It’s been 2.5 years since the first of the new gTLDs were launched and there have been some significant changes to the global domain name market since then, which have been outlined in a new report from ICANN called Phase II Assessment of the Competitive Effects Associated with the New gTLD Program [pdf].
The study found the share of registrations held by the top four, top eight and top fifteen registries and registrars by domain name registrations has declined. There was also movement in the largest 15 registries and registrars as ranked by total domain registrations, with some registries or registrars who were not among the largest 15 in Phase I being ranked among the largest 15 in Phase II. And the number of registrars located in China has become more prevalent among registrars with the largest shares of new gTLD registrations. It also notes that domain registration shares of legacy gTLDs have declined as the percentage of European and Asian registry operators have grown.
Since the Phase I Assessment new gTLD registrations increased from 3,483,064 registrations as of November 2014 to 16,570,035 registrations as of March 2016. New gTLD registrations accounted for approximately two percent of all gTLD registrations as of November 2014 and now account for nine percent. Overall domain registration levels have also increased since Phase I, since registrations among the legacy TLDs in total continuing to increase driven largely by .com.
But the study strangely finds no aggregate (worldwide) effect of new gTLD entry or on legacy TLD registrations. This appears to be plainly wrong as several legacy gTLDs have seen a significant drop in registration numbers, several of which commenced their decline at the same time as new gTLDs were introduced when looking at RegistrarStats.com.
Some of these include .biz who had around 3.3 million registrations in early 2014 and around 2.7 million now. The .info TLD saw its decline start in late 2011 with over 8 million registrations. Registrations declined by close to 2.5 million until a slight jump in early 2016 and then the decline started again. .TEL is another whose decline started earlier, around 2011, but the decline was accentuated in early 2014 as new gTLDs entered the market. And .xxx peaked in late 2012 at around 120,000 registrations, dropped and recovered to around 115,000, and has since trended downwards with a dip down in mid-2014. And gTLDs aren’t the only ones with .us also seeing a significant decline from around a peak of 2.3 million in early 2014 to 2.2 million before jumping by around 400,000 registrations in recent weeks on the back of a discounted promotion. Even .net hasn’t been immune with registrations slightly declining for about two years from early 2014 before an uptick in late 2015 and another in recent weeks.
The study made a number of findings on prices such as the average and median retail prices for registrations of legacy and new gTLDs have declined, retail mark-ups over wholesale prices generally declined, the overall wholesale price level of legacy gTLDs is lower than that of new gTLDs and wholesale price levels for legacy and new gTLDs remained stable from Phase I to Phase II. The study notes the presence of price caps on legacy TLDs may help to explain the absence of changes in legacy TLD wholesale prices.
On registrations, the study found the shares of domain name registrations across registries, and across registrars, continue to be more dispersed for new gTLDs compared with legacy gTLDs. The study also found the share of registrations held by the top four, top eight, and top fifteen registries and registrars by domain name registrations has declined. Among registrars for domains in new gTLDs, Chinese registrars are dominating reflecting that Chinese registrants are accounting for around half of all registrations and Chinese registrants accounting for seven of the top ten known registrants.
Overall the study is quite dense and in reality, only the dedicated and the masochists among us will delve too deeply into it. But if you must, the study can be downloaded from: