Google’s most popular and least popular top level domains

70 thoughts on “Google’s most popular and least popular top level domains”

1. Pingback: Search Engine Journal » Google’s Most Popular Top Level Domains

2. Thanks for your comments. I wonder how many websites are active on each of the “low hit” country extensions?

   Thanks again, Gregg.

3. Thank you Gregg.

   Not too many working sites on many of those, though the names for some of those none functioning sites are interesting, like these under the Iraq domain:

   “myshoesizeisalargernumberthanmy.iq/”

   “www.google.com.iq/”

   “theusashutourwebsitedown.iq/”

   “www.saddam.husseyn.gov.iq/”

   “www.thankyougeorge.iq/”

4. Pingback: Domain Editorial » Stats on Google results per TLD

5. Interesting — the .uk has the most pages indexed of any ccTLD, nearly three times as many as the next-highest ranked ccTLD (.ca). However, .de has nearly double the number of domain names registered as .uk. I wonder if this reflects a bias (perhaps inadvertent, or at least benign) towards indexing English-language web pages?

6. Excellent list!

7. .ru:

   67 900 000

8. Pingback: Dominios más populares en Google

9. Thank you, Yurij. That was a typo. I’ve changed the tables to reflect that result.

   Thank you, too, Rand and Nadir.

   Geoffrey,

   That is an interesting stat. I’m not sure that it is an indication of search engine bias, or if it could indicate something completely different.

   For instance, I’ve seen a single page on a site using a Lotus Notes multiple views feature get indexed in Google thousands of times.

   Google will also include in its index a URL to a page that it has not crawled, so a typo in a link could look like it points to a page with a certain tld, when that page really doesn’t exist. There are a number of examples of that in the table of the tlds with the least page results. For example, “www.bloomberg.com.ap/” is included with the results for a search of site:.ap. Notice that the rest of the results for that cctld use the URL as a page title, and lack descriptive snippets. Chances are that means that those were indexed as links pointing to pages, rather than as pages that may have actually been indexed in Google.

10. I always wondered where my terrific online email company was registered..
    http://www.fastmail.fm

    Federal State of Micronesia.

    Very interesting research Bill.

11. Actually the list is missing .su, yes surprise the Soviet Union still lives on the web, with

    3,530,000 pages in google.

12. Thanks, Bruce. I’ve seen at least one doctor’s site registered in the Republic of Moldova. Guess he liked the “.md” address. 🙂

13. site:.pl
    I’ve got 41.000.000 results and you have 20.100.000
    I wonder why?

14. Thanks, Francisco.

    That’s a fair amount, and it looks like many of those are still in actual use. Unlike the 177 .bu (Burma), 310 .cs (Czechoslovakia), 181 .dd (Deutsche Demokratische Republik), and 15 .zr(Zaire) results listed in Google though those are retired/expired.

15. That’s an interesting list William.

    Am not clear about what to make of it.

    How many bottles of Visine was that again? 🙂

16. Hi Tom,

    That is a really good question. Here’s what I see on a search for site:.pl

    Results 1 – 10 of about 20,100,000 for site:.pl. (0.07 seconds)

    Chances are that we are using different data centers from Google to see our results, but that’s a significant number of different results.

    Articles Reprint,

    I wasn’t sure exactly what information I would glean out of collecting this information, and placing it on the web.

    But I have learned something about the cctld process, a little about global activity on the web, and some about how errors might produce some odd results in Google.

    I’ll probably try this again in a few months to see if the numbers change significantly, and think about if there might be any meaning to that.

17. Pingback: Dan's Archive

18. I tried adding the third column for all 267 tld/cctlds. I get 11,870,015,634. Google claims it has 80+ billion documents indexed. Something’s not right 🙂

19. Hi Marc,

    When Google last stated the number of search results that they carried in their index, I believe that their claim was to a little more than 8 billion objects (pages, images, sound files, etc.) This was sometime around this last fall. Yahoo! then noted that they had more than 19 Billion.

    Since then, Google has stopped telling us how many objects are in their database. This is what they tell us now:

    At Google we believe the essential quality of an index isn’t the total number of documents, but its comprehensiveness – which unique documents are in the index. So we don’t count duplicate or uncrawled pages. According to our internal testing, our newly expanded search index is more than three times larger than that of any other search engine.

    I don’t know if that was written before or after the Yahoo! claim to 19 Billion. I added up the number of results, too, and came up with the almost 12 billion that you have. I felt pretty comfortable with that total.

20. Hi again,
    Yes, we are using different data centers.
    Please try http://www.google.com/ncr
    I’ve got:
    7,720,000,000 for site:.com. (0.07 seconds)
    1,500,000,000 for site:.gov. (0.12 seconds)
    3,630,000,000 for site:.org. (0.06 seconds)
    40,100,000 for site:.pl. (0.04 seconds)

21. Hi – great stats!

    I wonder if there is a way to track direct navigation/type-in statistics of internet users searching information on the internet?

    Sam

22. I’m not getting the same numbers that you are, but I am getting different amounts based upon which data center I look at:

    Address Results
    http :// 64. 233.161.105/ 5,500,000,000
    http :// 64. 233.171.99/ 5,610,000,000
    http :// 64. 233.167.99/ 4,040,000,000
    http :// 66. 102.9.104/ 5,740,000,000
    http :// 216. 239.59.104 5,650,000,000
    http :// 216. 239.53.104/ 5,180,000,000

23. Hi Sam,

    Great question. There are a few ways to track internet usage, including the type you are asking about:

    Site Specific – using logs from a site to see how people use those sites. Tools from companies like Clicktracks can make that easier to do than a generic log file analysis tool.

    Network Specific – looking at the log stats that ISPs share with companies like Hitwise, to see where people go, what they do there, and how long they spend doing it.

    Surveys – like the ones that the Pew Internet & American Life Project folks use.

    User Specific – Where you have a very large panel of users whose movements on the web are tracked, including places they visit, how they get there (type-in, following links), and so on. That’s a practice that Comscore follows.

    Browser plug-ins also have some ability to capture some of this type of information. For instance, the Google Toolbar needs to track where you are to supply you with information such as what the pagerank of a site might be.

24. Hi William – thanks for the response. I contacted each of these companies and none do specific research on this. There’s gotta be someone who is tracking this information – otherwise there won’t be a huge aftermarket for domain names or a huge traffic conference for “domainers” – there is a reason why this is on the rise…

    Sam

25. I don’t believe that there is any centralized manner of collecting this type of information, or even clearly distinquishing whether a site was visited based upon an address being typed in on a guess that there might be something at the address, versus copied from a physical object (print article, label on a product, billboard, mention on radio or television) or from being saved as a boomark or favorite.

26. Thats what I am starting to find out…

    Although, I think there may be some creative ways of coming close to “guessing” the stats..

    Thanks
    Sam

27. You’re welcome, Sam.

    Guessing might be as close as you can get. Some sites just seem like they would get a lot of type-in traffic, like a “business.com.” Others might rely upon branding, or timely news.

    Good luck, though. There are some forums that focus on domain names, and brokering them. You might be able to find some discussion on those about how they estimate the value of a name.

    Cheers.

    Bill

28. Bill,

    I read the “we won’t declare our index size anymore” release but they did say that they had X times more pages than their closest competitor so I’m just baffled where these indexed documents are. If they can’t be found using site: search, then how?

    Thanks!

    Marc

29. Hi Marc,

    It seems that document amount changes from data center to data center, and probably suffers from an inability to identify and account for canonical URLs, and indexing pages on dynamic sites more than once.

    Examples:

    1. These might all be the same page, and yet still in the index:

    “http://example.com”
    “http://www.example.com”
    “http://example.com/index.htm”
    “http://www.example.com/index.htm”

    2. Dynamic sites might allow search engines to index the same page multiple times.

    3. Other software might cause strange results. For instance, the expand/collapse views twisties in Lotus Notes sites, that allow visitors to open and close sections of a page show different URLs when those sections are expanded and collapsed. Those little “twisties” that trigger the expansion and collapse point to different URLs, which could be collected by crawlers. The expansion and collapse of several sections, in different orders in the URL could be indexed by the search engine, resulting in possibly thousands of URLs indexed for a single page.

    Also, errors in URLs seem to account for a good number of pages that are indexed by search engines. Simple things, such as forgetting to include the “mailto:” in an email link, or a typo in the URL. As I mentioned above, I’ve seen results in Google for nonsense tlds, such as .xyz. Those aren’t accounted for in the list above.

    This is the Google company line on index size these days:

    Search engines’ published metrics for index size measurement vary greatly and are no longer easily comparable. Often, for instance, web crawlers retrieve duplicate entries for one page or links to documents that they haven’t crawled, and whose content thus isn’t in the index.

    I think the total size of their index is something that they are trying to keep secret these days. I also think that it may be almost impossible to calculate with any ease, especially if you are trying to not count duplicate pages or sites that haven’t been crawled yet.

30. That’s fascinating research, Bill, and provides a rich appreciation of the size of the Internet and also its particularity. Thanks for putting in the effort.

    I still believe these domain names have zero or almost zero impact on SERPs. What’s your current take on that?

31. Thanks Barry,

    I’m reminded a little of the Steven Wright Joke, that goes something like this:

    “I have a full scale map of the United States…

    …I keep it folded in my back yard.”

    Google may have their copies of the web – but they really aren’t exact copies – instead they are compressed indexes, and they definitely have flaws and are missing large parts of what is out on the web, while also containing duplicate pages and other errors.

    There’s been a belief that some tlds may have a slight impact on search engine results, such as the ones that are restricted to certain institutions – the .gov, .mil, and .edu. That might be true, but I would want to study it before saying so with any certainty.

    A regional flavor of Google (http://www.google.de for instance) is likely to focus more on sites with a .de cctld. Is that because of the cctld, or because the site is in German or both? I don’t know.

    It is worth exploring some more though.

32. November 2nd, 2006 at 12:36pm CT

    .br:

    105 000 000

33. Oops,

    Now I am getting 148,000,000 for .br
    I think this is a moving target.
    Sorry for previous post.

    –ErnstB

34. Hi Ernst,

    I think one of the things that I learned from this post and comments upon it was how different the amount of results can be from one search to the next, especially if you end up on a different data center for each if Google is doing load-balancing.

35. Great post, Reveals some interesting facts too. The actual registeration for .uk and .de is close (According to zooknic) but check the indexed sites .uk is almost times more than .de.

36. 1.1 million .edu’s and I can’t get any — lol

    Arggghhhhhh!!!!

    Great work with this list!

37. I did a count of how many sTLDs had been registered using the 3-month-old data at http://www.icann.org/en/tlds/monthly-reports/ and here are some of the results:

    .aero = 4575
    .pro = 5146
    .coop = 6097
    .jobs = 8016
    .cat 22,479
    .travel = 27,358

    As to how many of these unique domains are in use is a different exercise.

38. I think the mobi domain space will explode over the next 6-12 months, more so than .asia or .eu because of it’s international applications.

39. I’ve been wondering if people will start adopting mobi more quickly than they have.

    I’m not sure if the preference is on creating special sites for mobile devices on a mobi tld, or making present sites more friendly for handhelds.

40. i wonder hiw this is all going to change now that ICANN has authorized the creation of new ‘arbitrary’ TLDs… I mena, will people’s preferences shift towards the new names or will .com remain as THE most sought after TLD?

41. Hi trespies69,

    Interesting article on that subject here:

    ICANN to Add New Top-Level Domains, World to Come to an End.

    If the cost for one of these new top level domains ends up costing a 5 or 6 figure fee, we might not see too many of these custom top level domains. It will be interesting to see how this develops. My guess is that .com will remain the most sought after tld.

42. Great list, one thing I’ve found is that the .mobi domains are booming like no other right now. Most major companies are trying to make a mobile friendly version of their site.

    It would be really interesting to see an updated version of these stats, thanx again this helped!

43. Thanks, Adrian.

    I’m going to think about doing another update of these stats. Though I’ve wondered if they tell us more about the search engine than they do about the use of the domains.

44. I’m wondering about those .mobi stats. They’re quite low, if you ask me. However, I have no doubt we’re going to see a significant increase in .mobi very quickly in the next year to three years. .mobi sites are fast moving from “nice to have” sites to imperative and essential sites if guys want to stay ahead of their competitors. I know guys who develop .mobi sites and the way their business is increasing can only show that .mobi is going to be very big, very soon. Especially when I look at their client base.

45. Something that I’ve wondered about too.

    There’s been a signficant increase since this post was written. A site search on Google for the .mobi extension is now showing 8,870,000 results.

    I suspect that the numbers are even larger, but many pages may not have been indexed yet.

46. Well it seems Guatemala has no much websites, I’ve always have curiosity about how local country SEO works, I’ve put in the Webmasters Tools of Google the location of the site but the site is still new so will wait to see how that affects.

    Very interesting post, I like your blog 🙂

    Pablo Rosales

47. Thanks a lot for the link, I’ll check it out. I came first to this blog by accident but started to like it a lot, nice job 🙂 will be here every day

48. Hi Pablo,

    Thank you. The amount of results isn’t a reflection of the number of sites in a country, or the number of sites that are using a specific country related top level domain, but rather how many pages Google estimated were in their index.

    So, the numbers may be telling us more about Google, and about which pages they are indexing. Google is now showing 2,240,000 estimated pages with a “.gt” country code, up from 904,000 when I originally wrote this post.

    I hope your use of Google’s Webmaster Tools does help you.

    Ricardo Baeza-Yates and Carlos Castillo of Yahoo, have done some research on local country search, and how search engines are treating web pages from different countries. One of their papers is somewhat technical, but you might it interesting:

    Characterization of national Web domains (Scroll down to the bottom of the page, and click on “View or Download” to read the paper).

49. Thank you, Pablo. 🙂

50. Good stats. If only Google had a feature to remove double references at his urls, we could have even better and more reliable results
    Ross

51. Hi GR,

    It’s difficult to try to find out how many results Google might have for different tlds based upon the search results estimates that they provide.

    The following paper, and some of the references in it describe ways to try to estimate the size of a search index, which accounts for things like just looking at single URLs:

    Random Sampling from a Search Engine’s Corpus (pdf)

    This paper is also very interesting when it comes to look at results from country level tlds:

    Characterization of National Web Domains (pdf)

52. The Isle of Man’s strict policy over who can use .im domains may reflect in the low number. However, all that strictness didn’t stop somebody registering shop.im ahead of a government sponsored initiative.

53. Hi Paul,

    That’s an important point that you’ve made. Each of the different country level tlds have their own requirements, and those vary from one to another, with some limiting the number of sites using those codes. Thanks.

54. It would be interesting to see stats on pages per head of population. I suspect the richer counties would still have more domains per head than the poorer countries due to better internet infrastructure.

55. Hi John,

    I’m not sure if there’s been much research on that specific topic, though you may find some research by Yahoo’s Ricardo Baeza Yates somewhat interesting. He and his collaborators look at information about sites in different country domains:

    Link Analysis in National Web Domains (pdf)
    Characteristics of the Web of Spain (pdf)
    Link Analysis in National Web Domains (pdf) presentation
    Crawling a Country: Better Strategies than Breadth-First for Web Page Ordering (pdf) presentation

56. I was looking for a list of all the domain extensions. thanks.
    The additional information was very enlightening, there were some surprising results. i would have thought some extensions would be a lot more popular.

57. Hi Jonney,

    I’m a little tempted to look at this numbers again. In some ways, people are choosing cc tlds not so much to show where a site might be located, or to show what it might be about, but rather as a vanity tag. For example, someone with a site that shows lots of video might choose a .tv cctld.

58. Pingback: 7 Of The Rarest TLDs Around

59. Wow! Thanks for this Bill. Great to be able to place the South African (where I’m from) online landscape into global proportions like this. I’d ve very interested to contrast the number from 01/2006 to 5 years later (now) and see what type of growth we’re looking at on that global scale and nation by nation vs. the mean…. could be interesting to cross reference that with socio-political trends; eg. China’s growth in digital footprint vs. the mean; Burma etc.

60. Hi Matthew,

    The numbers above are an estimate from Google of the number of pages that they may have contained in their index five years ago. The numbers are actually misleading in a few ways, so they may not be as helpful as they may seem.

    For instance, a number of those pages may be duplicates. Google may not have indexed a number of pages that it should have. The numbers themselves are estimates from Google, after looking at a small percentage of the total pages.

    There are potentially other problems as well, but I still find the numbers interesting, if for no other reason than to compare the numbers from the different domains.

    A “site:.za” search at Google today shows 296,000,000 results, as opposed to the 16,400,000 from 5 years ago.

61. We have had a co.uk domain for many years now and bought a .co site last year but through referencing them the wrong way round took a hige dip in our ratings which we still are struggling to recover from (even though we’ve now referenced the domains the other way round for about 6 months). I’m just wondering if .co in itself is not a strong domain to have and could this be affecting our results??
    thanks
    Mary

62. Hi Mary,

    There may be some implications of using a different country code tld that could possibly affect rankings, but other things may have gone on as well that impacted where your site ranked.

    For example, if you moved content from one domain to another, and didn’t set up permanent (301) redirects from the pages of the old domain to the new one, then you probably lost a lot of the link equity (or PageRank) from external links pointing to the pages of the old domain. Even if you do that correctly, it’s not unusual to lose some rankings and traffics for a period of time when changing domains. That can be anywhere from a few weeks to months.

63. Thanks Bill – That’s really interesting.
    I’ll get someone to check our redirects.

    Mary

64. Hi Mary,

    You’re welcome. Definitely a good thing to do.

65. Hi Bill
    I found this list of all the domain extensions through a fast Google research.
    Some results were quite surprising, for example, I thought there were many more domains with the .net extension. Would be great to have an updated list, at least to compare it with these 5-year old results.

66. Hi Sandro,

    This isn’t a list of the domains that use different extensions, but rather a list of how many pages Google was estimating that were in their index at the time that I did the searches. I was wondering if I could find any interesting patterns when I did the searches, and I was surprised at the very low numbers of pages for some of the cctlds.

67. Interesting, using Bills method of determining the number of domains .cn now has almost 3 Billion domains. Now thats rapid growth.

68. Hi Joe,

    I wasn’t trying to determine the number of domains for each type of domain based upon the number of search results in Google. Instead, I was trying to estimate how many URLs Google might have indexed for each domain type. The first two lines of my post are:

    What are the most popularly used top level domains, or at least, which are the ones that show up on pages indexed in Google?

    I wondered this yesterday after seeing a news article stating that the registration of .cn (china) top level domain names topped 1 million for the first time ever by the end of 2005.

    So, there may be almost 3 billion URLs with .cn domains estimated by Google, but definitely not 3 Billion domains.

69. Very informative list Bill thanks for your efforts for collecting and sharing 🙂

70. Good list Bill, how did you come about to compile this? But anyways, great list. Looking forward to read more informative posts from you.

Comments are closed.