Great point, thanks jmcc
the question is how to implement it: one can easily learn how many registrations there are for a specific TLD, but how can one - roughly - assess how many actively developed websites there are in that TLD?
or which one between two TLDs has more actively developed websites?
Are there any indicia that could be of help?
It is one of the most difficult things to do with gTLDs because many of them have multiple country-level markets. The .COM may appear to be a global TLD but that's just the tip of the iceberg. The .COM, as with most large gTLDs, is a composite market of a global market and many country level markets. The development/web usage rates of domain names from those countries reflect the economic conditions of the market. The largest player in .COM is the US market so web usage for US registrations tends to be quite good. The .COM is effectively the US country code TLD (it has .US as the official one but that's at about 1.9 million registrations).
For a large gTLD, the way that web usage is measured is sample based. A sample of domain names is taken (typically over 100K but I use 150K) and the usage of websites on these domain names is broken down by categories. The most basic are:
No site/no response
Active
Holding pages
Not found/forbidden
PPC parked
Sales
Affiliate landers
Compromised
Redirects
It is possible to break these down into simpler categories:
No content
Active
Templated content
Redirects
I run web usage surveys each month but only run large surveys on the new gTLDs every few months. The last major new gTLD survey was from October with full surveys for all new gTLDs with less than 500K registrations and 150K samples for those with more than 500K registrations. I've been working on adding them to the site. The 150K surveys for COM/NET/ORG/BIZ/INFO/MOBI/ASIA and some ccTLDs will also be added in the next two weeks.
Accurately measuring usage is extremely difficult. The claims put out by CENTR (the European ccTLD registry organisation) about usage in their ccTLDs and gTLDs is unreliable and based on simplistic methodology. The usage rate for .COM is approximately 30% but the .COM is a high quality gTLD. The ccTLDs are different because they have geographically concentrated markets and can have usage rates over 30%. The usage on other gTLDs varies and some of the new gTLDs have very low usage rates.
The web usage is only part of how a TLD's performance is measured. The renewal rate is also important. Some new gTLDs, the ones with very heavily discounted registration fees, can see over 80% of domain names in a zone deleted in a year. Others TLDs will have much better performance with over 30% renewing but the main gTLDs tend to have over 50% renewal for the first year. The blended renewal rate is based on all domain names renewing/deleting in a month and can be estimated from the ICANN registrar reports with the simple formula (renewals/(renewals+deletions)*100. Some of the main gTLDs have blended renewal rates over 70%. That's similar to many of the strong ccTLDs.
With the advent of freely available AI services, it is going to get a lot more difficult to measure web usage because it will be used to generate websites that will appear to be real websites.
Regards...jmcc