The DNA of the new gTLDs – why .xyz looks odd
HosterStats has compared the domains of all new gTLDs against the equivalent in COM/NET/ORG/BIZ/INFO/MOBI/ASIA to see what keyword patterns are present. These TLD comparisons show the DNA of a TLD and DNA of .xyz new gTLD is rather odd when compared with the TLD DNA of other new gTLDs such as .club and .guru.
The domain registration activity in .guru and .club shows that there is significant interest in these gTLDs because phrases that are registered across com/net/org/biz/info/mobi/asia dominate the registrations. These are often the valuable generic keyword domains and brand protection domains. What makes .xyz new gTLD so different is that the .xyz/.com pairing accounts for 33.678% of domain name registrations in that TLD. In .club, the .club/.com pairing only accounts for 11.278%. In .guru, the .guru/.com pairing only accounts for 7.432%.
The reason for this is because .xyz apparently did a deal with Network Solutions to promote the new gTLD by offering a free registration to some Network Solutions customers. The story has been covered in detail on TheDomains.com , DomainIncite.com and DomainNameWire.com.
In most new TLD launches, the first domains to be registered are brand protection domains (generally registered in other TLDs) and generic keywords. The latter are considered valuable and some of the new gTLDs have reserved a lot of the single word generic domains for themselves. This pattern of registrations appears in most of the active new gTLDs where there is significant interest in the gTLD. However the .xyz registry deal with Network Solutions ensured that the TLD DNA of .xyz would be quite different to the others. The .xyz registry and gTLD have taken some flak over this deal and the opt out nature of many of the Network Solutions registrations have led to claims of the .xyz registration figures being at best unreliable and at worst inflated. It is the kind of marketing that the new gTLDs could do without as the end user has to have some trust in the TLD in which they register domains and develop websites and services. The big test for .xyz gTLD lies ahead – will these registrations be developed or will they effectively turn the gTLD into a Dead Zone where few websites are found? Ordinarily, it could take slightly over a year to find out when the renewal fees for these special offer domains fall due. But with this kind of publicity, people might think twice about developing sites in .xyz gTLD and the next few weeks of registration activity will be interesting.
Nice analysis!
I like the “DNA” concept. Could you take this further to determine the “DNA of the new gtld registrant?
Could you look at the matching com/net/org/etc domain and check if it is likely owned by a domainer, by checking for a parked page or a page containing “for sale” versus an end-user with an active website?
*
Thanks, John, for your carefully reasoned article with stats to back it up!
*
I run a 350K domain usage and development survey each month. I could probably use the same software and databases to run a comparison on the new gTLDs compared to the main TLDs. The easiest to run would be a development survey that measures development in the new gTLDs.
The problem with a lot of registries is that they are relying on iffy analysis (COInternet) and extrapolation (Eurid’s dodgy usage survey that uses students to “classify” websites) when it comes to measuring usage.
It could be done but would take some time (probably more than 2 weeks given the mix of languages) to do it properly.
My understanding is that the XYZ registry sold thousands of domains to NetSol at wholesale prices. Although Specification 9 of the ICANN registry agreement governing the registry code of conduct is vague, I do not see anything in Specification 9 prohibiting such a practice as long as other registrars are afforded the same opportunity.
Specification 9 reads:
“directly or indirectly show any preference or provide any special consideration to any registrar with respect to operational access to registry systems and related registry services, unless comparable opportunities to qualify for such preferences or considerations are made available to all registrars on substantially similar terms and subject to substantially similar conditions.”
It could also be a kind of “Brand Champion” idea (getting a major registrar onboard to promote the TLD) that didn’t work out as expected. But after one registrar has been selected as a kind of Brand Champion would the other registrars be so eager to promote it? It might be interesting to see if other registrars were offered a similar deal and rejected it.
The bottom line: 25k domains were not registered by a human registrant. Let’s cut through the BS and call it what it is: an attempt to game the numbers. Whoever had this brilliant idea, must be having second thoughts right now.
It has backfired but there may be some renewals among the bunch. It might have been a good idea if nobody was watching the zonefiles. The CZDS (the centralised zonefile service) is currently stuck on yesterday’s zonefiles so we don’t yet know how it did yesterday. If it maintains enough new registrations over the next few days, then it might be ok for it. (These events are happening within the bubble of the Domain Business rather than the wider internet.) If it does not and new registrations collapse, then it might be one of the worst moves yet in the new gTLD business and it has seen some really bad ones.
John, I fully agree with your thought process and appreciate your expert data analysis. Indeed, you raise a great point with the fact that otherwise “premium” keywords were left untouched, while utter junk that existed in .com was matched with .xyz pseudo-registrations. Some of these .com’s were even in expiry!
It is a perfect storm for junk registrations. The offer seems to have been an opt-out offer to customers who had built up enough reward points. That would favour those with high numbers of registrations rather than brand owners and those with developed websites. It could have been NetSol’s implementation of the offer in that rather than targeting bluechip customers with older domains (the ideal brand protection registration customer), it targeted customers with high domain churn with loads of one year wonders. It is just a theory though.