Preliminary remarks on the assessment of registrar network performance

June 2015

The introduction of new TLDs (Top-Level Domains) on the market in 2014 brought relations between registries and registrars to the forefront. If some such as Michele Neylon (Blacknight) stressed that the balance of power would be reversed in favor of registrars, the possibility for registries to create their own registrars may prove that prognosis to be incorrect. For all the newcomers, effective management of their distribution networks, i.e. on which their ability to access the market depends, is a strategic issue of vital importance.


The study of the quantitative data at our disposal is nevertheless a source of astonishment that we should like to share with the community. Without claiming to provide definitive answers to the issues they raise, these findings are nonetheless an initial step toward a better understanding of the domain name market, backed by the analysis and development of the most relevant indicators to help stakeholders build their strategies and assess the results.

An increasing number of registrars active on nTLDs

We make a distinction between three indicators to study the registrars for a given TLD:

  •  a) registrars referenced in ICANN reports, i.e. which are liable to market nTLDs;
  • b) operational registrars with at least one name in their portfolio, and which have therefore proved their technical ability to process orders from their customers;
  • c) active registrars, i.e. that have filed at least one domain name in the period in question. The threshold of 1 name filed is very low, but avoids debate about the most relevant limit for deciding whether a registrar is "active" or not.

 Evolution of registrars operating nTLDs Fig. 1 – Change in registrars operating on nTLDs
Source: ICANN / Registry Analytics by Afnic

Fig. 1 shows that the number of active operational registrars increased from 50 in January 2014 to 200 in February 2015, multiplication by a factor of 4, while the number of referenced registrars was "only" multiplied by 2 in the same period. This corresponds to the period of implementation of the nTLD market. Overall, the proportion of active registrars compared with operational registrars is greater than 90%. Surveys carried out on various nTLDs, however, show that the variations are greater in specific cases. For example, for the .xyz, only 58% of the operational registrars were "active" in February 2015 (the date of the last ICANN report available).

The ratio of operational registrars against referenced registrars illustrates the motivation of the former to start marketing nTLDs. The ratio of active registrars compared with operational registrars measures the level of involvement of registrars with the technical capacity to market nTLDs to their customers. It also reflects the extent of demand. A low or declining ratio may suggest that the registry does not sufficiently drive its registrar network, or that the latter, once they are capable of responding to any request, do not have the success they hoped for with customers. This is unfortunate for both the registry, which may not have the distribution channels it needs to reach its targets, and for the registrars, who have invested in implementing nTLDs in their systems for nothing.

"Leaders" and role models

The study of the market share of the first five registrars in the portfolio of the top 10 nTLDs (by volume) leads to some surprising findings. Table 1 shows the proportion in February 2015 and in June 2015.


Table 1 - Market share of the top 5 registrars in the top 10 nTLDs Table 1 - Market share of the top 5 registrars in the top 10 nTLDs
(according to the ranking by in June 2015)
Source: ICANN / Registry Analytics by Afnic (02/2015); (06/2015)

We can see that the average for the top 10 nTLDs is 90%, more than twice the average of all the nTLDs. The concentration of these portfolios is therefore extreme, making the .club an outsider with 66%. But this proportion of "leaders" for each nTLD can be seen as a reflection of the adoption of the nTLD by the community of registrars, or the "craving" they have to market it. In some cases such as XN--SES554G (.网址) or the .realtor nTLD, the level of 100% can be explained by the fact that the registry is its own registrar. This is also the case for many .corp nTLDs.

The .xyz, .science, .party, .top and .link and perhaps .wang belong to the category of nTLDs "pushed" by a handful of registrars conducting mass filings as part of highly aggressive strategies.

The case of .berlin is more complex because it is a geo-TLD. Its target being located geographically, its "natural" distribution network mainly consists of German registrars and very large international registrars.

Finally, the .club remains significantly more concentrated than the average but at a level significantly lower than those of its peers. But we know that it is also that of an nTLD that has been less a "victim" – for the time being – of free or semi-free marketing strategies by registrars. Its ratio therefore shows a higher level of adoption by registrars and ultimately, it may be hoped for its sake, the greater efficiency of a network more securely meshed than those of the "mushroom nTLDs," by reference to those towns in the Far West that rose out of the ground in a few days before sinking back into oblivion, their inhabitants abandoning them as quickly as when they arrived.

Measurement of the "normal" concentration level

The main problem when constructing a new indicator is to situate the levels enabling its interpretation. From what values is it reassuring, worrying, or simply neutral? With little hindsight on nTLDs which have little past, the average market share of the 5 leaders on all of the nTLDs (see Fig. 2) provides a landmark even if it is biased by the mass filing strategies outlined above. If its overall trend is downward, that is a good sign for nTLDs, because it shows that they are attracting a growing number of active registrars.

  Figure 2 - Evolution of nTLDs market shares for the 5 first registrars Fig. 2 – Change in market share of the top 5 registrars in nTLDs
Source: ICANN / Registry Analytics by Afnic


An nTLD with a concentration rate higher than average may be seen as a good sign, because it being propelled by some particularly active stakeholders. But this coin has another side: is it not better to be in the situation of an nTLD that has persuaded a large number of registrars to get really involved in its development, therefore depending much less on the "whims" or reversals in strategy of its "leaders"?

Lessons to be learnt

The main impression one gets from this study is that it will be difficult to create a single indicator to measure the performance of nTLD distribution strategies (and Internet Top Level Domains in general). Performance must be measured against targets, but they are not the same depending on whether the TLD is a "pure generic" with global reach or an nTLD targeting a specific niche market.

It is certain that a low proportion of operational registrars registering names indicates a problem of leadership or a mismatch between network and target (and therefore a risk of under-performance). Similarly, too great an offset between referenced registrars and operational registrars reflects an abnormal "gap" between registrars’ intentions and their move to the actual act.

Depending on the profile of the nTLD, it will be logical to see it work with a small but motivated network of registrars, or with a larger network while remaining focused for example on a continent or country. Reference to overall averages is therefore necessary but insufficient to give an exploitable picture of performance in terms of strategic actions.

The indicators presented in previous editions of our Domain Name Industry Report to measure TLD performance (DN creation rate and DN retention rate) can also be very useful in quantifying the performance of individual registrars and controlling the distribution strategy, taking into account the bias linked to incoming and outgoing transfers.






Lire cette ressource en français Top of the page