The sample of 1.7 million domains comprised IDNs in gTLDs (both top and second level), and .eu (second level).
The report found that on average, 30% of IDNs (regardless of level) were both “in use” (ie had active name servers) and had sufficient web content to identify the language. In other words, more than half of IDNs with active name servers do not have web content, so measuring active nameservers alone would overstate the numbers for language of web content.
The largest gTLD IDN space is .com which (combined with .net) comprises 1.3 million
IDNs (second level), of which 30% had active web content. The percentage of active
websites in other second level gTLD IDNs (.asia, .biz, .info and .org), comprising
160000 domains, is 47%, similar to that of .eu (second level) which is 48%.
Within new gTLDs, there were 35000 IDNs registered at the second level. Of these,
only 16% have active websites.
From the ccTLD world, the Viet Nam ccTLD registry (VNNIC) publishes usage statistics
for IDNs registered under .vn (second level). VNNIC operates a give-away policy for
.vn IDNs, resulting in high levels of registration – .vn is the second largest IDN space in
the world. However, only a small proportion of .vn domains have active nameservers
(16%) and even fewer have active web content (13%). Therefore, the policy of a TLD
is relevant, and in particular “giveaway” policies are likely to be associated with lower
levels of active usage in the medium to long term.
Top level IDNs
The majority of domains in our sample of 1.7 million IDNs were at the second level, with
only 91000 full IDNs (all recently launched new gTLD IDNs). Of these, 18% had active
web content. There is substantial variation in usage rates between 0% (lowest) and 64%
Evidence indicates that the age of the TLD influences rates of active nameservers.
Variations in usage by script
There were variations between script, with Han, Katakana and Hiragana (associated
with Japanese language) domains most likely to have active web content (49%) and
Devanagari domains least likely (4%). Evidence from the ccTLD world (second level
domains), also shows differences in usage rates by script. In the sample of 50000 .eu IDNs, we found active web content for 50% of Latin script .eu IDNs, 34% for Cyrillic
script and only 10% for Greek script. However, there would be limited value in drawing
conclusions from such low numbers of Cyrillic and Greek script .eu IDNs.
Red.es informed us that 68% of IDNs under .es (Latin script, second level) had active
web content, including redirects.
Domains with active web content
The Russian ccTLD registry publishes detailed statistics on usage, which indicate that
up to 57% of .рф domains are in use, according to our definition. This assumes that
“parked” domains (21%) have enough content to determine language, and that “under
construction” domains (14%) do not. If parked domains are excluded from the analysis,
36% of .рф domains have active web content.