ICANN has put the DNS-centric community on notice for quite some time that the upper limit of its ability to conduct simultaneous evaluations of new gTLD applications would be around 500, and if a greater number of applications was received in the opening round it would need to resort to a batching system that used some means of prioritizing applications for processing. But it was always vague about how batching might work.
At last month’s Costa Rica meeting it gave a verbal explanation of the system its staff had come up with, and many attendees were baffled – while nearly all agreed that there had to be a better way, although they were having trouble coming up with an alternative. Pulling numbers out of a hat wouldn’t pass muster, as ICANN legal staff had determined that a pure chance approach would violate California law against private lotteries.
On March 30th ICANN issued an “Update on New gTLD Batching” which even promises a future video providing a real world demonstration of the chosen method. (http://www.icann.org/en/news/announcements/announcement-2-30mar12-en.htm). The methodology’s description in set forth in writing and ICANN has dubbed it “digital archery” as it involves an applicant coming closest to its own designated target. (http://newgtlds.icann.org/en/applicants/tas/batching-basics)
Here’s how ICANN describes this new mode of archery:
How does it work?
The batching process is activated if significantly more than 500 applications are received. Each applicant will be notified to register into an online system to set a future time target. The ability of applicants to "hit" that target at the selected time will be used to determine which applications are placed into the first, or subsequent batches.
Next, applicants will return to the online system on that day and time and try to hit submit as close as they can to their target time. It's kind of like a game of digital archery. First you set the target and then you try to hit it with as much accuracy as you can.
These two times (i.e., the target time and recorded time) are compared, and the absolute difference between the two is an applicant's score in the secondary timestamp. This is the number that will be used to determine placement within a batch.
In the example above, the target time is 12:00:00 and the actual time the target was hit is 12:00:01. That means the secondary timestamp is 00.00.01. Time variance numbers are absolute, which means that there are no negative numbers. An applicant who submitted the final entry 00:00:01 before the target time will have the same secondary timestamp as one who submitted 00:00:01 after the target time.
ICANN’s announcement also makes clear that it will guard against any one region dominating the archery tournament by grouping applicants into geographic regions. It also will permit any applicant that is in no hurry to get its new gTLD through the process to opt out and accept placement in a lower priority batch. Finally, the announcement provides no estimate of how long each batch will take to process – but if each batch took a full year that would still mean adding about ten new gTLDs to the authoritative root each week (or slightly less, presuming that some applications encounter objections or contention).
The announcement has brought some grumbling that this staff-authored, Board-blessed approach is something that should have gone through the standard policy development process but is instead being imposed from above. It has also engendered speculation that, since this is a game of skill rather than pure chance, some applicants will be renting facilities in Marina del Rey or using other means to improve their chances of coming closest to the target. That said, a random lottery is apparently forbidden and an auction process to buy a favored slot in the opening batch would no doubt engender all sorts of criticism. So, unless someone quickly comes up with a workable and popular alternative, this digital archery approach will be implemented very soon.
From our perspective the most important aspect of this approach is that it does not discriminate between gTLD types, and thereby assures the maximization of diverse types of new gTLDs and the greatest introduction of new price competition between registries.
Indeed, it helps assure that hotly contested generic new gTLDs will find their way to the first batch:
Multiple Applications for Similar Domain Name Extensions
To ensure stability and security of the domain name system, each domain name extension, or "string," must be unique. In the event that more than one organization applies for the same or similar top-level domain, all applications for those contending strings will be placed into the earliest batch designated. For example, if one contending applicant is selected in the first batch and the other contending application is selected in the second batch, both applications will be placed in the earlier batch. This will help ensure that one application is not unnecessarily held up by another.
What this means is that if multiple applicants apply for a general purpose new gTLDs like .web, or compelling verticals like .sport or .music, so long as one applicant makes it to the first batch all applicants for that string will all be evaluated in the first batch, regardless of their individual level of digital archery skills.
This is very much preferable to a prioritization system that has been floated by some trademark interests, who would have ICANN process IDNs first, .geos second, .brands third, and all .generics dead last. We have nothing against new gTLDs in non-ASCII scripts or city, state, and nation gTLDs. And we believe that some savvy brand owners are going to find that owning their brand at the top level of the DNS provides multiple benefits at very reasonable costs. But new .generics are beneficial as well, especially to foster domain price competition among registries.
As we noted a few days ago, ICANN has unveiled a proposed renewal agreement for VeriSign’s continued operation of .com and it permits domain price increases of 7% in four out of the contract’s six years. The other incumbent gTLDs are permitted to raise domain prices by 10% in each and every year -- or have no price limitations at all. (http://internetcommerce.org/DotCom_Deja_Vu) Effective price competition from new gTLDs can help temper future price increases by incumbents.
New gTLDs also will not be subject to pricing limitations, but we hope to see a variety of new monetization models that include low cost or even free domains. In any event, new gTLD choices in the general purpose and vertical generic spaces should provide meaningful competitive benefit, especially those that are operated by new entrants into the registry space. Right now ICANN has an open comment soliciting the best metrics for measuring Consumer Trust, Consumer Choice and Competition -- and the rapid introduction of new generic gTLDs certainly seems to punch the choice and competition buttons. (http://www.icann.org/en/news/public-comment/cctc-draft-advice-letter-23feb12-en.htm )
It appears all but inevitable that ICANN will need to utilize a batching system, as there will almost surely be more than 500 initial applicants. As of March 25th there were 839 registrants in its TAS new gTLD registration system, and by the time TAS registration closed on March 29th there were almost surely additional registrations. (http://www.icann.org/en/news/announcements/announcement-29mar12-en.htm ) Each TAS registrant can submit up to 50 new gTLD applications, which would yield a mind-blowing but most unlikely total number. The educated guesstimate in San Jose seemed to be 1400 total new gTLD applications, but given the large number of TAS registrants we wouldn’t be surprised to see the total go somewhat higher.
ICANN has also just announced that its target date for announcing the names of new gTLD applicants and their proposed strings is April 30th. (http://www.icann.org/en/news/announcements/announcement-02apr12-en.htm ) We look forward to the unveiling so we can begin to judge the potential benefits of the competitors. After that, let the archery tournament begin!