Berkman - open ccTLDs
Note: a previous version of this article
identified the report as a "Berkman" report.
This was an error; the report is Ben Edelman's
alone and not an official product of the Berkman Center.
The Berkman Center’s Ben Edelman has released another of his mini-research
reports, “Registrations in Open ccTLDs.” This one examines registration patterns
in the open ccTLDs as policy guidance for what might happen in any additional
new unrestricted domains.
The Edelman reports, which purport to be academic, always invoke mixed feelings
from this academic observer. One the one hand, they are welcome and constructive
contributions because they tend to bring some empirical analysis and real data
into issues that are usually just debated on ideological grounds. On the other
hand, the methods are often much sloppier than the academic dress would suggest,
making it possible for numerous interpretations to be drawn from the data. More
seriously, the studies tend to occupy an ambiguous middle ground between empirical
reports and interpretive advocacy. They purport to be purely “documentary” yet
the reports often contain interpretations that seem biased in a particular policy
Below, I take a closer look at the conclusions of “Registrations in Open TLDs”.
While I take issue with the interpretation of the data, I encourage Berkman
and Ben to continue working in this vein, hopefully tightening up their methodology
and logical rigor in the future, and perhaps making it clearer to themselves,
as well as the rest of us, whether they intend for these reports to be objective
data dumps or policy analyses supporting some options over others. Either option
is legitimate, but you can't straddle the fence.
The main point of the latest Edelman report is to make a case that there is
no need for new unrestricted top-level domains in the DNS. Toward that end,
the report says that the results “support the following claims:”
- That there are “substantial” defensive
registrations in the open ccTLDs.
- That there is “substantial” cybersquatting
in the open ccTLDs.
- That there
is “substantial” speculative
registration and warehousing in the open ccTLDs.
- That there is little need for
additional unrestricted gTLDs.
Let’s take up these claims one by one.
- Substantial defensive registrations.
The word “substantial” is not scientific. One could easily use the same data
to make the claim that “a surprisingly low level” of defensive registrations
is found in these TLDs. It’s all a matter of interpretation.
Consider: the study took a highly selective list of 1271 of the world’s most
famous brand names, and discovered that nearly 70 percent of them were not
defensively registered in the open ccTLDs. This was surprising to me; I would
have expected more defensive registrations. It is apparent by now that no serious
adverse consequences have come from not defensively registering these names.
The data show, for example that genuity.tv and starbucks.tv
have been registered by some guy in Dallas, but most of us didn’t know that
and have no apparent reason to care; neither company’s brand equity seems to
be impaired. The data also show that a significant portion of famous mark registrations
are waiting to be deleted. Thus, one could use the same data to argue that trademark
owners gain very little from defensive registration across all TLDs, and that
both they and the speculators are gradually learning this. Obviously the proportion
of speculation and defensive registration would be even lower with less well-known
The study mentions, but in a peripheral way, that .tv and .ws are subject to
ICANN’s UDRP. Which means that not only did an overwhelming majority of famous
mark holders choose not to register their names in those domains, but
that they haven’t bothered to challenge the obvious cybersquatters who did register
them, either – again, with no documented adverse consequences.
- Substantial cybersquatting. Of the registered .CC and
.TV famous name domains, many are registered to entities other than the respective
That word “substantial” again. What exactly does it mean? By any objective
measure, there are orders of magnitude more cybersquatted registrations in .COM,
.NET, and .ORG than there are in the open ccTLD domains. So by what standard
is cybersquatting a problem unique to them? And to what degree does this tell
us anything about what will happen in new domains?
The real issue here is not how many of the “famous” names are registered to
entities other than their rightful owners, but what percentage of the total
number of registrations in these TLDs consists of illicit and defensive
registrations. In other words, the important policy issue is: can a TLD operator
make a sustainable business solely out of defensive and potentially illegal
registrations? If the answer is “yes,” then there might be some reason to limit
entry. If the answer is “no,” then there is no reason, given the remedies of
the UDRP and the law, to block additional TLDs if entrepreneurs are willing
to add them and customers are willing to patronize them.
The Berkman study does not address this more precise formulation of the problem.
However, nothing in the evidence suggests that the answer is yes. If only 30%
of the 1,271 most famous names in the world are registered defensively that’s
a grand total of 370 registrations. Even if all of the famous names were registered
either defensively or illegitimately, and we multiply the number of names by
five to err on the safe side it’s hard to see how you get more than 6500 registrations
by selling infringing and defensive registrations. And you can’t sustain a commercial
registry on 6500 registrations.
- Substantial speculative registration and warehousing.
The report complains, “Top registrants of common nouns in .CC, .TV, and .WS
have registered as many as 50 common nouns in a single open ccTLD.” Well, so
what? This is a bogus point, for three reasons. First, what’s wrong with registering
lots of common nouns? Some people collect baseball cards, is this any worse?
We really need to resist the idea that it is somehow ICANN’s responsibility
to tell people how to use domain names and how many registrations are appropriate.
ICANN’s job is to coordinate how the root name servers are administered. Second,
much more extensive patterns of speculation and warehousing are found in .COM,
.NET and .ORG. Speculators have registered hundreds and even thousands of domain
names in the gTLDs; the speculative numbers for the open ccTLDs look pretty
puny in comparison.
Finally, and most important, the Berkman analysts seem to be utterly innocent
of basic principles of economics. The existence of speculation contradicts the
report’s conclusion that there is no need for additional open domains. Speculation
and warehousing are driven by perceived scarcity. People spend $100 or more
to sit on a name and not use it because they are hoping that someone else will
want to buy it from them in the future. If that kind of speculation is considered
a “problem,” (and it’s not clear why it is a problem for anyone but the person
paying the rent) the obvious solution is to expand the supply of TLDs to reduce
the incentive for speculation and warehousing.
web content see little need for additional unrestricted gTLDs.
Here again, the Berkman analysis seems to be based on assumptions that are
inimical to or ignorant of market processes. The implicit conclusion is that
if web content providers “see little need” for additional TLDs, then ICANN shouldn’t
permit them. But that’s a non sequitur: there is no reason at all for ICANN
to decide what providers of web content need or don’t need. Web content providers
can make this decision for themselves by patronizing or failing to patronize
new TLD services. Maybe .CC and .WS are unpopular because they are meaningless
and ugly domains that no one associates with anything. Maybe people don’t trust
them operationally. Maybe they are less popular because, indeed, content providers
believe in sticking with .COM and nothing will change that. Perhaps new names
or alternative service concepts will succeed in the market, perhaps they will
not. To this policy analyst, the issue regarding new open TLDs has never been
about ICANN picking the “right” character strings or the “right” business model
(restricted or unrestricted) but about defining a method for new entry that
allows consumers and producers to arrive at suitable names and business models
for themselves, in the free marketplace. It is only within ICANN’s amateurish
regulatory apparatus that technologists with no economic expertise or regulatory
experience have managed to convince themselves that the domain name space needs
a central planner.
This discussion has been archived.
No new comments can be posted.
New TLDs and the Ben Edelman Report
Log in/Create an Account
| 33 comments
The Fine Print:
The following comments are owned by whoever posted them.
We are not responsible for them in any way.
If you're worried about methodology, start following some of Bob Connor's links to supposedly unique .info sites and see how many are just redirects or mirrors of other sites (typically but not only to .com). If you're worried about what someone will do with supposedly scientific information, here is Afilias (the thoroughly discredited Roland LaPlante) spinning Bob Connor's research, and even he admits .info has not lived up to expectations. And if you're worried about accuracy, 900,000 was not 5 times 500,000 when I went to school. -g
[ Reply to This | Parent
"No one else out there is really making an attempt to do the same kind of research that Ben does."
ICANN has claimed not enough direct participation by stakeholders from the private sector and has certainly singled out private industry that - per ICANN - directly benefit from its existence and hence should be "contributing".
This is the same entity that has slammed the door on private industry of any opportunity to innovate the addressing space without going through regulatory-type hoops for no apparant or logical means for doing so (i.e. no identifiable outcome). Generally, private industry does not participate where there is not reasonable and clear opportunity, especially in these economic times where every expense projection is under internal scrutiny. Qualified technology and other market research funded by private industry? What possible reason has ICANN supplied to the private sector for such a "contribution"? It is noticeably lacking for a reason.
[ Reply to This | Parent
Milton doesn't seem to like the UDRP. Milton studies the UDRP and finds it less than perfect. Does that mean his research should be considered invalid? No. The statistics stand on their own, and anyone is free to attempt to draw different conclusions. If you can show flaws in the methodology or errors in the data, that's one thing, but if the science is sound, then one (including its author) is left to interpret it as one wishes. Ditto for Ben. I don't see anyone coming up with credible arguments regarding the methodology or data for either of them. I respect them both and value their research. If you don't like their ideologies, do your own research and publish it and draw whatever conclusions you wish. -g
[ Reply to This | Parent
|3 replies beneath your current threshold.|