ICANN ICANN Email List Archives


<<< Chronological Index >>>    <<< Thread Index >>>

Two Possible Allocation Methods for Single-Letter 2nd-level domains

  • To: allocationmethods@xxxxxxxxx
  • Subject: Two Possible Allocation Methods for Single-Letter 2nd-level domains
  • From: George Kirikos <gkirikos@xxxxxxxxx>
  • Date: Tue, 16 Oct 2007 11:55:48 -0700 (PDT)


Here are two possible mechanisms to allocate these domains:

1) Randomized Allocation to Existing Registrants -- under this method,
all the value is passed on to existing domain name registrants, and not
to registries, ICANN, or unrelated 3rd-parties/middlemen. For example,
to allocate M.com, all registrants of domain names beginning with the
letter "M" in the .com space (e.g. ma.com, maa.com, mb.com, mother.com,
movies.com, and maybe 5 million others!) that have been in existence
for more than 6 days (to avoid the Add-Grace Period Issues) will
receive either 1 ticket equally (i.e. unweighted), *or*, preferably, a
number of tickets proportional to the length of time their domain name
has been registered (a domain that was created 5 years ago would
receive twice the weight as one created 2.5 years ago, and so on). The
1-letter domain name then would be randomly allocated to someone who
was an existing registrant, who could then shorten their domain name.
(NB: Since there is no money involved, this allocation method should
not be considered an illegal lottery, like .biz was found to be; here,
existing registrants are simply allocated an unexpected "bonus"

2) Sequential Auctions Open To Anyone for Places in a Queue to Pick
Names - Unlike an auction that others might suggest whereby people bid
on A.com, B.com, C.om, etc. separately (which I'm sure others will
suggest), I believe a set of auctions for the right to choose an
unallocated domain name would raise the most money.

For example, at the beginning of the series of auctions, there are 23
1-letter .coms (i.e. excluding the Q.com, Z.com and X.com that were
previously allocated and in use) available to be registered. *ALL*
parties would bid on the right to choose *any* of the remaining
available names. So, for example, Microsoft might be bidding in order
to secure M.com, but they would be competing with folks like Google who
might want G.com, against Yahoo who might want Y.com, and against a
multitude of others who want a 1-letter .com. The winner of the first
round gets to pick first.

So, suppose the winner is Microsoft, at $20 million, and they pick
M.com. That name is taken out of the list of available names, and a
month later, another auction is held. Suppose Google wins, at $15
million. They take G.com out of the set of available names. Now the
list is down to 21 domain names, and another auction is held a month
later, and so on (for simplicity, I've excluded the single-digits .com,
and other TLDs).

The beauty of this scheme is that it would promote fierce bidding at
the beginning. Parties who would not be bidding against each other for
M.com, for example, like Google or Yahoo, would still bid high, as they
would want to make sure that they have their choice of the remaining
available names (i.e. they would worry that someone else might take
Y.com or G.com, if they don't get to pick first). Furthermore, as the
number of domains taken rises (e.g. only 5 1-letter .coms left
available to bid upon), the bidding will also be fierce, as the
scarcity value has increased further, and anyone wanting a particular
1-letter .com out of the remaining 5 will want to ensure they get to
pick next,  lest someone else take it instead.

Compared to auctioning off the individual names (i.e. auction only for
A.com, auction only for B.com), this method also allows people who have
a choice between 2 or more 1-letter .com domains the opportunity to bid
for both. For example, if a company can do with either A.com *or*
B.com, but the auctions for both names were held at the same time, the
possibility would exist that they end up with *both* names by mistake,
and so they have to bid conservatively. If instead they were bidding in
the sequential auctions for the right to pick any available remaining
name, if they saw that A.com was taken in the 3rd auction, they know
that in the 4th auction and beyond that they still have a chance for
B.com, and can bid accordingly.

By holding these auctions a month at a time, it would be an orderly
release of the domains, too. e.g. Google might bid a premium to get to
pick G.com first, and launch it a month before other corporations
launch their competing sites (and up to 22 months before the last
single-letter .com is allocated). Also, holding these auctions
sequentially improves information available to all participants. They'd
see the results of the prior auctions, and gain a better sense of what
the names are worth.

In conclusion, I expect if domain name registrants were polled, they'd
go for the random lottery method, as that gives anyone a chance at a
domain based on their existing registrations. Probably the goal is to
raise the most cash, though, for ICANN, in which case I believe the 2nd
method, auctioning the ability to pick which domain you want out of the
remaining ones, and having sequential auctions for that right to pick
next, maximizes revenue.


George Kirikos

<<< Chronological Index >>>    <<< Thread Index >>>

Privacy Policy | Terms of Service | Cookies Policy