The first in a multi-part story about the geography of new TLD registrants.
Who registers domains? It’s a question we might answer in a thousand various ways.  One approach is to examine geography.  Where do most registrations come from?  Who are the “registrant nations”?  Who sits idly on the sidelines?  Which countries favor which  TLDs?  How strongly?  Why?
Painting this picture will take me several articles.  Here is the first.  I’ll be analyzing only nTLDs â the new domain endings released since 2014.  This study is reliant on raw data from nTLDStats, though most of what you find here you won’t find there.  Country information is derived from whois records â i.e. from the mailing address volunteered by each registrant.
It isn’t enough to be told that China and the USA are the largest registrant nations. Of course, we’d expect that.  They’re big.  They dominate the domain market.  (China has been grabbing my headlines since 2014.)  No, what we want to know is whatever we wouldn’t expect â what details weren’t always a foregone conclusion.
From nTLDStats, we learn that these 2 superpowers between them account for more than half of all registered nTLD domains: China at 44%; the USA at 10%.  But knowing that isn’t enough.  We need some benchmark to measure against.  Is 44% higher or lower than China’s due?  What about 10% for the United States?  Seems a poor showing, given the overwhelming preponderance of English keywords among the new extensions.  If registries were banking on a large American footprint, did they fall short?
Let’s connect registration volume with each country’s online population. This way, we can gauge what’s abnormally high or low; and we can compare big countries with smaller.
# | Country | Domains | % Share 1 | % Share 2 | % Users | Over |
---|---|---|---|---|---|---|
1 | China | 12.16M | 44.09 | 60.63 | 21.41 | 2.83 |
2 | USA | 2.79M | 10.12 | 13.92 | 7.42 | 1.88 |
3 | Germany | 611.1k | 2.22 | 3.05 | 2.19 | 1.39 |
4 | Russia | 466.3k | 1.69 | 2.32 | 3.26 | 0.71 |
5 | UK | 402.8k | 1.46 | 2.01 | 1.84 | 1.09 |
6 | Japan | 360.0k | 1.31 | 1.8 | 3.65 | 0.49 |
7 | France | 267.9k | 0.97 | 1.33 | 1.69 | 0.79 |
8 | New Zealand | 218.3k | 0.79 | 1.09 | 0.12 | 9.08 |
9 | Netherlands | 208.9k | 0.76 | 1.05 | 0.49 | 2.14 |
10 | India | 206.0k | 0.75 | 1.03 | 11.36 | 0.09 |
11 | Indonesia | 203.7k | 0.74 | 1.02 | 1.75 | 0.58 |
12 | Canada | 177.3k | 0.64 | 0.88 | 0.98 | 0.9 |
13 | Australia | 155.9k | 0.57 | 0.78 | 0.63 | 1.24 |
14 | South Korea | 128.8k | 0.47 | 0.65 | 1.4 | 0.46 |
15 | Turkey | 103.3k | 0.37 | 0.51 | 1.31 | 0.39 |
16 | Italy | 101.7k | 0.37 | 0.51 | 1.21 | 0.42 |
17 | Ukraine | 95,862 | 0.35 | 0.48 | 0.68 | 0.71 |
18 | Spain | 89,738 | 0.33 | 0.45 | 1.12 | 0.4 |
19 | Switzerland | 85,939 | 0.31 | 0.43 | 0.23 | 1.87 |
20 | Vietnam | 81,460 | 0.3 | 0.41 | 1.52 | 0.27 |
21 | Brazil | 77,921 | 0.28 | 0.39 | 3.8 | 0.1 |
22 | Austria | 69,212 | 0.25 | 0.34 | 0.22 | 1.55 |
23 | Cayman Islands | 46,067 | 0.17 | 0.23 | 0 | BIG |
24 | Belgium | 46,057 | 0.17 | 0.23 | 0.3 | 0.77 |
25 | Poland | 45,630 | 0.17 | 0.23 | 0.81 | 0.28 |
26 | Gibraltar | 38,602 | 0.14 | 0.19 | 0 | BIG |
27 | Hong Kong | 34,115 | 0.12 | 0.17 | 0.19 | 0.89 |
28 | Pakistan | 31,044 | 0.11 | 0.15 | 1.05 | 0.14 |
29 | Sweden | 30,165 | 0.11 | 0.15 | 0.27 | 0.56 |
30 | Panama | 26,501 | 0.1 | 0.14 | 0.06 | 2.33 |
31 | South Africa | 25,521 | 0.09 | 0.12 | 0.87 | 0.14 |
32 | UAE | 25,311 | 0.09 | 0.12 | 0.26 | 0.46 |
33 | Thailand | 25,311 | 0.09 | 0.12 | 0.83 | 0.14 |
34 | Armenia | 24,841 | 0.09 | 0.12 | 0.05 | 2.4 |
35 | Mexico | 24,368 | 0.09 | 0.12 | 2.26 | 0.05 |
36 | Norway | 24,145 | 0.09 | 0.12 | 0.16 | 0.75 |
37 | Romania | 23,319 | 0.08 | 0.11 | 0.34 | 0.32 |
38 | Czech Republic | 23,008 | 0.08 | 0.11 | 0.27 | 0.41 |
39 | Cocos Islands | 19,479 | 0.07 | 0.1 | 0 | BIG |
40 | Denmark | 19,027 | 0.07 | 0.1 | 0.17 | 0.59 |
41 | Israel | 17,969 | 0.07 | 0.1 | 0.2 | 0.5 |
42 | Singapore | 15,903 | 0.06 | 0.08 | 0.14 | 0.57 |
43 | Ireland | 14,271 | 0.05 | 0.07 | 0.12 | 0.58 |
44 | Greece | 13,760 | 0.05 | 0.07 | 0.23 | 0.3 |
45 | Cambodia | 13,654 | 0.05 | 0.07 | 0.09 | 0.78 |
46 | Lithuania | 12,933 | 0.05 | 0.07 | 0.06 | 1.17 |
47 | Malaysia | 12,523 | 0.05 | 0.07 | 0.67 | 0.1 |
48 | Hungary | 11,945 | 0.04 | 0.06 | 0.22 | 0.27 |
49 | Chile | 11,862 | 0.04 | 0.06 | 0.36 | 0.17 |
OK.  I’d better explain.  At first glance, nTLDStats lists the top 50 countries by nTLD registration volume.  On closer inspection, it’s only 49. There’s a gap: China is #1 while the USA is #3.  What’s going on? Why, it turns out 27.3% of all nTLD domains are masked by whois privacy.  With country data unknown in more than 1/4 of cases, we can’t declare precisely how many domains are registered in China or the USA or anywhere else.  But clearly China accounts for much more than 44% of nTLD registrations.  Among domains without privacy, where country info is exposed, China’s share is 60.6%; and the USA claims 14% not 10%.  (That illustrates the difference between the 2 columns above, “% Share 1” and “% Share 2”.) Together, these 2 giants own 3/4 of all visible nTLD domains.
Next, you’ll notice each nation as a fraction of the world’s internet users (“% Users”).  Large as it is, China constitutes only 21.4% of people who go online.  Hence China is overrepresented among nTLD registrations by a factor of 2.83 (“Over”).  Likewise, the United States comprises 13.9% of nTLD domains, though it contains only 7.4% of global web users; so, in a sense, the USA is roughly twice as interested in buying nTLDs as might be expected.
This begs the question: For which countries is nTLD presence most exaggerated? Answer: Small territories that are home base to large domain industry companies: the Caymans (0.17%) with Uniregistry, Gibraltar (0.14%) with AlpNames, and the Cocos Islands (0.07%) with .CC … and only 596 inhabitants.  Yet their share, however outsized, is minuscule.  Scraped together, this trio owns fewer than 1 in 200 nTLD domains without whois privacy.
New Zealand is overrepresented by a factor of 9 and currently owns more than 1% of visible nTLD domains.  That’s bizarre but explicable.  Almost all of those (92%) are .KIWI.  And .KIWI domains were registered in the hundreds of thousands quite recently as part of a controversial registry promotion.  After New Zealand, the next most overrepresented countries are (in descending order) China; Armenia; Panama; the Netherlands; the USA; and the German-speaking cluster of Switzerland, Austria, and Germany.
Here’s another way to look at things: internet users per nTLD domain registered (“Users Per”).  In the United States, 1 nTLD has been registered for every 86 connected citizens.  China has 1 nTLD domain per 57 people online.  New Zealand’s glut puts the ratio at 18 kiwis per .KIWI.  Meanwhile, there’s just 1 nTLD domain in India for every 1784 potential online visitors.  And Mexico, most of all, has spurned the nTLDs, with only 1 domain per 2994 Mexicans.
# | Country | Rank Users | Users Per |
---|---|---|---|
1 | China | 1 | 57 |
2 | USA | 3 | 86 |
3 | Germany | 9 | 116 |
4 | Russia | 6 | 226 |
5 | UK | 10 | 148 |
6 | Japan | 5 | 328 |
7 | France | 12 | 204 |
8 | New Zealand | 83 | 18 |
9 | Netherlands | 37 | 75 |
10 | India | 2 | 1784 |
11 | Indonesia | 11 | 278 |
12 | Canada | 22 | 179 |
13 | Australia | 33 | 130 |
14 | South Korea | 14 | 351 |
15 | Turkey | 15 | 409 |
16 | Italy | 17 | 386 |
17 | Ukraine | 29 | 230 |
18 | Spain | 18 | 404 |
19 | Switzerland | 55 | 85 |
20 | Vietnam | 13 | 605 |
21 | Brazil | 4 | 1576 |
22 | Austria | 57 | 104 |
23 | Cayman Islands | 194 | 1 |
24 | Belgium | 47 | 209 |
25 | Poland | 27 | 575 |
26 | Gibraltar | 202 | 1 |
27 | Hong Kong | 64 | 182 |
28 | Pakistan | 20 | 1095 |
29 | Sweden | 48 | 294 |
30 | Panama | 110 | 76 |
31 | South Africa | 24 | 1109 |
32 | UAE | 50 | 330 |
33 | Thailand | 26 | 1056 |
34 | Armenia | 114 | 71 |
35 | Mexico | 8 | 2994 |
36 | Norway | 72 | 209 |
37 | Romania | 45 | 467 |
38 | Czech Republic | 49 | 373 |
39 | Cocos Islands | N/A | N/A |
40 | Denmark | 69 | 287 |
41 | Israel | 61 | 354 |
42 | Singapore | 78 | 289 |
43 | Ireland | 84 | 263 |
44 | Greece | 54 | 532 |
45 | Cambodia | 91 | 217 |
46 | Lithuania | 108 | 159 |
47 | Malaysia | 31 | 1721 |
48 | Hungary | 56 | 601 |
49 | Chile | 43 | 973 |
Above, you’ll see each country ranked both by nTLD registrations (“#”) and by the number of internet users (“Rank Users”).  Thus, although India is second only to China in terms of citizens online, India’s nTLD footprint ranks only 10th.  Among nations large and active enough to appear in the top 49 by nTLD registrations, the 3 proportionately most underrepresented â or, at any rate, least interested in buying nTLDs â are Mexico, India, and Brazil.  A fascinating discrepancy, I think … worth more research.
Want a thumb rule for a country’s proportional nTLD footprint?  Here you go!  New Zealand = 9.  China = 3.  USA = 2.  Netherlands = 2.  German speaking = 1.5.  UK / Canada / Australia = 1.  Brazil = 1/10.  India = 1/10.  Mexico = 1/20.
To be continued…
nTLD In-for-mat-ion o-ver-load!
Interesting data though.
No China, bad news for nTLD’s. No surprise there.
@Mark Thorpe,
Think that’s a lot? My full study is being parceled out in tiny installments!
In today’s domain industry, there aren’t many professionals interested in research. Mainly a lot of part-time door-to-door salesmen, which is what domainers are (let’s face it). And that’s fine as far as that goes.
But at some point, somebody’s got to drag the data into the daylight. That’s what will attract serious professionals from outside industries â professionals who are used to rigorous reports and who, without that, tend to dismiss the domain industry as nothing but a pack of “cybersquatters”.
The housing market doesn’t consist only of realtors. At some point, analysts show up and REITs. E-commerce isn’t all online shops. Economists are commissioned by investment firms and universities to conduct studies.
Many domainers will hate this. But, like it or not, this kind of vivisection is part of the growing up process for the industry as a whole.
Couldn’t agree more.
@Whois
Not all domainers are door-to-door salesmen, most of my sales are inbound. But thanks for assuming otherwise.
I also don’t hate the data, i am just not a fan of nTLD’s.
I used to be a financial analyst for a big corporation, along with being an Investment and loans specialist for a large bank.
Just don’t forget to show some love for the original TLD’s like .com .net .org. If it was not for them, there would not be any nTLD’s.
I have been seeing a lot of .XYZ .Club and other nTLD bromance lately among domainers.
But In a place where I call reality, most end-users want .Com.
Cheers
@Mark Thorpe,
“Just donât forget to show some love for the original TLDâs like .com .net .org.”
When I write in praise of .COM, people call me a shill and an enemy of progress.
When I write about nTLDs, people interpret that as an “nTLD bromance”.
As long as I can alienate everybody for diametrically opposite reasons, I think I’m on the right track.
Interesting data. I am a new domainer with passing interest on countriesâ registration and interest level in the DNS industry and I have been closely following the registrations via namestats specifically for .shop.
.shop seems to be a big winner with the cluster of Netherlands-Switzerland-Austria-Germany. An interesting trend, given that these are countries that have traditionally been content working within their ccTLD namespace. The top 10 doesnât seem like a surprise at all except for that New Zealand oddity but that has been clarified đ
Don’t forget about .Store as well. đ
No thanks đ
I just donât feel .STORE but .shop just seems right for the ecommerce websites of the future, especially mid level ecommerce stores. I am also buying a couple of .VIPs. It is a risky investment given that I am new and buying something which is really a âwildcardâ but the upside is that I get domain names I could only dream of in a .COM namespace. Maybe some day, someone might find a use for them. Or NOT. But if I lose money, it wouldnât the worst investment mistake I have ever made.
Don’t feel .Store? Apple store, App store, iTunes store, Google Play store and so on.
Well all those are mostly trademarked names. The really good generic ones have been reserved by registry or taken. With .shop, you simply have a goldmine. I know uptake will be slow unless they do .xyz style marketing blitz but I am confident they will be worth something 2-3 years from now. Though I am not too closed minded about .store. I can still buy a good .store as long as it’s not ridiculously priced.
“a lot of part-time door-to-door salesmen, which is what domainers are ”
Nothing like insulting your readership.
Many of the part-time domainers I have met are skilled by profession* in crunching data.
* engineers, mbas, accountants, real estate agents, etc.
I was waiting for someone else to say something about door-to-door salesmen. Lol
@Meyer,
Don’t be so touchy.
I’m a domainer, and I knock on doors too. You might call me self-deprecating â but “insulting”? Hardly!
Why so defensive about how other domainers “are skilled … in crunching data”? As if I said nobody could perform long division …
The point I was making is simply this:
Most domainers are busy trying to sell stuff. Whether they’re “skilled in crunching data or not”, how many of them have the patience to thoroughly read this kind of article? Only a small fraction.
And how many domainers PAY for dense, numerical articles? Zero. Within the domain industry, there is no market for market analysis. In a real sense, the domain industry cannot afford this and doesn’t support the activity. I always write at a loss.
Of course I know that many domainers can read and understand a piece like this. A few of you may even find it interesting. Would I trouble to write it otherwise?
But the fact remains: How many domainers take a day off to perform studies like this? Almost nobody. Financially, it makes no sense for them to do it. So the industry doesn’t produce this kind of content.
Unflattering perhaps, but it’s true: In 2017, the domain industry cannot economically support many roles outside of sales. There’s a thin layer of news sites, but most bloggers struggle to justify even part-time writing. Compared to other industries, the domain space is underdeveloped in this regard. No surprise: It’s a young industry.
It doesn’t matter how many domainers can read an article like this. Nor even how many domainers could produce articles like this. What matters is whether the industry can fund articles like this. It can’t.
@shopman
I would own both .store and .shop version of a domain name if possible IMO. That way you have leverage.
For the leverage, that would be a smart strategy!
I would be careful when trying to sell .shop domain names.
In GMO’s (.shop registry) application to buy .shop extension, GMO said they would frown upon people profiting from the selling of .shop domains.
The company said it would do random spot checks to make sure no registrants are breaking this rule.
I do not know if this rule is still in effect or not, seeing they have been giving away their .shop domains lately for free or almost free.
.Store (Radix Registry) does not have this no selling rule đ
When applicants were submitting applications for their new gTLDs, they said a lot of nice things that would put them in good standing with ICANN. I think GMO would be committing suicide if it is closing the registry to domain name investors. They initially said that but I think they changed their minds a long the way and did not include that in the TOS. I donât think there is any stipulation in GMO TOS that says .shop domains cannot be resold or parked which is really the lifeline of the industry https://nic.shop/policies/registration_policy.pdf. Domainers add vibrancy to any namespace.
Every registry wants a productive namespace and in so far as they want to push their objectives in making ,shop an ecommerce namespace that hosts mostly ecommerce stores, they also recognize they need âbottom feedersâ like domainers for the growth of the registry :-). It will be sort of âdonât ask donât tellâ relationship between GMO and domainers.
The messaging might be strongly towards developing domain names while looking the other way when a domainer also buys some domains and lists for resale; unless of course they are shi**ing in GMOâs backyard like registering trademarks.shop. Maybe I need to ask them directly before buying more domain names.
A couple of things to consider when determining country – maybe ensuring that registrant data in the whois is accurate. Not every registrant provides accurate information. How you do define country if content of website displays physical address in country A and registers domain name using Country B (assuming its accurate) – so my point is, I think there are a number of variables to consider when determining country.
Also, how do you determine country for domains registered using masked whois information? Use content of website?
@Samantha Frida,
You raise good points. People ought to be critical of these numbers and conclusions â because I’m fallible, because interpretation is partly subjective, and because my statistics are only as true as the underlying data.
It’s important to note here, that I’m making no attempt to determine country. Rather, I accept the country determinations and counts made by nTLDStats as a starting assumption. My study only derives conclusions from their data as an undisputed premise.
Any flaws in nTLDStats’s country determinations or counting methodology DO affect my statistics and conclusions. And there are bound to be some flaws with nTLDStats.
Personally, I suspect that some countries may have been over-counted based on whois proxies that weren’t labeled as such by nTLDStats. At this point, I haven’t confirmed that hunch; and I may be mistaken. It’s only an hypothesis. But my statistics for .CC (see above) are a red flag. More than 19,000 domains for fewer than 600 registrants? Dubious. One reason for conducting a study like this one is that, by teasing out ratios like User Density per Domain, we can identify outliers like .CC and .KIWI as suspicious.
“Also, how do you determine country for domains registered using masked whois information?”
nTLDStats sets aside domains with whois privacy. In fact, “whois proxy” is listed as the #2 country by nTLD registration volume â more than 1/4 overall. Consider that a big lump of “indeterminate”.
In this article, I have made no attempt to guess the country behind whois proxies. (In a later article, I will examine that question.) What I HAVE done here is to express each country’s nTLD % share in 2 ways: (1) known percentage of all domains; (2) known percentage of domains without privacy. (You’ll find the numbers side by side in the first table.)
A simplistic assumption would be that all countries use whois proxies equally; and, in that case, you can read “% Share 2” as the true fraction. But in reality, different countries use different registrars; and different registrars have different rates of whois privacy usage. So a country’s true share of nTLDs cannot be precisely known. What we do know is that it’s AT LEAST equal to “% Share 1” and, on average, will be equal to “% Share 2”.
Some countries will be above “% Share 2” (but never above % Share 1 + 27.3% because no country hogs 100% of whois privacy). Other countries will be below “% Share 2” â in fact, somewhere between “% Share 1” and “% Share 2”.
Hope that answers your question.
Joseph Peterson.
This type of reporting is exactly what the industry should be putting out vs
pure spec on both sides. I support both old and new and greatly appreciate just the facts, globally. So Thanks.
Yes indeed, “part of the growing up process for the industry as a whole”.
Quite a bit of sales and use information is being posted on NP to track New “G”s so your not alone in the quest for relevant data.
Happy Hunting!
@168,
Thanks. Yes, a lot of different people are contributing useful information about the market â news, sales reports, etc.
That goes all the way back to Ron Jackson at DNJournal and continues with services like NameBio and nTLDStats; researchers like Jamie Zoch at DotWeekly; and, of course, journalists like Andrew Allemann here at DNW; and lobbying / activist groups like the ICA. Obviously, they’re not alone. Lots of people engaged in different ways with more than just selling their own stuff.
Imagine if we were all in the dark, just selling privately with no bird’s eye view of the industry!
Thank you for your work, Joseph. Nom, nom, data. All and any.
Yes, those who are not door-to-door salesmen are often escaped professionals from other research-focused disciplines. Old people, to put a finer point on it. Those who came later to domaining don’t have the luxury of sitting on a chaise longue eating grapes and just answering incoming emails. It reveals the insularity, lack of empathy, and brittleness that can infect the career domainer to take offense at your characterization. (What can one do. Eckhart Tolle doesn’t work overnight. And we’ve gotten too drinky for hot yoga.)
China is over-represented in nTLDs because in China it is still a bit 1998 and $1000 is a worthwhile payday there. (Pre-ACPA, pre-Verizon lawsuit, from an enforcement standpoint.) A scan of UDRPSearch respondents backs this up. So does a scan of the Fortune 500. Virtually all of the 500–US, France, Japan, Spain, Russia, Germany, UK, etc.–have their .com locked down, but China stands alone in that a couple of its giants are still stuck with .com.cn. Sitting on their hanzi, as it were.
Any time someone discusses nTLDs as investment assets, as you’ve done here, the real concern is for newbie domainers. It is important not to recklessly over-invest, because the value of these assets is capped by the success level of the end user. When startups use nTLDs, they do it because they lack money and foresight. They rarely succeed, in the absence of both, but in the event that they do get some investment, they adopt the .com. The Fortune 500, being virtually all .coms. tells you all you need to know about the axes you’re plotting. Companies do not broadcast that they are undercapitalized for one second longer than necessary.
@C. S. Watch
I was a bit surprised to see people taking offense at my description of “part-time door-to-door salesmen”.
Unquestionably, many domainers and domain brokers spend time contacting strangers and inviting them to buy a domain. That’s the way this works. Active outreach means knocking on doors. Some people earn a living doing this. Most do it part time. And, of course, there are domainers who only wait passively for inbound offers; but often they too entrust their premium assets to brokers who go out and knock on doors for them.
Why on earth should anybody look down on this activity? Do the people who took offense at my remarks believe that knocking on doors â i.e. approaching strangers to sell domains â is shameful? I certainly don’t.
Why would anybody think that? It’s not as if “poor newbies” are the only people whose business model relies on active outreach. Top brokers at Sedo like Dave Evanson definitely knock on a few doors (metaphorically speaking) in order sell premium domains for 6 or 7 figures. Elliot Silver invites people to buy what he owns. Many very successful brokers and domain investors are (among other things) active salesmen. Does anybody really think that’s undignified? Or seriously believe that I think that?
“Any time someone discusses nTLDs as investment assets, as youâve done here, the real concern is for newbie domainers.”
Of course, newbie domainers ought to be careful and do their research.
But, for the record, this article of mine literally makes no reference to domain investment or resale whatsoever. I don’t mention the aftermarket or even the primary market (registration and renewal fees). You’ll notice I haven’t discussed money or pricing or quality at all. This article is about the name space â not about domain investment.
Righto, apologies. I lost the reins on my own subjectivity long ago. (Such is the nature of addiction.)