New US trademark rules likely to exclude many dot-brand gTLDs

The US Patent and Trademark Office plans to allow domain name registries to get trademarks on their gTLDs.
Changes proposed this week seem to be limited to dot-brand gTLDs and would not appear to allow registries for generic strings — not even “closed” generics — to obtain trademarks.
But the rules are crafted in such a way that single-registrant dot-brands might be excluded.
Under existing USPTO policy, applications for trademarks that consist solely of a gTLD cannot be approved, because they don’t identify the source of goods and services.
If “.com” were a trademark, one might have to assume that the source of Amazon.com’s services was Verisign, which is plainly not the case.
But the new gTLD program has invited in hundreds of gTLDs that exactly match existing trademarks. The USPTO said:

Some of the new gTLDs under consideration may have significance as source identifiers… Accordingly, the USPTO is amending its gTLD policy to allow, in some circumstances, for the registration of a mark consisting of a gTLD for domain-name registration or registry services

In order to have a gTLD trademark approved, the applicant would have to pass several tests, substantially reducing the number of marks that would get the USPTO’s blessing.
First, only companies that have signed a Registry Agreement with ICANN would be able to get a gTLD trademark. That should continue to prohibit “front-running”, in which a gTLD applicant tries to secure an advantage during the application process by getting a trademark first.
Second, the registry would have to own a prior trademark for the gTLD string in question. It would have to exactly match the gTLD, though the dot would not be considered.
It would have to be a word mark, without attached disclaimers, for the same types of goods and services that web sites within the gTLD are supposed to provide.
What this seems to mean is that registries would not be able to get trademarks on closed generics.
You can’t get a US trademark on the word “cheese” if you sell cheese, for example, but you can if you sell a brand of T-shirts called Cheese.
So you could only get a trademark on “.cheese” as a gTLD if the class was something along the lines of “domain name registration services for web sites devoted to selling T-shirts”.
Third, registries would have to present a bunch of other evidence demonstrating that their brand is already so well-known that consumers will automatically assume they also own the gTLD:

Because consumers are so highly conditioned and may be predisposed to view gTLDs as non-source indicating, the applicant must show that consumers already will be so familiar with the wording as a mark, that they will transfer the source recognition even to the domain name registration or registry services.

Fourth, and here’s the kicker, the registry would have to show it provides a “legitimate service for the benefit of others”. The USPTO explained:

To be considered a service within the parameters of the Trademark Act, an activity must, inter alia, be primarily for the benefit of someone other than the applicant.
…
While operating a gTLD registry that is only available for the applicant’s employees or for the applicant’s marketing initiatives alone generally would not qualify as a service, registration for use by the applicant’s affiliated distributors typically would.

In other words, a .ford as a single-registrant gTLD would not qualify for a trademark, but a .ford that allowed its dealerships around the world to register domains would.
That appears to exclude many dot-brand applicants. In the current batch, most dot-brands expect to be the sole registrant as well as the registry, at least at first.
Some applications talk in vague terms about also opening up their namespace to affiliates, but in most applications I’ve read that’s a wait-and-see proposition.
The new USPTO rules, which are open for comment to people who have registered with its web site, would appear to apply to a very small number of applicants at this stage.

Mystery gTLD applicant to take Google fight to lawmakers

An as-yet unidentified new gTLD applicant plans to lobby Washington DC and Brussels hard to get dozens of Google’s new gTLD bids thrown out of ICANN on competition grounds.
Phil Corwin of the law firm Virtuallaw, who is representing this applicant, told DI yesterday that his client believes Google plans to use new gTLDs to choke off competition in the web search market.
“They’re trying to use the TLD program to enhance their own dominance and exclude potential competitors,” Corwin said. “We think this should be looked at now because once these TLDs are delegated the delegations are basically forever.”
He’s planning to take these concerns to “policy makers and regulators” in the US and Europe, in a concerted campaign likely to kick off towards the end of the month (his client’s identity will be revealed at that time, he assured us).
Corwin’s client — which is in at least one contention set with Google, though in none with Amazon — reckons ICANN’s new gTLD program is ill-suited to pick the best candidate to run a gTLD.
If objections to new gTLD applications fail, the last-resort method for deciding the winner of a contention set is auction. Google obviously has the resources to win any auction it finds itself in.
“On any TLD Google has applied for, nobody can beat them,” said Corwin. “They have $50bn cash, plus the value of their stock. If they want any of the TLDs they’ve applied for, they get them.”
“A string contention process that relies solely on an auction clearly favors the deep-pocketed,” he said.
Google applied for 101 new gTLDs, 98 of which remain in play today. A small handful of the strings are dot-brands (such as .youtube and .google), with the majority comprising dictionary words and abbreviations.
Some of its generic bids propose open business models, while others would have “closed” or single-registrant business models. As we reported on Friday, this has kicked off a firestorm in the ICANN community.
Corwin said that Google appears to be planning to close off not only individual TLDs, but entire categories of TLDs.
For example, Google has applied for .youtube as a brand, but it’s also applied for .film, .movie, .mov, .live, .show and .tube with a variety of proposed business models.
“You can pretty well bet that they’ll exclude those that will pose a competitive threat to YouTube,” Corwin said.
Search will become much more important after the launch of hundreds of new gTLDs, Corwin reckons, as consumers are “not going to know that most of them exist”.
“Generic words are the perfect platform for constructing vertical search engines that can compete against Google’s general search engine,” Corwin said.
“Google is trying to buy up not just one but multiple terms that cover the same goods and services in key areas of internet commerce, and in effect control them so competition cannot arise and challenge Google’s dominance as a search engine,” he said.
Google has not yet revealed in any meaningful way how its search engine will handle new gTLDs.
The US Federal Trade Commission, at the conclusion of an antitrust investigation, recently gave Google a pass for its practice of prominently displaying results from its own services on results pages.
With that in mind, if Google were to win its contention set for .movie, but not for .film, is it possible that .movie would get a competitive advantage from preferential treatment in search?
Corwin reckons that Google anti-competitive intentions are already suggested by its strategy in ICANN’s new gTLD prioritization draw, which took place in December.
Of the roughly 150 applications for which Draw tickets were not purchased, Google is behind 24 of them — including .movie, .music, .tube and .search — 22 of which are in contention sets.
As a result, these contention sets have all been shunted to the back of ICANN’s application processing queue, adding many months to time-to-market and costing rival, less-well-funded applicants a lot of money in ongoing overheads.
“We see Google playing a rather different game here to most other applicants in terms of their motivation, which is not to enter the market but to protect their market dominance,” Corwin said.
Corwin said the game plan is to taken all these concerns to policy makers and regulators in the US and Europe in order to get governments on-side, both inside and outside of the ICANN process.
Corwin is also counsel and front-man for domainer group the Internet Commerce Association, but he said that the new anti-Google drive is unrelated to his work for ICA.
So why is his client only bringing up the issue now? After all, we’ve all known about the contents of every new gTLD application since last June.
My hunch is that Google is playing hard-ball behind the scenes in settlement talks with contention set rivals.
Contention sets can be resolved only when all but one of the applicants drops out, either following an ICANN auction or private buy-outs. Most applicants favor private resolution because it offers them the chance to recoup some, all, or more than the money they splashed out on applying.
That game plan probably does not apply to Google, of course, which is not wanting of funds. The company may even have good reason to prefer ICANN auctions, in order to to discourage those who would apply for new gTLDs in future just in order to put their hands in Google’s pockets.
The topic of closed generics and competition is likely to be a hot-button topic at ICANN’s next public meeting, coming up in Beijing this April.
Members of ICANN’s Governmental Advisory Committee have already expressed some concerns about many “closed gTLD” applications made by Google, Amazon and others.
ICANN’s board of directors is currently mulling over what to do about the issue, and has thrown it open to public comment for your feedback.

Closed gTLD debate threatens Google and Amazon

Howls of criticism about Google, Amazon and others’ plans to grab huge swathes of new gTLD real estate and keep it to themselves seem to have spurred ICANN into action.
A public comment period opened this week seeks community feedback (indirectly) on applications such as Amazon’s .music, L’Oreal’s .beauty and Google’s .blog, among many others.
These gTLDs have all been proposed with “single-registrant” business models, in which the registry controls all second-level domains and regular registrars cannot sell them to anyone else.
It’s the “dot-brand” model, but applied to generic dictionary words for which the applicants have no trademark rights.
Scores of such applications have been made, notably by Google and Amazon, but they have drawn criticism from many in the ICANN community, such as a small group of registrars and others led by Blacknight Solutions.
Members of the Governmental Advisory Committee, most vocally Australia, have also expressed serious concerns about the model, saying it could be anti-competitive.
ICANN’s board of directors is currently mulling over these complaints, and has thrown the issue open to public comment to aid in its deliberations.
What it wants from you is:

proposed objective criteria for:

• classifying certain applications as “closed generic” TLDs, i.e., how to determine whether a string is generic, and
• determining the circumstances under which a particular TLD operator should be permitted to adopt “open” or “closed” registration policies.

The way the public comment request is phrased should be quite worrying to applicants for closed generic gTLDs.
It seem to assume that ICANN should be classifying gTLDs, something it has steadfastly refused to do for all of the years these kinds of debates have been raging.
What is a “closed generic” anyway?
The DI PRO New gTLD Application Tracker classifies gTLD applications into three buckets: Open, Restricted and Single-Registrant.
We made no attempt to segregate dot-brands from other Single Registrant bids, precisely because there’s currently no such thing as a dot-brand under ICANN’s rules.
There doesn’t seem to be much community concern about the apps we have classified as “Restricted” — applications for .lawyer that propose to vet registrants for their lawyerly credentials, for example.
The concern is all directed at Single Registrant bids. We have 912 of these in our database.
Many of these are dot-brands, where the applied-for string is an exact match with a famous trademark, but many are for dictionary words for which the applicant has no preexisting rights.
In order to sanely operate a dot-brand, applicants must request an exemption to the ICANN rules that oblige them to offer their gTLDs via accredited registrars on a non-discriminatory basis.
This Code of Conduct is a part of the base Registry Agreement for new gTLDs, but it contains a carve-out for single-registrant applicants:

Registry Operator may request an exemption to this Code of Conduct, and such exemption may be granted by ICANN in ICANN’s reasonable discretion, if Registry Operator demonstrates to ICANN’s reasonable satisfaction that (i) all domain name registrations in the TLD are registered to, and maintained by, Registry Operator for its own exclusive use, (ii) Registry Operator does not sell, distribute or transfer control or use of any registrations in the TLD to any third party that is not an Affiliate of Registry Operator, and (iii) application of this Code of Conduct to the TLD is not necessary to protect the public interest.

This provision was added specifically in order to enable “dot-brands” to exist.
It would be pretty weird if, for example, L’Oreal was forced to make .loreal domains available via hundreds of registrars. By requesting an exception, L’Oreal has the chance to keep .loreal in-house.
However, because ICANN deliberately has made no distinction between commonly used words and brands (.amazon could be both), L’Oreal was also able to apply for .beauty as a single-registrant gTLD.
It’s not really a loophole — the possibility of companies applying for closed generics was envisaged by ICANN and the policy-making community long before the application window even opened.
Make no mistake, this is well-trodden ground. ICANN had plenty of opportunities to address the issue before the new gTLD application window opened a year ago and it quite consciously decided not to.
The feeling over the last couple of years has been that objection mechanisms such as the Community Objection, as well as GAC Advice, would be sufficient to close down these problematic gTLDs bids.
During the year-long community discussion about registry-registrar vertical integration, the possibility of closed generics was acknowledged and heavily debated.
The GNSO’s Vertical Integration Working Group failed to reach consensus on almost everything, but most of the recommendations emerging from it included some Code of Conduct exemptions for dot-brands.
Some in the WG suggested that the exemptions should only apply to true dot-brands (ie, those back up by a trademark) but ICANN decided against referring to trademarks when it wrote the Code of Conduct due to the very real possibility that it would encourage gaming by speculators.
That problem has not disappeared. While there’s no such gaming in the current batch of applications, there will be second and third and fourth application rounds that the rules being hastily debated at the last minute right now will also (presumably) apply to.
What do closed generic applicants want?
Some ICANN community members assumed that it would be the big domainer-backed companies (later emerging as Donuts, Uniregistry et al) that would attempt these kinds of land-grabs.
But that (so far) hasn’t turned out to be the case. The domainers have generally proposed registration policies that are super, super liberal in comparison to Google, Amazon and other closed-generic applicants.
I believe it’s partly because it’s these massively powerful e-commerce companies that are the ones making the land-grabs, and the scale of the grabs, that the issue of closed generics has reemerged now.
There are two broad use cases of concern here.
First, the .beauty scenario: L’Oreal keeps all the second-level .beauty domains to itself, essentially converting the word “beauty” into a brand name as far as the DNS is concerned.
Second, the .blog scenario: Google implements a policy that all .blog domains must use its Blogger service, potentially to the detriment of competitors such as WordPress or Tumblr.
In both scenarios, the bids could be rejected in their entirety as a result of formal objections, ICANN board action or Governmental Advisory Committee advice.
If the applications were approved, ICANN could also subjectively apply the ill-defined “public interest” test outlined above to force compliance with the Code of Conduct.
But that would merely lead to the bizarre scenario where 1,700 accredited registrars all qualify to sell .music domain names, but the only potential customer is Amazon.com’s intellectual property management department (which wants to run .music as a single-registrant gTLD).
As ICANN points out in its public comments request, the Code of Conduct regulates who can sell domain names in new gTLDs, not who they can sell them to.
The .blog scenario is a little different.
This is what Google, which has applied via its Charleston Road Registry subsidiary, has proposed (with my emphasis):

Should ICANN grant Charleston Road Registry’s exemption to the Code of Conduct, and the proposed gTLD operate with Google as the sole registrar and registrant, members of the public will not be able to directly register domain names in this new gTLD. Users will, however, be given the opportunity to make use of a vanity second-level domain as a memorable identifier linked to content in Blogger.

In other words, Google will “own” all the second-level .blog domains, but will allow Blogger customers to “use” them.
It looks like what it is: a transparently bogus attempted workaround of the Code of Conduct, designed to let Google exclude rival blogging services and independent, self-managed bloggers from .blog.
(Disclosure: DI is an independent, self-managed blog.)
However, I can’t see how what Google has proposed could possibly qualify for an exemption, which is only supposed to be granted provided the registry does not “transfer control or use of any registrations in the TLD to any third party”.
If sanity prevails, Google probably won’t qualify for an exemption.
But that won’t stop it tying .blog to Blogger.
The Code of Conduct, remember, is only concerned with equal, non-discriminatory access for accredited registrars. It does not speak to registry services or registry policies.
Google could possibly still have a registry policy stating that all .blog domains must point to Blogger.
In addition, Google could make the registration fee $0, making it unattractive for most registrars to carry (though I guess registrars could use it as a loss-leader, they wouldn’t be able to up-sell hosting and other services if all .blog domains have to use Blogger).
In conclusion
Applicants for closed generics paid millions of dollars to apply, using the rules set down in the Applicant Guidebook at the time, and I can’t see them being too happy about this eleventh hour surprise.
However, there can be little doubt that ICANN, if its role is to protect the public interest and consumer trust, has to seriously tackle the issue of closed generics.
But it has to address it in 2011.

Consumer Watchdog slams “outrageous” Google and Amazon keyword gTLD bids

Consumer Watchdog, a California-based consumer rights advocacy group, has attacked Google and Amazon’s new gTLD applications in a letter to an influential senator.
The organization has asked Sen. Jay Rockefeller, chair of the Senate Commerce, Science and Transportation Committee, to “thwart” the “outrageous” plans for single-registrant dictionary-word gTLDs.
Google and Amazon have separately applied for dozens of gTLDs — such as .music, .blog and .book — that they would exclusively use to market their own products and services.
Consumer Watchdog said in its letter (pdf):

If these applications are granted, large parts of the Internet would be privatized. It is one thing to own a domain associated with your brand, but it is a huge problem to take control of generic strings. Both Google and Amazon are already dominant players on the Internet. Allowing them further control by buying generic domain strings would threaten the free and open Internet that consumers rely upon. Consumer Watchdog urges you to do all that you can to thwart these outrageous efforts and ensure that the Internet continues its vibrant growth while serving the interests of all of its users.

As we reported yesterday, a number of domain name industry participants are planning to complain to ICANN about these applications on pretty much the same grounds.

Microsoft, Yahoo and others involved in new dot-brand gTLD group

HSBC, Microsoft, Yahoo and jewelry maker Richemont have told ICANN they plan to form a new GNSO stakeholder group just for single-registrant gTLD registries.
The group would comprise dot-brand registries and — potentially — other types of single-user gTLD manager.
A letter (pdf) to ICANN chair Steve Crocker, signed by executives from the four companies, reads in part:

As a completely new type of contracted party, we do not have a home to represent our unique community. In addition, the existence of conflicts with other contracted parties makes it challenging for us to reside within their stakeholder group.

Combined, the companies have applied for about 30 single-registrant gTLDs, mostly corresponding to brands.
Richemont, which is applying for dot-brands including .cartier, is also applying for the keywords .jewelry and .watches as single-user spaces.
The group plans to discuss formalizing itself at the next ICANN meeting, in Toronto this October.
During the just-concluded Prague meeting, the GNSO’s existing registries stakeholder group accepted several new gTLD applicants — I believe mainly conventional registries — into the fold as observers.
How the influx of new gTLD registries will affect the GNSO’s structure was a hot topic for the Governmental Advisory Committee during the meeting too. I guess now it has some of the answers it was looking for.

Most new gTLDs could be closed shops

ICANN’s new generic top-level domain program could create almost 900 closed, single-user namespaces, according to DI PRO’s preliminary analysis.
Surveying all 1,930 new gTLD applications, we’ve found that 912 – about 47% – can be classified as “single registrant” bids, in which the registry would tightly control the second level.
Single-registrant gTLDs are exempt from the Registry Code of Conduct, which obliges registries to offer their strings equally to the full ICANN-accredited registrar channel.
The applications include those for dot-brand strings that match famous trademarks, as well as attempts by applicants such as Amazon and Google to secure generic terms for their own use.
Our definition of “single registrant” includes cases where the applicant has indicated a willingness to lightly share second-level domains with its close affiliates and partners.
It also includes applications such as those for .gov-style zones in non-US jurisdictions, where domains would be available to multiple agencies under the same government umbrella.
But it does not include gTLD applications that would merely require registrants to provide credentials, be a member, or agree to certain restrictions in order to register a domain.
Since there’s been a lot of discussion this last week about whether the single-registrant model adds value to the internet, I thought I’d try to measure the likely scale of the “problem” when it comes to eventual delegation into the DNS root zone.
How many closed registries could we see?
According to the DI PRO database, of the 912 single-registrant applications, 132 are in contention sets. There are 101 contention sets with at least one such applicant.
Some are up against regular multiple-registrant applications (both open and restricted gTLDs), whilst others are only fighting it out with other single-registrant applicants.
Let’s look at a couple of hypothetical scenarios.
Scenario One – Single-Registrant Applicants Win Everything
First, let’s assume that each and every applicant passes their evaluations, does not drop out, and there are no successful objections.
Then let’s imagine that every contention set containing at least one single-registrant bidder is won by one of those single-registrant bidders.
According to my calculations, that would eliminate 31 single-registrant applications and 226 multiple-registrant applications from the pool.
Another 264 multiple-registrant gTLD applications would be eliminated in normal contention.
That would leave us with 881 single-registrant gTLDs and 528 regular gTLDs in the root.
Scenario Two – Single-Registrant Applicants Lose Everything
Again, let’s assume that everybody passes their evaluations and there are no objections or withdrawals.
But this time let’s imagine that every single-registrant applicant in a contention set with at least one multiple-registrant bidder loses. This is the opposite of our first scenario.
According to my calculations, that would eliminate 117 single-registrant applications and 140 multiple-registrant applications.
Again, normal contention would take care of another 264 multiple-registrant applications.
That would leave us with 795 single-registrant gTLDs in the root and 614 others.
In both of these scenarios, at either extreme of the possible contention outcomes, single-registrant gTLDs are in the comfortable majority of delegated gTLDs.
Of course, there’s no telling how many applications of all types will choose to withdraw, fail their evaluations, or be objected out of the game, so the numbers could change considerably.
As another disclaimer: this is all based on our preliminary analysis of the applications, subject to a margin of error and possible changes in future as we refine our categorization algorithms.