On July 25th ICANN announced the publication of the Draft Report of the Independent Review of the Trademark Clearinghouse (TMCH). This study was coordinated for ICANN by the Analysis Group, in conjunction with researchers from the Center for Internet and Society at Stanford as well the University of Pennsylvania’s Wharton School.

The Report’s Executive Summary provides a background on its origin and purpose:

The Trademark Clearinghouse (“TMCH”) was established in March 2013 and serves as central repository for information to be authenticated, stored, and disseminated, pertaining to the rights of trademark holders in ICANN’s New Generic Top-Level Domain (“new gTLD”) program. Analysis Group was commissioned by ICANN to undertake an independent review of TMCH services based on the Governmental Advisory Committee (“GAC”) recommendation in May 2011 that a comprehensive, post launch review be performed. The purpose of this review is not to make policy recommendations, but to assess the strengths and weaknesses of the TMCH services in conjunction with the specified areas for review proposed by the GAC. Specifically, our review is focused on the TMCH matching criteria, as well as the Claims Service and Sunrise Services. (Emphasis added)

That section makes clear that, while public comments on the draft report will be accepted through September 3rd, this Report was triggered by GAC concerns expressed before the Applicant Guidebook for the new gTLD program was even completed, and is not the work product of a GNSO-created working group and therefore will not directly result in the establishment of any new ICANN policy.

However, this Report will be of use to the Working Group (WG) established to review all Rights Protection Mechanisms (RPMs) in all generic Top Level Domains (gTLDs), of which I am one of three Co-Chairs. A sub-team of the WG is currently engaged in identifying available data to inform its policy discussions focusing on the TMCH and the related Claims Service and Sunrise Registration programs. The full WG will engage in an intensive review of all the available data relating to these subjects, including this Report, and then will reach its own consensus conclusions on policy recommendations.

As summarized by ICANN, the Report made three key findings (which are reproduced below along with relevant quotes from the Report’s Executive Summary):

• Expanding Matching Criteria to include non-exact matches may be of limited benefit: The dispute rate of completed registrations that are variations of trademark strings is very low.—“We also find that trademark holders infrequently dispute registrations that are variations of trademark strings. Given the low dispute rates, an expansion of the matching criteria may bring little benefit to trademark holders and only harm non trademark-holder domain registrants, who may be deterred from registering trademark string variations that would otherwise not be considered a trademark infringement by trademark holders or authorities who make such determinations.”
• Extending the Trademark Claims Service may have diminishing value: Registrations of names matching trademarks decline after the required 90-day Claims service period ends.—“[E]xtending the Claims Service period or expanding the matching criteria used for triggering Claims Service notifications may be of limited benefit to trademark holders and may be associated with costs incurred by other stakeholder groups, such as registries, registrars, and non-trademark-holder domain registrants. Although our data do not permit us to perform a cost-benefit analysis of extending the Claims Service or expanding the matching criteria, the tradeoffs felt by different stakeholder groups should be considered when weighing those policy decisions. The effectiveness of Claims Service notifications depends on how many registration attempts are being made. We find that registration activity declines after the 90-day Claims Service period ends, so any additional months added to the Claims Service period will likely have diminishing value.”
• Few trademark holders utilize the Sunrise period: Most users of the Trademark Clearinghouse submit proof of use to gain access to the Sunrise period. However, across eligible trademark holders, fewer than 20 percent have used the Sunrise period to date.—“[W]e find that although trademark holders expressed valuing the Sunrise period through questionnaire feedback and many trademark holders apply for Sunrise eligibility by submitting proof of use when recording their marks in the TMCH, many trademark holders do not utilize the period. This could be due to the expense of Sunrise registrations or because other protections of the TMCH services, such as the Claims Service, reduce the need for trademark holders to utilize Sunrise registrations.”

These are valuable findings and observations and will receive due consideration by the RPM Review WG.

In regard to any proposed expansion of the matching criteria, I lean against that option for two separate reasons. First, the TMCH is supposed to be a repository of registered trademarks meeting certain qualitative criteria, and permitting the registration of inexact matches would erode its value as a global trademark database. Second, after all the divisive debate within ICANN over adoption of Trademark-plus-Fifty, which permits trademark owners to register inexact matches of their marks in the TMCH corresponding to typosquat or domain-plus-product/services domains that were recovered in trademark litigation or a UDRP actions, the actual registration of Trademark-plus-Fifty terms in the TMCH has been quite low.

Finally, the Report was unable to quantify the extent to which the transmission of Claims Notices by registrars to individuals attempting to register domains that matched marks in the TMCH was either deterring infringing registrations, or causing non-infringing registrations to be abandoned out of fear of unwarranted resulting legal action. In this regard the Report states:

Although it is possible that the Claims Service and matching criteria may help deter rights-infringing registrations that are exact matches to trademark strings recorded in the TMCH, it is also possible that some good-faith registrations are being deterred by the current Claims Service system, which may be detrimental to the registration activity of non-trademark-holder domain registrants. Limitations of our data do not allow us to definitively conclude whether Claims Service notifications have a deterrent effect on either type of registration activity.

In relation to this inconclusive analysis, the report also documents that:

• As of April 1, 2016, there were 40,465 records in the TMCH, of which 32,528 were current and have been verified to have accurate and correct information meeting TMCH guidelines.
• There were 113.2 million unique download requests for TMCH records between October 2013 and February 2016 (note: registrars download a record each time there is an attempted registration of a new gTLD domain—but registrars may also download records when no registration attempt has been made).
• Roughly 26,405 unique, verified trademarks in the TMCH (81% of all verified trademarks in the TMCH) have been downloaded during the Claims Service period at least once.
• 93.7% of the 1.8 million registration attempts that received a Claims Service notification were abandoned, and only 6.3% went on to complete the domain registration. Of the nearly 114,000 registrations receiving a notification that were completed, only 0.3% were subject to subsequent domain disputes as of December 2015.

In regard to that very low dispute rate involving completed registrations of domains that triggered generations of a Claims Notice, the Report suggests possible reasons:

There are several possible reasons why the dispute rate on Claims Service notifications is so low. First, bad-faith registrations may be largely abandoned when a Claims Service notification is received, so very few domains are registered that trademark holders would wish to dispute. Second, there may be a lag between the time a domain is registered and discovered by a trademark holder and when a dispute is filed, causing us to see some registrations as non-disputed when they may become disputed in the future (i.e., we do not observe a dispute in the dispute data because it is limited to disputes that occurred before the end of 2015). Third, trademark holders may be generally less concerned by the domain registrations in the Claims Service data, either because the domain names are low-priority for disputes or because exact match registrations made in new gTLDs are less threatening to trademark holders than registrations in legacy TLDs like .com.

To evaluate whether the first explanation explains our results, we would need information on the domain names that were attempted in abandoned registrations. However, the Claims Service data only contain domain names for completed registrations. Therefore, we are unable to evaluate the characteristics of abandoned registration attempts.

I have been concerned since Spring 2015 that receipt of a highly legalistic Claims Notice may have scared off potential new gTLD registrants with no infringing intent but who are not sophisticated about trademark law, writing at that time:

No doubt there have been attempts by intentional cybersquatters to register trademarked names that have been effectively deterred when they received a Claims Notice and realized that the trademark owner would be notified of the domain registration immediately and might well take some form of legal response.

But there also may have been lots of potential registrants for non-infringing uses of short and meaningful generic dictionary words as domain labels who were spooked enough when they received the Claims Notice to abandon the registration. While the Claims Notice does provide a prospective registrant with information regarding the Jurisdiction where the trademark is registered and the class of Goods and Services that the trademark covers, most prospective registrants of non-infringing domains are not well versed in trademark law, don’t want to have to spend money to consult a lawyer to see if their registration will be infringing or not, and don’t want to risk being hit with a cease-and-desist letter, UDRP or URS filing, or a trademark infringement lawsuit. The same could be true even for potential registrants well versed in trademark law who simply don’t wish to expose themselves to a potential legal action, regardless of its merits—especially since continuing on to registration after receipt of the Notice might be alleged to constitute proof of bad faith registration…the TMCH has almost surely been quite effective in deterring infringing domain registrations at new gTLDs. But it appears to also have been a substantial damper on total new gTLD domain registrations. The unanswered question is how big of a headwind it has been.

The RPM Review WG may well try to find a more definitive answer to that question, and one promising area of inquiry is finding out what relevant data may lie outside of the TMCH in any records of attempted and abandoned registrations held by registrars. The WG may also consider making the language of the Claims Notice more understandable and less intimidating for the domain purchasing general public.

Overall, the Report provides much useful data and analysis for the Working Group’s further consideration, as it proceeds to comprehensively review the new gTLD PRMs and considers whether to recommend any modifications of them.