.EU Web Survey Results

[jmccormac]

I've just finished a websurvey of 2.3 million .eu domains which had, approximately 1.98M websites.
The results are below. In real terms, the .eu ccTLD is not a ccTLD at all and is more like a fragmented set of markets.

A - Active / unclassified: 13.28%
B - Brand Protection Registrations: 8.43%
C - Compromised: 0.01%
D - In page Redirects: 4.36%
E - External TLD Redirect: 9.22%
F - 403 etc: 2.06%
H - Holding Pages: 23.39%
I - Internal Site Redirect: 0.55%
M - Matched External TLD Redirect: 4.22%
N - Duplicate Content: 3.00%
O - Other TLD site served as .eu: 0.95%
P - PPC parking and monetisation: 10.44%
R - Redirects / unclassified: 1.74%
S - For sale/rent: 1.41%
U - Unavailable: 0.21%
W - Possible hijacked site: 0.02%
Z - Redirect to other .eu website: 2.23%
NS - No website: 14.50%

Regards...jmcc

[Webwork]

Always interesting jmcc. Keep 'em coming. Knowledge is . . an aid to wiser decisions . . sometimes . . ;)

[jmccormac]

Eurid came out with mickey mouse survey of 5000 .eu sites and tried to claim that it had 36% business usage or something equally ridiculous. As a ccTLD, .eu is dead. The only way to describe it now is a gateway TLD that is used to guide users to the relevant sites. About 21.87% of sites are either Brand Protection/External TLD redirects/Matched External TLD redirects. The approach Eurid uses is the Finite Monkeys one - get a bunch of students/employees to sit down and categorise sites. This is sheer muppetry. Large scale surveys have to be automated and they have to be run and designed by people who understand search engines and categorisation. These people do not. The way that they waffled (obvious Belgian pun) about how their 5K survey scales up statistically is rubbish. This is because people break HTML in all sorts of wonderful ways and almost every small webhoster has their own variant of a holding page. There are over 1500 different REGEXPs used as part of the analysis here and that's only part of the process of categorisation of domain usage. The O factor, where sites from other TLDs are served as .eu sites is user transparent. It actually takes some linguistic analysis and link analysis to detect them.

It has replaced .com, to some extent, as the second choice TLD in some of the Eastern EU countries but it has effectively flatlined in most of the EU countries where growth in their ccTLD and com would take up about 80% of growth in any month.

I've been working on a timescape algorithm for assessing web growth in TLDs and it is showing some promise. But at least 40% of .eu has not been updated in 2011 if the early results are correct. (Combination of Last-Modified data from webservers and copyright date cues from the HTML.)

I've dug out the data from the .eu survey I ran in 2007 to compare and about 33K sites were unchanged (quick comparison of HTML hashes). I think I have another survey from 2008 around somewhere. In the monthly Irish webscape survey that I run (about 9K .eu doms out of about 350K domains surveyed), .eu always resembles .biz in terms of usage.

What is interesting is that Ovidio and some of the other early market speculators have dropped their portfolios or have stealthed them. Most of the UK front company .eu domains that are left (about 17K or so) are on an aftermarket domain sale site and few of them are getting any interest at all.

I should have spent my time surveying the few million .co.uk sites that I track. :)

Regards...jmcc

[HuskyPup]

Interesting...I have been constructing my new company example.eu flag ship site this year and I am more than pleased with the results.

GooBinHoo have all treated it as a golbal TLD and the site has done extremely well even now displacing my 17 year old example.com for some of my widget terms.

It has been very noticeable that real world business enquiries have increased through this extension however I do specifically state that the office is located in one of our major producing centres in Italy.

[jmccormac]

It has become a gateway TLD, HuskyPup,
While it doesn't have a lot of content in terms of Active/unclassified sites, a lot of big companies now use it as a gateway to redirect users to the relevant ccTLD site. It is also a very cheap way to access the EU market.

There's a vast difference between the statistics and the topology of the web. This is the point that I was making (using the muppetry and Finite Monkeys terms). Site usage tends to be highly specific so trying to classify the usage of a bunch of sites as business/community/institutional etc and then extrapolating that to the entire TLD causes major problems simply because many sites are unique.

It also tends to freak people like me out (especially after having just finished categorising almost 2 million .eu websites) because we see the whole of the web and know how people break HTML. Categorising sites on a large scale (5000 sites is not a large scale) is complex work. One person could categorise the 5K sites. This is how many web directory operators do their seed set. However it can be very boring work.

The imprecision of describing a site as a business site is also a major problem because what differentiates a one page site for a carpenter from a site like Homebase or Argos (or any big brand shop with loads of products)? Now you can look for shopping cart signatures where links to known shopping cart software will appear but it also requires analysis of other links to be sure. Last year in Ireland, a SEO agency quoted a government stats agency claim about n% of businesses having sites that were e-commerce ready. This was rubbish because much of the new website development comes from Sole Traders rather than companies and the survey methodology of the government agency was simply to ring and ask if they had a website or something equally iffy. The web is far more complex than these people realise. Not every website is a shopping site and not every website should be. The one page website with contact details can often be a far more effective sales tool for electricians, plumbers, carpenters, accountants, lawyers than a well designed, multi-page website. This is because people have different requirements when searching for such services. An all-singing, all-dancing website with online ordering will only complicate matters when people just want to call a plumber to fix a leaking pipe.

I think that I mentioned a few years ago that it would take between five and ten years for the .eu to get over the Landrush fiasco. It is beginning to do so but it will take a long time for it to gain serious market share in Ireland and the UK. Though for accessing the Eastern European countries, it is better than a .com because of the way that they've adopted it as the second choice TLD after their local ccTLD.

Regards...jmcc