Two comments on the NSI proposal from the first consultancy

[Eric Brunner-Williams <ebw@xxxxxxxxxxxxxxxxxxxx>]

Two comments on the NSI proposal from the first consultancy.

These are made in a personal capacity, though like some others also 
making comments in a personal capacity, I draw upon experience from the 
late 1980s to the present, originally as a USG contractor, an IAHC 
contributor, and later as an ICANN "stakeholder", throughout most of the 
ICANN period, both as registry and as registrar.
1. p2, para 1, suggests that via cross-ownership in a registrar (or 
registrars) a registry could diversify its revenue base, strengthen its 
financial position and reduce risk to its registry operation.
When NeuStar/NeuLevel was preparing to go live with .biz (and I worked 
for incompetent buffons) I proposed that we (Nu*) have a "registrar in a 
can" for testing and registrar continuity purposes. I can't imagine the 
buffons (then CEO Ganick included) going long on spending NuStar and 
Melbourne's investment capital on selling Verisign's three brands, 
and/or the Afilias or GNR or RCOM brands, but not their own. I can't 
imagine even them using .biz money to sell anything but the .biz 
inventory (though I don't have to imagine them using .biz investment 
capital to complete the NANPA build-out).
I similarly can't imagine Elliot coming to a radically different choice, 
to flog com/net/org and biz but not his brand, to use his limited 
resources to promote competing brands, for some diversity of income that 
necessarily has to have a greater ROI than spending the same money on 
his brand.
So what about the other end of the initial new gTLD registry 
participants? I'm shaking my head there too. A cooperative registrar 
that sells everything but cooperative product, to make a bigger buck? A 
museum registrar that sells everything but museums? An aviation 
registrar that sells everything but what SITA could have sold but 
managed to fail on for internal reasons? I'm still shaking my head.
I just don't see how this rationale for cross-ownership, the registrar 
lifts registry that created or acquired it, works.
Would it have made sense when .org was given to ISOC? If so, would the 
correct threshold have been a million registrations or a hundred 
thousand registrations, and then only of the remaining VGRS inventory, 
and the then rather fictional inventories of NuStar and Afilias, but not 
the .org inventory?
Similarly, during the .net redelegation fraud?

For the subsequent new gTLD launches, up to the present?

If the "why is a dollar better spent on not marketing .foo by the foo 
registry" not enough of a headache, there's the fact that the registrar 
market is degenerating along two axis -- the larger registrars account 
for more of the total market than in past years, and ICANN's managed to 
allow a simply staggering number of shell registrars, owned by 
domainers, to come into existence via several well-known gaming routes. 
To be a good long-term bet, the foo registry's registrar property must 
either be a consolidation winner in the com market or a gaming 
operation, in the same market.
I'm involved with several current applications, all of which are 
concerned about the basic one time 185k fee and the annual 75k fee, and 
avoiding additional costs such as those for RSTEP and objection. The 
idea that these applicants are going to "diversify [their] revenue 
base[s], strengthen [their] financial position[s] and reduce risk to 
[their] registry operation[s] by starting up highly competitive 
traditional or domainer registrar operations, both for VGRS inventory, 
seems wicked unlikely to me.
2. p2, para 2 suggests that cross-ownership is the appropriate mechanism 
to relieve the restrictions which prevent a competitive registry from 
"reach[ing] a sustainable level of registrations in order to [become 
and] remain competitive in the market".
There are a number of candidate mechanisms to relieve the existing 
restrictions, the .museum operator sought one (5k direct), the .coop 
operator sought another (6 months direct), and we haven't even mentioned 
mandates for existing com/net/org/info registrations by museums or 
cooperatives to be transfered to those registries and their prior 
registrations marked "reserved"
to protect the transferred registrant, at any present price, or with any 
retroactive refund from the monopoly operator for the prior multi-year 
payments, at up to the 1995 level of $50/yr, or paid future-year 
payments, or the creation of a public interest registrar and a transfers 
mandate from some equally exploitive registrar price points. Yet this is 
the only mechanism for which an existence proof exists. The .org brand, 
existing independent from the prior incumbent monopoly operator.
We have to solve the registry capture problem. Restated, there is no 
ICANN gTLD brand, there is only the pre-ICANN C/N/O brands.
A decade ago there were two propositions on the table. The first one was 
to reallocate existing registry inventory into competing registries, 
without necessarily creating new namespaces. The second one was to leave 
existing registry inventory with the incumbent monopoly, and necessarily 
create new namespaces, and coincidently, create a distinct retail 
channel for all the namespaces, which we call "registrars". This is the 
genesis of "vertical separation" and "equal access".
The second proposition has not yet created an ICANN gTLD brand, and 
cooperatives and museums and aeronautical registrants remain 
overwhelmingly captured by the pre-ICANN, pre-competitive, C/N/O brands.
If our goal is to relieve the restrictions which prevent a registry from 
"reach[ing] a sustainable level of registrations in order to remain 
competitive in the market", ending the pre-ICANN capture of their 
registrants, and transferring their market share from the incumbent 
monopoly to the competitive registry goes to the root cause of existing 
inventory capture. The incumbent monopoly can elect to retain a 
registrant, but must pay the sponsored registry which ICANN considered, 
and authorized, as the competitive sponsored registry, the value of the 
registration, or it can cooperate with the transfer and prevent domainer 
seizure of the registrant's name and good will in the namespces of the 
incumbent monopoly.
Solutions that require the competitive registry to leave pre-competition 
captured inventory in place and "compete" with the incumbent monopoly by 
creation of new inventory, with, or without, any of a number of 
variations on the "vertical separation" and "equal access" rules for new 
inventory creation, are simply less effective than solutions that do not 
leave the existing inventory stranded at the incumbent monopoly and 
unable to transfer to the competitive registries.
Dressing the problem up with dancing registrars doesn't solve the 
problem, and as others have pointed out, the dancing registrars may have 
a greater interest in acquiring a greater share of the incumbent 
monopoly market share than the viability of new, and competitive registries.
I understand how a registrar would arrive at a registrar-centric 
solution to the registry start-up problem, but "competitive registrars" 
only exist in a market defined by the pre-competitive registries, 
everything else is simply too small to matter, with apologies for that 
observation to my friends Jeff and Ken.
Eric Brunner-Williams
again, in a personal capacity.