The .za Domain Name Authority (ZADNA) recently issued a public discussion document requesting public submissions on direct second level registrations (SLRs) in the .ZA domain name space. What this effectively means is that, at some stage in the future, you could be able to register "yourname.za" in addition (or alternatively) to "yourname.co.za". Shortly after the Discussion Document was released, we published a blogpost on the complexities of introducing SLRs into .ZA.
The deadline for public submissions was 23 April 2018 and we submitted our final submission on the same day. You can view our full submission here: DNS' SLR Response to ZADNA:
One of the key focuses of the ZADNA Discussion Document (section 3) was the recent SLR experiences of other ccTLDs, such as Colombia (.co), United Kingdom (.UK), New Zealand (.NZ), Kenya (.KE) and Australia (.AU). We took a closer look at these examples and discovered some interesting facts:
(Colombia) We believe that it is not useful to benchmark the developments of .CO to what is happening in .ZA or other traditional ccTLDs. The .CO TLD has completely lost its identity as the country code domain for Colombia, rather opting to exist as a generic commercial alternative to .COM. In essence .CO has become more of a quasi-gTLD and its comparison to SLR developments in traditional ccTLDs, such as .ZA, is not helpful.
(Kenya) The .KE SLR experience has been, by all accounts, a disaster. Not only has the take-up of SLRs been disappointing, but the administrators also managed to severely compromise the technical functioning of some of the existing SLDs, including CO.KE. The .KE experience is an example of how not to implement SLRs and it is safe to say that if we repeat their experience in .ZA, it would undoubtedly lead to massive economic losses and disruptions, which will in turn severely compromise the trust and goodwill already developed in the .ZA namespace.
Most of the criticism that participants had of the .KE SLR experience relates to how the SLR policy was conceived, communicated and implemented. Specific issues, include:
- The registry announced the launch with no clear plan or rules set out.
- Originally they did not even have a grandfather period for any other [sld].ke registrants and they did not understand the sunrise, landrush, GA phase model.
- They took on suggestions of a grandfathering period as well as an auction mechanism, however, they did not understand these concepts and proceeded to announce these only after the Sunrise launch.
- They did not understand how the grandfather period's rights work and rejected any application for domains they felt should be charged at a premium (regardless of registrant's pre-existing rights).
- There was actually no landrush phase as they added a list of domains to the auction phase which meant the registrant had to bid and open an auction for the domains on the auction list whether there were other bidders or not. If a domain name desired during this 'landrush' phase was not on the auction list then you were required to contact the registry and they would add this to the auction list.
- They did not block/reserve any domain names from being registered. This meant that any new registrations made by a registrar were successful. The registry then had to manually compare these registrations with the list of reserved names and then unilaterally deleted the conflicting domain names even if these names had already been made active.
- The auction system required the registrant to apply and participate and pay the fee (if won) directly to the registry. Once this was done, the registrant then had to find a registrar and pay them a fee to 'register' the domain. Again, anyone could register the domain name won at auction and only after the GA, the registry reviewed and removed any domains which were not won at auction or were registered by another registrant.
- The GA period was delayed (with no notice provided to registrars or registrants) as the registry managed to somehow delete 100's of .ke and, in some cases, co.ke domain names while approving the previous phase applications. No communication was provided and they subsequently launched the GA phase after 22:00 that night without any prior warning.
(United Kingdom) The SLR experience of .UK is still a work in progress, with the grandfathering of CO.UK domains only ending in 2019. The following graph charts .UK registrations since the introduction of their SLR program in 2014:
As can be seen, CO.UK registrations started dropping off immediately after the introduction of SLR’s, and three years down the line, the total .UK domain space has only grown by a mere 1.61% as opposed to 3.3% in the year prior to SLRs being introduced.
As at May 2017 (3 years after their initial launch) the total number of .UK SLRs only amounted to 637,360 registrations. This accounted for approximately 5.9% of the total .UK domain space and according to some commentators, the .UK SLR program was widely regarded as a “flop”
.UK 3LRS = 10,039,279 (94.1%)
.UK SLRs = 637,360 (5.9%)
Total .UK = 10,676,639
Something drastic had to be done. For July, August and September 2017, Nominet (the .UK administrator) offered its registrars a 2-year free registration period for unclaimed (grandfathered) SLRs. This lead to what is clearly an abnormal spike in the SLR registration rate, adding a further 1.2 million new SLRs into the zone. The number of ‘paid’ 3LRs continued their decline during this period.
It appears this ‘large’ spike in registrations was created by automated registrar systems ostensibly on behalf of registrants. In other words, because it was free, the registrars took preemptive advantage and registered these names, even without the registrant’s permission or knowledge.
.UK 3LRS = 9,980,134 (82.5%)
.UK SLRs = 2,119,904 (17.5%)
Total .UK = 12,100,038
At first glance, it appears that this promotion was an effective way of increasing the number of SLRs and the total .UK zone count. However, the truth of the matter is that this ‘promotion’ merely served to obscure the true uptake of .UK SLRs while at the same time decreasing the number of actual paid (3LR) names. There is simply no way of knowing how many of these new ‘free’ registrations will be renewed and retained after the expiry of their initial periods.
What is clear is that since the SLRs launch in 2014 the ‘paid’ domain registration rates appear to be down for the total .UK space. The total paid domain growth in the whole .UK has only been 2.11% over the 3.5 year period.
In conclusion, even though the SLR numbers may appear to be substantive, they still only account for roughly 18% of the total .UK namespace. More concerning is that they have only reached this number after 4 years of availability and basically giving away more than half of these registrations.
As the 2019 grandfathering deadline approaches there is no-doubt substantial concern in the UK as to what will happen to the remaining grandfathered domain name strings that have not yet been claimed by the 3LR registrants. Will we see a last minute rush to claim these names, or will there be a massive landrush by speculators and cybersquatters? If the latter, could we see further degradation of the .UK namespace?
As a matter of interest, .UK 3LRs still consistently show more new registration numbers (creates) each month:
A more detailed analysis of the .UK experience is available [here]
(New Zealand) The SLR experience of .NZ has matured a bit more than that of the .UK, primarily because the Preferential Registration Eligibility (PRE) period has already expired and the the bulk of .NZ SLRs are currently available on a first-come-first-served basis. Although the SLR take-up has produced a significant number of delegations (over 134,781) the question is what has happened to the balance of the CO.NZ domains that are still available as corresponding SLRs?
The graph below shows the year-on-year domain count and total growth since just before the introduction of SLRs, until 2017. There was a substantial SLR growth spike in the first 2-years, after which it tapered off dramatically. Essentially 77% of the current SLR domain count is attributable to the first two years of registration (2014 & 2015).
However, the growth statistics belies the fact that New Zealanders still prefer CO.NZ above .NZ SLRs. This is substantiated by the fact that monthly new creates are consistently 3x higher for CO.NZ domain names as opposed to .NZ SLRs. The demand is still clearly on the 3LR front even through SLRs have been available for a number of years.
(Australia) As mentioned in the Discussion Document, AUDA has not yet implemented its SLR program. Based on the Final Report of the 2015 Names Policy Panel it is almost certain that auDA will implement some sort of priority SLR mechanism for existing 3LRs… if they elect to proceed with SLRs. The 2017 Policy Review Panel is expected to deliver a proposed implementation policy for SLRs in August 2018.
Recent developments in Australia indicates that the .AU SLR program has been put on hold and may even be scrapped entirely. The Australian government has insisted on urgent reforms within AUDA citing, amongst other reasons, the lack of consultation from AUDA on SLRs.
“On Wednesday, the federal government released 29 recommendations off the back of its investigation into how Australian web addresses will be best managed in the future, after submissions to this inquiry took aim at the consultation and management processes of current domain registry authority auDA.”
“Conflicts had been brewing over the organisation’s consultation process for months, with many small business stakeholders believing auDA’s plan to roll out a top level “.au” web address would cause havoc for businesses and brands that would have to stake a claim on new website domains.”
Furthermore, the suggestion in the Discussion Document that a tender was issued for the provision of an SLR registry system is misleading. The AUDA tender that was issued was for a new registry service provider (RSP), for the entire .AU space, including SLRs and SLDs. In other words .AU SLRs will not have a different registry system to that of the existing SLDs.
(Japan and Malaysia) Even in ccTLDs where SLRs and 3LRs have been available for some time, 3LRs still play a significant part. In .JP (Japan), where SLRs have been available for over 17 years, 3LRs still account for over 30% of the total registrations. .JP had ~200,000 CO.JPs at the introduction of SLRs in 2000, a figure which has risen to over 400,000 currently. In .MY (Malaysia), where SLRs have been available since 2007, 3LRs currently account for 50% of the namespace.
(South Africa) Although South Africa has not yet introduced .ZA SLRs it is worth looking at the registration numbers in the SLDs. This will provide us with a good indication of the growth rates compared to ccTLDs that have introduced SLRs, over the same period. Subsequent to the launch of the new gTLDs in 2012, .ZA has continued to perform well with CO.ZA nett growth amounting to an average of approximately 4,500 registrations per month for the past 2 (two) years. This growth is underpinned by an average of 22,100 new CO.ZA creates per month over the 2-year period. Compare this to the average nett decline for the combined .CAPETOWN, .JOBURG and .DURBAN gTLDs over the same period.
The following charts plot CO.ZA domain count, domain growth and new registrations since 2012.