Wanted: a gTLD to ban

ICANN may have failed so far to deliver a way for the world to create any more gTLDs, but it’s about to pick a string that it will resolve to never, ever delegate.

It’s going to designate an official “private use” string, designed for organizations to use behind their own firewalls, and promise that the chosen string will never make it to the DNS root.

IP lawyers and new gTLD consultants might want to keep an eye on this one.

The move comes at the prompting of the Security and Stability Advisory Committee, which called for ICANN to pick a private-use TLD in a September 2020 document (pdf).

ICANN hasn’t picked a string yet, but it has published its criteria for public comment:

1. It is a valid DNS label.
2. It is not already delegated in the root zone.
3. It is not confusingly similar to another TLD in existence.
4. It is relatively short, memorable, and meaningful.

The obvious thing to do would be to pick one of the 42 strings ICANN banned in the 2012 new gTLD round, which includes .example, .test and .invalid, or one of the three strings it subsequently decided were too risky to go in the root due to their extensive use on private networks — .corp, .mail and .home.

The SSAC notes in its document that ICANN’s two root server constellations receive about 854 million requests a day for .home — the most-used invalid TLD — presumably due to leaks from corporate networks and home routers.

But .homes (plural) is currently in use — XYZ.com manages the registry — so would .home fail the “confusingly similar” test? Given that it’s already established ICANN policy that plurals should be banned in the next round, .home could be ruled out.

ICANN’s consultation doesn’t make mention of whether gTLDs applied for in subsequent rounds would be tested for confusing similarity against this currently theoretical private-use string, but it seems likely.

Anyone considering applying for a gTLD in future will want to make sure the string ICANN picks isn’t too close to their brands or gTLD string ideas. Its eventual choice of string will also be open for public comment.

There don’t seem to be a massive amount of real-world benefits to designating a single private-use TLD string.

Nobody would be obliged to use it in their kit or on their networks, even if they know it exists, and ICANN’s track record of reaching out to the broader tech sector isn’t exactly stellar (see: universal acceptance). And even if everyone currently using a different TLD in their products were to switch to ICANN’s choice, it would presumably take many years for currently deployed gear to cycle out of usage.

Interview: Sandeep Ramchandani on 10 years of Radix and new gTLDs

It’s over a decade since ICANN’s last new gTLD application round, and naturally enough many companies in the industry are celebrating their 10th anniversaries too. Radix has been putting a lot of effort into promoting its own birthday, so a couple months ago I had a long chat with CEO Sandeep Ramchandani about the last decade and what the future holds.

We discussed Radix’s business model, rivalries, performance, blockchain-based alt-root gTLDs, the company’s plans for the next application round, and the TLDs he wishes the company had bought.

Measuring success

Radix is based in Dubai but has most of its 75-person headcount located in Mumbai, India. It also has satellites, mainly focused on registrar relations and marketing, in the US, South America (where it markets .uno) and Asia.

Across 10 gTLDs, it has amassed over 5.6 million registrations, according to its web site. If you exclude pre-2012 TLD .info, that’s more than Identity Digital, which has more than 20 times as many TLDs in its stable.

“Donuts went for the long tail, category-specific names,” Ramchandani said. “Our idea was to launch TLDs that had mass-market potential.”

More than half of the regs to date have been concentrated in two TLDs — .online and .site, each of which measure their volumes in seven figures. The TLD .store is approaching a million names also.

More than half of the company’s sales are coming from the US, with 20% to 30% from Europe. It’s pretty much the same mix across premium sales and basic regs, he said.

Radix has been focusing most of its marketing effort on .store, .tech and .online, but Ramchandani says he thinks .site, currently at around 1.2 million domains and the company’s second-biggest seller, has a lot of untapped potential.

“We have about six million domains right now, but I don’t think that’s the best metric, as you can easily spike volumes by selling cheap,” Ramchandani said.

“The real metric is domains that are renewing every year,” he said. “Our first year registration price is still fairly low, but we optimize it to maximize our renewals.”

There’s also the matter of live web sites, of course. Radix estimates there are over 725,000 live sites on its domains, according to its web site.

On premium renewals

If you’re a domain investor, imagine you have a portfolio of tens of thousands of domains that you price at between $100 and $10,000, and you get to sell them not once, but every single year.

That’s Radix’s “high-high” business model, where domains in premium tiers are priced for users and renew at premium prices.

Ramchandani says that between 10% and 15% of Radix’s revenue comes from premiums, but it’s growing faster than regular-price regs. So far, it’s sold about 5% to 6% of its premium inventory. Many thousands of domains remain.

But the problem with premiums is of course whether or not they will renew at all, particularly if they’ve been sold to a domain investor who failed to secure the quick flip.

Ramchandani said premium renewals have been running at about 55% for the first renew, 75% for the second and above 90% for the third. The second and third-time figures are very respectable indeed for any TLD.

Premiums are typically held by end-user registrants rather than investors, he said. Probably lower the one in 10 premiums are owned by domainers, he guessed.

“We don’t have a lot of domainer interest, because the holding cost is too high,” he said. “A lot of the best web sites we see on our TLDs are on premiums.”

On industry consolidation

One of Ramchandani’s regrets over that last decade is that Radix didn’t manage to pick up some of the gTLDs that changed hands as the industry began to consolidate.

“We could have gone a bit harder to acquire some of the larger TLDs that did sell over the last few years,” he said. He would have to loved to have gobbled up .club or .design, he said, but these were bought by deeper-pocketed GoDaddy.

He said Radix sees itself as a buyer rather than a seller “for sure”, but the problem is: “We are interested in buying, but there aren’t so many out there that are really good TLDs.”

The company is not interested in the business model of buying up a dormant dot-brand and repurposing it to mean something other than its original meaning, which other registries have tried.

Ironically, that was where Radix started out, selling Palau’s .pw ccTLD as a domain for the “professional web”, which was a hard sell.

On the next round and alt-root TLDs

The long-touted next application round has been in policy development hell at ICANN for a decade, and Ramchandani agrees that “it’s a couple years away at this point and could very well be longer than that”.

“We will participate,” he confirms, adding “we’ll have to look at which TLDs we think are worth going for.”

“I think the best ones are already on the market, but there may be a few — based on recent trends — that make really good TLDs that qualify to have the scale and global impact that we look for,” he says.

“But honestly if we end up with none I think we still think have a very, very exciting business opportunity ahead of us for the next 10 years at least with the TLDs we already have, so it’s not something we’re betting the business on,” he says.

But how big will the next round be? There were 1,930 applications in the 2012 round, and plenty of anecdotal evidence today about pent-up demand, particularly from brands. That said, many say the first round wasn’t as successful as some had anticipated, which could lower turnout.

“A lot depends on the barrier to entry,” Ramchandani says. “Last time there was an investment of $185,000 for an application so there was a decent barrier to entry, but there are talks about potentially reducing that spectacularly. If that happens, I think the floodgates will open.”

(I should note that our conversation took place before ICANN announced that applications fees will likely be closer to $250,000 in the next round.)

“Last time this process ran there was less confidence that there was a sustainable business around new gTLDS, but given how some of the domainers in that round have performed — there are a bunch of TLDs that have done substantially better than everyone’s expectations — there might a lot more coming in to fight for those in contention with us in the next round,” he said.

He’s expecting to see “really high numbers” in dollar terms when strings come up for auction, but “a dozen, max, that will be really highly contested”.

One factor that could push down applications are blockchain-based alt-roots, where the likes of Unstoppable Domains throwing its legal weight around to prevent versions its TLDs appearing in other roots.

That said, Ramchandani would not rule out applying for TLDs that exist in alt-roots.

New gTLDs grow in China as .cn regs slide

China-based registrations of .cn domains decreased in the first half of last year, while new gTLD swelled to pick up the slack, according to the local registry’s semi-annual report.

CNNIC published the English translation of its first-half 2022 statistical report in December, showing a steep decline in .cn regs, from 20,410,139 at the end of 2021 to 17,861,269 at the end of June last year.

These appear to be registrations made by registrants based in China. Verisign’s Domain Name Industry Brief for Q2 2022 shows .cn at 20.6 million.

While .cn slumped, new gTLDs saw an uptick of almost a million names in China, from 3,615,751 domains to 4,590,705 over the six months. New gTLDs accounted for 13.6% of all China-registered domains, the CNNIC report says.

The report also shows that the number of Chinese-registered .com names dropped by about half a million, to 10,093,729 from 10,649,851, over the period.

The full report can be viewed here (pdf).

Identity Digital sees abuse up a bit in Q3

Identity Digital has published its second quarterly abuse review, showing abuse reports up slightly overall.

The report, which covers the third quarter 2022, also shows that the registry only released the private Whois information for a single domain during the period.

ID said it closed 3,225 abuse cases in Q3, up from 3,007 in Q2, covering 4,615 domains, up from 3,816. The vast majority — almost 93% — related to phishing. That’s in line with the previous quarter.

In about 1,500 cases, the domains in question where suspended by the registry or registrar in the first 24 hours, the report says. In 630 cases, the registry took action after the registrar failed to act within 72 hours.

The company received five complaints about child sexual abuse material from the Internet Watch Foundation during the period, up a couple on Q2, but all were remediated by the registrars in question.

It received four takedown notices from the Motion Picture Association under the registry’s Trusted Notifier Program, all of which resulted in suspended domains.

There were requests for private Whois information for 20 domains, three of which were intellectual property related, but only one resulted in disclosure. In 12 cases ID took the decision not to disclose.

The company has over 260 gTLDs in its stable and over 5.5 million registered domains.

The full slide deck can be viewed here (pdf).

.music gets its first live web site

The .music gTLD may be still officially unlaunched, but it got its first live anchor tenant this week after the DotMusic registry joined a partnership aimed at making translated lyrics more accessible.

DotMusic said it has become part of an initiative called BELEM, for “Boosting European Lyrics and their Entrepreneurial Monetisation”. Given that the entire namespace of .music is currently available, one wonders why such a contrived acronym was chosen.

The project, which is funded with €2 million of European Union money, has a live web site managed by DotMusic at belem.music. It’s the first .music site to go live other than the mandatory nic.music registry hub.

BELEM is out to get lyrics in various European languages translated by humans and the translations licensed to streaming services.

It has 14 other partners, including Canadian lyrics licensing company LyricFind and French streaming service Deezer, which plans to roll out one-click translations based on the new service.

The aim, the group says, is to “break down cultural barriers and further support artists’ monetisation of their works”.

For DotMusic, it’s an anchor tenant perhaps more noticeable to the music industry than to the public at large.

The .music gTLD has been live in the root for over two years now, and there’s still no published launch plan.

Content police? ICANN mulls bylaws change

ICANN could change its bylaws to allow it to police internet content to an extent, it emerged this week with the publication of the Operational Design Assessment for the next stage of the new gTLD program.

Currently, ICANN’s bylaws state that the Org may not “regulate (i.e., impose rules and restrictions on) services that use the Internet’s unique identifiers or the content that such services carry or provide”, and it’s been adamant that it is not the “content police”.

But the community has recommended that future new gTLD applicants should be able to agree to so-called Registry Voluntary Commitments, statements of registry policy that ICANN would be able to enforce via contract.

RVCs would be much like the Public Interest Commitments many registries agree to in the 2012 application round, implemented before ICANN’s current bylaws were in effect.

As an example I’ve used before, Vox Populi Registry has PICs that ban cyberbullying and porn in its .sucks gTLD, and in theory could lose its contract if it breaks that rule by allowing .sucks sites to host porn (like this NSFW one, for example).

ICANN’s board of directors expressed concern two years ago that its bylaws may prevent it from approving the RVC recommendation.

But Org staff have now raised, in writing and on a webinar today, the prospect that the board could change the bylaws to permit RVCs to go ahead. The ODA published on Monday states:

The Board may wish to consider how and whether it can accept the recommendations related to PICs and RVCs. One option may be to amend the Bylaws with a narrowly tailored amendment to ensure that there are no ambiguities around ICANN’s ability to agree to and enforce PICs and RVCs as envisioned

How worrying this could be would depend on the wording, of course, but even the chance of ICANN meddling in content is usually enough to raise eyebrows at the likes of the Electronic Frontier Foundation, not to mention supporters of blockchain alt-roots, many of whom seem to think ICANN is already censoring the internet.

It’s not clear whether the change is something the board is actively considering, or just an idea being floated by staff.

ICANN bloat to continue as new gTLD program begins

ICANN expects to hire so many new staffers over the next few years that it’ll need to rent a second office in Los Angeles to store them all in, according to a newly published new gTLD program planning document.

We’re looking at about 100 more people on the payroll, about 25% above the current level, judging by ICANN figures.

The Org said in the Operational Design Assessment published last night that the next new gTLD round will need it to hire another 25 to 30 dedicated staff during implementation of the program, along with 10 to 15 contractors, and then an additional 50 to 60 permanent staff to help manage the program going forward.

The number could be even higher if the board of directors and community encourage ICANN to speed up the roll-out of the round by reducing automation and relying more on the manual processing of applications.

The ODA says that ICANN has already identified an option to lease more office space close to its LA headquarters, to house the newcomers.

The budget for ICANN’s current fiscal year expects the Org to average 423 operational staff and another 25 employees dedicated to the new gTLD program.

ICANN reckons that the next round will require 125 full-time equivalents (FTE) during the implementation phase, reduced to 114 after the application phase kicks off.

For comparison, in May 2012, shortly after ICANN closed the application window for the last round, the whole organization comprised just 143 people. A year later, it had grown to 239.

The ODA does not break down how many additional staffers it will need to hire if the community plumps for the low-automation “Option 2”.

ICANN spunks a year, $9 million, on new gTLD plans destined for trashcan

ICANN has published the Operational Design Assessment for the next round of the new gTLD program, a weighty tome of 400 pages, most of which are likely destined to be torn up, burned, or used as toilet paper.

The ODA is the document, prepared by staff for board consideration, that lays out how the Org could implement the community’s policy recommendations for the next application round, how much it would cost, and how long it would take.

As I wrote last week, the paper outlines two options, the more expensive of which would take five years and cost $125 million before a single application fee is collected.

This option “reflects the goal of delivering on all outputs of the SubPro Final Report [the community’s 300-odd policy recommendations] to the maximum extent possible”.

This would see the clock ticking the moment ICANN gets the board’s nod and begins the implementation work — best case scenario, probably the first half of next year — and the first applications accepted at least five years later.

So, no new gTLD applications would be received until the first half of 2028 at the earliest. The first registry go-live would not happen until the 2030s, three decades after the first application window closed.

The second option, which was discussed on a webinar last week, would take about 18 months to roll out and cost half as much in up-front costs, but would not necessarily give the community every last thing it has asked for.

In this scenario, the next application window could open as early as 2025, followed by windows in 2026, 2027 and 2028. There’d be no per-window limit on applications, but ICANN would only start to process 450 each year, with the lucky applications selected by lottery.

What’s surprising about the ODA is how little airtime is given to the second option — known as the “cyclical” or “batching” option — which doesn’t really get a serious look-in until page 354.

The large majority of the document is devoted to the single-round, long-runway, more-expensive option, which Org surely knows will prove repellent to most community members and would, if approved, surely confirm that ICANN is mortally unfit for purpose.

Yet ICANN has nevertheless spunked over a year and $9 million of domain buyers’ money assessing an operational design it surely knows has no chance of ever going operational. It’s pure, maddening, bureaucratic wheel-spinning.

ICANN will hold two webinars tomorrow to discuss the document, so if you’re interested in the debate, best settle in for a night of tedious and rather frustrating reading.

The ODA itself is here (pdf).

New gTLD applications to cost about $250,000

Getting hold of a new gTLD could cost applicants well north of a quarter million dollars in base application fees alone in the next round, according to ICANN.

Presenting the results of its year-long Operational Design Phase to the GNSO Council via Zoom last night, staffers said application fees are likely to be either around $240,600 or $270,000 next time, higher than the $185,000 it charged in 2012.

Those would be the base fees, not including any additional evaluations or contention-related fees.

The Org next week is set to present its board and the community with a stark choice — one big expensive round along the lines of 2012, with a potential five-year wait for the next application window to open, or a cheaper, staggered four-stage round with maybe only 18 months of development time.

The Operational Design Assessment — a 400-page tome the Org has spent the last 14 months developing — is set to be published early next week, outlining two options for how ICANN should proceed on the next round.

One option is to build a highly automated system that fully implements all of the GNSO’s policy recommendations but costs up to $125 million up-front to build and roll out over five years. Application fees would be about $270,000.

The other would cut some bells and whistles and require more human intervention, but would be cheaper at up to $67 million up-front and could be rolled out within 18 months. Application fees would be about $240,600.

ICANN CFO Xavier Calvez, responding to exclamations of surprise via Zoom chat, said that a decade of inflation alone would lead to a 28% price increase to $237,000 if the next round were opened today, but in two or three years the price could be even higher if current economic trends continue.

While many expected the fact that technical evaluations will be conducted on a registry service provider basis rather than a per-application basis would wipe tens of thousands from the application fee, ICANN pointed out that building and executing this RSP pre-evaluation process will also cost it money.

ICANN wants to operate the program on a “cost-recovery basis”, so it neither makes a profit nor has to dig into its operational budget. It expects “more than three dozen vendors will be required” to help run the round.

It seems that the portion of the fee set aside to deal with “risks” — basically, anticipated litigation — is expected to be around a fifth of the total, compared to about a third in the 2012 round.

ICANN is asking its board and the community to decide between what it calls “Option 1 — One Big Round” and “Option 2 — Four Annual Cycles”.

Option 1 would essentially be a replay of 2012, where there’s a single unlimited application window, maybe a couple thousand applications, and then ICANN processes them all in a highly automated fashion using custom-built software.

Option 2 would allow unlimited applications once a year for four years, but it would cap the number processed per year at 450 and there’d be a greater degree of manual processing, which ICANN, apparently unfamiliar with its own history of software development, thinks poses additional risk.

My hot take is that the Org is presenting a false choice here, much like it did in January with its ODA on Whois reform, where one option was so unpalatably time-consuming and expensive that it had most of the community retching into their soy-based lattes.

There’s also an implicit criticism in both ODAs that the community-driven policy-making process has a tendency to make big asks without adequately considering the resources required to actually get them done.

I might be wrong, but I can’t at this early stage see much support emerging for the “One Big Round” option, except perhaps from the most ardent opponents of the new gTLD program.

ICANN expects to deliver the ODA — 100 pages with 300 pages of appendices — to its board on Monday, with wider publication not long after that. It will hold two webinars for the community to discuss the document on Wednesday.

Macy’s scraps .macys gTLD

US retailer Macy’s has dumped its dot-brand gTLD .macys.

The company told ICANN recently that it no longer wishes to hold a registry contract, noting that it never used the gTLD.

ICANN last week agreed that as a dot-brand with no third-party users, the domain will not be redelegated to another registry.

It’s the seventh gTLD to scrap its contract this year, lower than ICANN’s budget estimates.