second use of a comparative evaluation process by ICANN was in connection with the .ORG reallocation RFP.4 The RFP set forth eleven criteria5 by which a team of four evaluators6 ranked each of the individual applicants. In the preliminary ICANN staff evaluation report, it was proposed that “a recommendation to the Board that the ISOC proposal be selected and that the Board authorize the President to proceed to negotiate an agreement based on that proposal.”7 The ICANN Board later voted to approve the proposed registry contract between PIR and ICANN for the operation of the .ORG TLD.
C. .NET Allocation
The next use of a comparative evaluation process was in connection the 2004-2005 .NET RFP in which five applicants submitted responses to become the next .NET registry operator.8 ICANN designated Telcordia as the independent evaluator to analyze the submissions on a set of 31 criteria.9 Telcordia used a color coded ranking criteria of Red (unacceptable); Yellow (has serious flaws or issues); Green (acceptable); and Blue (exceeds requirements).10 The Telcorida report found that all applicants had the capability to run the .NET registry, but identified VeriSign as the highest ranked applicant having two more “Blue” marks than the nearest competitor in the 31 criteria.11 The close and competitive nature of this reallocation was reflected in the ICANN Board vote where eight directors voted in favor of approving VeriSign as the .NET registry operator, four voted against it, and three abstained.12 This fractured vote was in contrast to the .ORG RFP in which the ICANN voted eleven in favor, one against, and three abstentions.
AUCTIONS AND THE ALLOCATION OF SCARCE GLOBAL PUBLIC RESOURCES
The allocation of scarce global public resources has been the topic of extensive economic and public debate. In the current Applicant Guidebook, ICANN proposes the use of an auction process to resolve contention when multiple parties are seeking the same string, except in the case of geographic and community-based applications. While auctions can serve an important resolution mechanism of last resort, ICANN may wish to reanalyze its rush to auction off TLD strings among multiple applicants.
While many people have sought to analogize ICANN’s allocation of contested TLDs to spectrum, there are significant differences. Certain TLDs (.BANK, .DOCTOR, .POLICE etc.) have an inherent and intuitive meaning that Internet users are likely to associate with specific goods and/or services. This is unlike radio spectrum where 94.1 MHZ can be used by a radio station equally for either rock, country, jazz or classical music.
Therefore, there are important public policy considerations that come into play when ICANN is selecting the appropriate steward for such a TLD. This is not a simple equation but a complex n-factorial equation where ICANN must factor in potential multiple meanings, among multiple communities for the same string. In spectrum allocation there are no higher order equations to solve because a specific spectrum frequency is content/service neutral, thus lending itself nicely to a pure auction allocation mechanism.
The following examples are provided to highlight some of the potential unintended consequences of a purely auction-driven resolution model that fails to take into account legitimate public policy considerations.
A. Hypothetical String Contention in Connection with “.BET”
The following two applicants are seeking the .BET TLD:
Applicant A: The Black Entertainment Television, Inc. which has a US national trademark registration for BET and which has filed a community based application for a .BET TLD. In the community-based application, Black Entertainment Television alleges that they intend to propose a hybrid use of the TLD to advance its commercial interests in connection with radio and television productions, as well as as a social network for Black culture globally.
Applicant B: A not-for-profit consortium of casino operators that has applied for the .BET TLD with the intention of enabling second level domain name registrations to provide an online gaming platform.
During the evaluation process, Applicant A only succeeds in scoring a 13 out of 16 on the Community Evaluation Criteria (14 needed to be declared a “clear winner”).13
B. Hypothetical String Contention in Connection with “.GAY”
The following three applicants are seeking the .GAY TLD:
Applicant A: A leading internationally recognized non-profit organization dedicated to human rights advocacy on behalf of people who experience discrimination or abuse on the basis of their actual or perceived sexual orientation, gender identity, or expression. Applicant A files an application for a community-based TLD and proposes to use the .GAY TLD to advance its core goals.
Applicant B: A leading adult pornographer that intends to primarily use the .GAY TLD for the distribution of pornography to the gay and lesbian community.
Applicant C: An internationally recognized and well financed religious organization believes that homosexuality is an illness that can be cured. They have proposed to use the .GAY TLD to help educate gay, lesbians and bisexuals on how they can be “cured” of their illness.
Applicant A failed to score the required 14 point threshold in the Community Evaluation criteria, and thus was not designated a “clear winner.” As a result of fending off a number of frivolous opposition challenges as provided for in the Applicant Guidebook, the non-profit has limited resources to participate in an auction against Applicants B and C, which both have far greater financial resources.
C. Hypothetical String Contention in Connection with “.GOP”
The following three applicants are seeking the .GOP TLD:
Applicant A: Is the Republican National Party which is seeking to register the string .GOP in connection with its long-standing identification as the Grand Old Party to be used for online campaigning, education, and outreach.
Applicant B: Is the city of Gop, Cameroon (population of under 1,000) which is seeking to apply for a City TLD for the benefit of each of its inhabitants.
Applicant C: Is Contextvision, a Swedish company with a US federal trademark registration for the mark GOP, which intends to use the TLD to promote its products on a global basis.
An additional factor that is not publicly known, because the answer to Question 11.C is held in confidence is that George Soros (a notable philanthropist focused on supporting liberal ideals and causes) is the financial backer of the Gop, Cameroon city TLD. Mr. Soros has pledged to commit up to $10 million of his personal funds to secure the .GOP TLD in any type of auction.
D. Hypothetical String Contention in Connection with “.KIDS”
There are two potential applicants for the .KIDS TLD:
Applicant A: A leading internationally recognized non-profit organization dedicated to promoting the protection of children online. Applicant A files an application for a community-based TLD and proposes to use the .KIDS to provide an online safe harbor for children and adolescents. The application sets forth strict criteria of who can register in the .KIDS TLD, and prompt take down mechanisms in connection with questionable content.
Applicant B: An entrepreneur with substantial venture capital funding is seeking to secure the .KIDS to provide a commercial platform to market goods and services to children/adolescents. The applicant has not filed a community application, and the policies are by and large duplicative of the current gTLDs with no safeguards regarding registration and take down mechanisms.
Applicant A failed to achieve the required 14 points in the Community Evaluation criteria, and thus was not designated a “clear winner.” Applicant B knows Applicant A as a non-profit lacks the financial resources to meaningfully participate in an auction, and thus refuses to engage in any meaningful dialog to resolve their dispute knowing that he will win at auction.
POTENTIAL RESOLUTION
While the .ORG RFP demonstrates the strengths of the traditional ICANN comparative evaluation process, the .NET RFP highlights how the use of an auction in a more closely contested allocation could provide a more equitable allocation framework. Recognizing the short comings of a purely auction resolution mechanism in a string contention set, it is proposed that ICANN modify the current Applicant Guidebook to adopt a hybrid solution along the following lines.
A. Comparative Analysis
In a situation where multiple parties seeking the same string are unable to reach an agreement, it is proposed that these applicants be subject to a comparative analysis to be conducted by a neutral third party evaluator retained by ICANN. This comparative analysis would be self-funded by the applicants, and the criteria could be modeled upon pre-establish objective criteria that ICANN has used in the past. In the alternative ICANN could incorporate objective criteria which factor into consideration additional safeguards designed to mitigate malicious conduct and safeguard intellectual property rights.
The third party evaluator would prepare a report providing a numerical score against the given objective criteria, just as currently envisioned in connection community based applications. A “clear winner” would be determined where the top applicant scored at least 20% percent higher than the next applicant, e.g. Applicant A (24 points), Applicant B (18 points) and Applicant C (15 points).
B. ICANN Board Validation
The third party evaluator’s report would then be forwarded to the ICANN Board. In the case of an evaluator’s recommendation identifying a “clear winner” it would require a supermajority of the Board to approve the evaluator’s recommendation (i.e. similar to the Board’s supermajority vote in connection with the .ORG RFP). However, if the Board was unable to reach a supermajority in connection with the evaluator’s recommendation (i.e. similar to the Board’s bare majority in connection with the .NET RFP) the string contention would move to an auction between the parties. If the evaluator’s report concluded that there was no clear winner, it would only require a simple majority of the Board to allow that string contention to move forward to an auction.
OPPOSITION TO THIS PROPOSAL
A. Deep Pocket Applicants
Those applicants with deep pockets seeking the more highly sought generic TLD strings clearly would oppose this proposal. Instead of having to make affirmative representations in their application to mitigate malicious conduct and intellectual property abuse, those applicants can provide the minimal requirements knowing that they have the deep pockets to win any auction.
In the current Applicant Guidebook there is a disincentive for applicants to undertake affirmative obligations above the minimal requirements, because these obligations do not provide any benefit in an auction and only represent additional operational costs that are likely to limit the number of registrations. Unfortunately ICANN has created a paradigm where instead of applicants investing in affirmative obligations to enhance its role as a trustee of a public resource, they are incentivized to do the minimal and maximize financial resources to win an auction and write a big check to ICANN.
B. ICANN Staff
The straight auction is the simplest path forward for ICANN staff on a number of levels. First, it mitigates ICANN’s litigation risk profile, since a pure auction contention resolution mechanism is the easiest to defend in court, even if the auction winner is clearly the least suited/desirable trustee of a given TLD. Second, undertaking any type of comparative analysis will require additional work by staff in overseeing the new gTLD process. Third, ICANN staff will also have additional ongoing compliance obligations to monitor those affirmative obligations made by applicants to mitigate malicious conduct and protect intellectual property rights.
THE ICANN BOARD MUST EXERCISE ITS FIDUCIARY DUTY
The decision that the ICANN Board makes in connection with this issue will have far reaching consequences on ICANN’s legitimacy as a trustee of a global public resource. While there is a clear desire to establish procedures ( aka “black boxes”) that produce predicable results and which can scale, the ICANN Board cannot place form over substance. One of the reasons behind ICANN’s initial creation was to alleviate the burden on government in coordinating this global resource.14 However, if the result of this new gTLD initiative is ICANN staff turning a crank on a black box that produces an unlimited number of new gTLDs with an unlimited number of unforeseen consequences, then ICANN has failed. While this proposed contention set resolution process will place an additional burden on the ICANN staff and Board workload, it is a burden they have accepted as a trustee of a global public resource.
1 http://www.icann.org/en/tlds/app-index.htm
2 http://www.icann.org/en/tlds/report/report-iiib1a-09nov00.htm
3 http://cyber.law.harvard.edu/icann/la2000/archive/scribe-icann-111600.html
4 http://www.icann.org/en/tlds/org/
5 Need to preserve a stable, well-functioning .org registry; ability to comply with ICANN-developed policies; enhancement of competition for registration services; differentiation of the .org TLD; Inclusion of mechanisms for promoting the registry’s operation in a manner that is responsive to the needs, concerns, and views of the noncommercial Internet user community; level of support for the proposal from .org registrants; the type, quality, and cost of the registry services proposed; ability and commitment to support, function in, and adapt protocol changes in the shared registry system; transition considerations; ability to meet and commitment to comply with the qualification and use requirements of the VeriSign endowment and proposed use of the endowment; and the completeness of the proposals submitted and the extent to which they demonstrate realistic plans and sound analysis.
6 (1) Gartner, Inc. performed an evaluation of technical aspects of the bids, which were identified in the published criteria (based on Board comments at its Accra meeting) as of primary importance. (2) An international team of Chief Information Officers from major academic institutions performed an independent evaluation of those technical aspects using a different methodology. (3) The Noncommercial Domain Name Holders Constituency of ICANN’s DNSO evaluated “usage” aspects (as detailed below) of the bids. And (4) ICANN’s General Counsel evaluated certain procedural aspects. See http://www.icann.org/en/tlds/org/preliminary-evaluation-report-19aug02.htm
7 http://www.icann.org/en/tlds/org/preliminary-evaluation-report-19aug02.htm
8 http://www.icann.org/en/tlds/dotnet-reassignment/net-rfp-final-10dec04.pdf
9 ICANN Policy Compliance (absolute); ICANN Policy Compliance (relative); Equivalent access for registrars (absolute); Equivalent access for registrars (relative); Registry Operations (absolute); Registry Operations (relative); Revenue and Pricing Model; Pricing; Outline Technical Capabilities (relative); General description of proposed facilities and systems; Stability of resolution and performance capabilities; Operational Scalability; Registry-Registrar model/protocol and shared registration system Database Capabilities; Geographic network coverage; Zone file generation; Zone file distribution and publication; Billing and Collection systems; Backup; Escrow; Publicly Accessible WHOIS Service Peak capacities; System reliability; Support for IDNs, IPv6, DNSSEC; Outage Prevention and Recovery Technical and Other Support; Provision for Registry Failure; Provision for Registry Failure; Security; Additional Relative Criteria; and Transition or Migration Plan, see
http://www.icann.org/en/tlds/dotnet-reassignment/net-rfp-finalreport-issue4-27may05.pdf
10 Id.
11 Id.
12 http://www.icann.org/en/minutes/minutes-07jun05.htm
13 Section 4.2.3 of the Applicant Guidebook, see http://www.icann.org/en/topics/new-gtlds/draft-rfp-clean-12nov10-en.pdf.
14 ICANN made this following representation when seeking non-profit tax status, “ICANN’s tax-exempt purpose of lessening the burdens of government”, see http://www.icann.org/en/financials/tax/us/appendix-4.htm
Sponsored byCSC
Sponsored byIPv4.Global
Sponsored byVerisign
Sponsored byWhoisXML API
Sponsored byRadix
Sponsored byVerisign
Sponsored byDNIB.com
Good analysis Michael.
Could some of the same goals not be achieved by merely tweaking the CPE to lower the threshold score required to avoid an auction?
Thanks for the constructive comment.
Under the current guidebook if two applicants score 14 points (or some lower agreed upon threshold) the result is still an auction. By adding more evaluation criteria designed to measure trust and affirmative steps to mitigate malicious conduct you allow for a better/broader potential picture of the applicants, which hopefully would help in selecting the better trustee.
I am not proposing to totally eliminate the auction aspect. As I noted in the article an auction would have been the ideal resolution mechanism given the close bid in the .NET RFP. I am just concerned about the potential unintended consequences of relying solely on auctions.
Thoughtful analysis Michael
Isn’t the actual problem much deeper and with the allocation itself rather than the actual allocation mechanism?
.org .net and .biz allocations were all to entities NOT seeking to compete within the sting.
Doesn’t each of your examples .bet, .gay .gop .kids involve [some] applicants seeking to secure from ICANN an implicit DNS branding advantage to be allowed to compete from the top level against their competitors in the same industry who are forced to compete from the second level?
Michael,
Your comments are right on the money. We have long argued in each of the Applicant Guidebook (AGB) public comments fora as well as at the open mike at ICANN meetings that the evaluation scoring for community needs to be lowered to allow for a more reasonable threshold. Community-based application scoring is the most subjective aspect of the AGB evaluations, yet ICANN staff continue to rebuff our arguments along with those of the BC and IPC constituencies, which make up two of the major ICANN institutions amongst the many that feel as we do. A cynical mind would believe that staff would prefer to have all multiple applicant applications be forced into auction for all the good reasons you have noted… Unless the evaluation scoring is revised in the final book, your thesis may well come true. As one of those looking for a fair and predictable process, I hope not.
Again, a welcome analysis. Thanks for shining a reasoned light on auctions.
Nice post Michael,
I have posted a CircleID article regarding TLD auctions in 2009. Using auctions to determine winners might backfire on ICANN. It has taken years to create the gTLD guidebook but nothing was addressed to make ICANN accountable in regards to selection of winners. Auctions seem like the easiest and simplest path for ICANN to take to avoid taking any responsibility of selecting winners, thus reducing its potential legal liability. ICANN also wins by taking additional money from the pockets of the applicant with the deepest pocket. The question then becomes, where does this money go and why should ICANN be profiting from auctions? Is this consistent with their Affirmation of Commitments? Accountability and transparency has comparative analysis written all over it but auctions still loom on the horizon.
As an applicant, I am ready for any auction that may arise, but certainly not pleased that monies will be “donated” to a non-for profit California corporation without justification why a better winner-selection process was not adopted. Just as gaming, the TLD applicant has no protection from ICANN that a fair process will be followed.
If ICANN is all about transparency and accountability, it will be interesting to see whether community TLDs will score the 14 points required. Given the subjectivity of evaluation and the incentive for ICANN to profit from auctions, scoring will be a much contested event. What is the real difference between a 13 (fail) or a 14 (pass). Will some applicants demand a recount?
Applications for generic, non-community TLDs seem to be the most susceptible to auctions. Perhaps the reasoning there is fair by ICANN to conduct auctions in the occasions where applications are nearly exactly the same.
Comparative analysis is not a beauty contest. I disagree with Michael on that. Beauty is superficial. You can not judge the book by its cover. You need to read the book and compare with a methodology that discredits applicants with a “wishlist” or a hopeful “to-do” list. Anyone can write a perfect application and promise the world. True comparative analysis is about showcasing results, innovative technology and milestones to back it up.
Will a true comparative analysis model be adopted? I hope so. However, ICANN has too much too lose financially as well as does not want to take any responsibility in selection. All the power to the evaluators. Our business and future depends on them. We will all be watching though. Once again, ICANN’s trust and credibility will be on the line.
“Instead of having to make affirmative representations in their application to mitigate malicious conduct and intellectual property abuse, those applicants can provide the minimal requirements knowing that they have the deep pockets to win any auction.”
This is what the RIAA/music coalition is worried about as indicated in their most recent letter. I have voiced the same concerns and that is why I am firm with my commitment that .music must be a restrictive community based TLD and not generic or open. We have been developing strict guidelines and policies to address the issues of malicious conduct and artist trademark and name disputes. The last thing that content creators want is their related-TLD being open for the public and anyone to register. The risks are too high for this to happen and abuse will certainly happen.
There are some cases where auctions are warranted. However, in TLDs where copyright and IPs are critical, comparative analysis that is both realistic and accountable is required.
Constantine Roussos
.music