Over the course of the last week, ICANN has released several pieces of information that taken together begin to allow us to piece together the overall gTLD landscape. ICANN is releasing partial information, without explanation or context, in dribs and drabs, and rumors are flying that we won’t get the “Reveal” (the release of full information on applications and applicants for new gTLDs) until the ICANN meeting in Prague at the end of June.

This partial information and delay from ICANN is creating consternation and confusion among the many applicants and those watching the new gTLD scene. Therefore some analysis, however imperfect, will help. What follows is what we at Minds + Machines consider a likely scenario based on the facts we have, along with some estimates and calculations that we think are reasonable. This is back-of-the-envelope stuff, but it shows how we see the new landscape forming.

First, let’s start with what we know for sure:

• There are 1,268 total applicant profiles in ICANN’s system (the TAS), each of which allows 1-50 applications
• There were 2,091 applications in the TAS as of April 12, 2012
• There were 214 unpaid applications

Given that the application window was suspended with just twelve hours to go, we believe that anyone who hasn’t paid yet isn’t going to, and that people who are paying for their applications aren’t leaving any unpaid. In other words, we don’t think the 214 unpaid applications are ever going to happen, and that there’s a rough one-to-one correspondence between unpaid apps and number of applicants with unpaid apps. Therefore there are roughly 200 applicants who are not actually going to submit 214 apps.

This leaves us with about 1,000 bona fide applicants and 2100 applications, or roughly 2 TLDs for every applicant. But we can break down these numbers a little further based on industry information.

Categorizing 2100 new gTLD applications

We can divide up the applicant groups into five buckets:

1. Geographical applicants, applying for cities, states or regions.
2. Applicants applying for a specific individual project: one or in some cases two names.
3. IDN applications, applying for generic names in a non-Latin script.
4. Portfolio applicants, applying for 20 TLDs or more.
5. Brand applicants, applying for their trademarked names.

There will be exceptions to these generalizations, but overall these categories are clear. But how many of each? The breakdown is important, because it will determine how the competition for high-value generic names will play out. We assume that geos, IDNs and brands are going to face little or no competition because of the favorable ICANN rules in (for geos and brands) or language specificity (IDNs). Therefore the contention is most like to occur among and between the smaller applicants and portfolio applicants who are applying for English-language generic-word TLDs. Looking at each category, we can identify some further characteristics:

Geo Domains

We estimate 50 geographic domains. Some of these have been in the news, but many others have not. This category includes US cities such as Las Vegas and NYC to European cities such as London, Paris, Berlin and Madrid, states such as Bavaria and NRW, and Japanese cities such as Tokyo and Osaka, and of course the entire continent of Africa. Altogether we do not predict more than 50 geo domains. We are estimating one application per applicant (for these purposes, we consider the geographic entity to be the applicant, even if the official “applicant” is another entity with a letter of support).

Niche Applicants

These are individual non-IDN applicants, applying for just one or two generic-word names such as .KIWI, .MUSIC, .HIV and .RADIO. We estimate that these probably represent another 100 applicants and 150 total applications. This is a very rough estimate based on who has been present at ICANN, or .NXT, or other industry meetings. These applicants will compete with one another and often, but not always with portfolio investors for generic terms—as the projects, such as .HIV, may serve a very specific purpose.

IDNs

IDN applicants, in our calculation, represent another 100+ applicants. We know already of applicants for Russian, Chinese and Arabic strings. While we don’t have very good knowledge about these applicants, our cloudy crystal ball says that the number of IDN applications will be around 10% of the total, or 200, and that on average IDN applicants will have 2 applications each. We foresee limited competition between IDNs, and by definition little competition with Latin-characters gTLDs. Note that IDNs could represent brands as well as generic terms and niche products, this is factored into our estimate of 100 applicants, 200 apps.

Portfolio Applicants

Portfolio applicants will be few. With the high cost of applications, there are probably no more than 20 portfolio applicants (such as ourselves) applying for more than 10 TLDs. We will know better when ICANN provides definitive information, but that is our estimate. By estimating 25 apps (on average) per portfolio, we predict a total of 500 applications from portfolio applicants. TLDH is applying for significantly more than the average for its own account; we don’t think that ours is the highest number, but we believe it is in the top five.

Brands

Brands will, with some exceptions, apply just for their trademarks. Based on our previous analysis, anything that isn’t a geo, niche project, IDN or portfolio app is an english language brand app. This gives us a total of 1,200 brand apps, and 730 brands. This implies about 1.65 applications per brand—which is in the general ballpark of the earlier estimate by Fairwinds of 2.2 applications per brand. Verisign has announced 220 applications, the vast majority of which we assume to be brands, many of them from Melbourne IT. We assume that other major registry service providers, particularly Neustar with their loss-leader $10,000 application consulting fee, but also non-U.S. providers like GMO, will have the rest of the brands.

Note that the biggest variable in our analysis is the number of portfolio applications. The number is not likely to be significantly larger than 500, however, as it would reduce the number of applications per brand below credible levels. For example, if there were 1,000 portfolio apps done by the small number of portfolio players, then it would imply 700 brands with 700 apps—not a likely ratio in our opinion. It seems more likely to us that the number of portfolio apps are in the 300-600 range.

Conclusion

The foregoing gives us the following view:

• Geos: 50 applicants – 50 apps (3%)
• Niche applicants: 100 applicants – 150 apps (7%)
• IDNs: 100 applicants – 200 apps (10%)
• Portfolio: 20 applicants: 500 apps (25%)
• Brands: 730 applicants: 1,200 apps (60%)

Total: 1000 applicants, 2100 apps

We therefore see about 700 applications being likely candidates for contention (niche plus portfolio apps). Some strings (.WEB, .MUSIC) are known to have multiple applicants. Others, such as .KIWI, are likely to face no competition. Overall we think that about two thirds (466) of our 700 of these will end up in contention, and of those 466 about half (233) will have more than two contenders.

So, our very rough back-of-the-envelope calculation says that of the 2100 applications, about 233 will have exactly two parties in contention, and these have a much higher likelihood of being settled without an auction. Another 233 will have more than two contenders, which makes a deal more difficult, but by no means impossible. At the end of the day, then, we estimate that ICANN is going to have about 100-125 auctions.

Batching

A word about batching. This is a very difficult area because ICANN hasn’t released much information about how it’s going to happen (which by the way makes any “digital archery” service somewhat suspect). ICANN has said they will pay attention to geographic diversity as well as to “skill” in the digital archery game, but they have also said that applications in contention sets will go in the first batch, which would mean that brands, geos, and IDNs would be sparsely represented in the first batch.

But it is also not at all clear to us that the batches cannot be staggered instead of back-to-back; work on a subsequent batch can begin even before the previous one is finished. Concluding that 2000 apps will take until 2014 is by no means a given if ICANN uses their new-found $350 million windfall wisely. The evaluations should go quickly, particularly since the estimate of how long it will take is based on the fiction that evaluators will be looking at 500 unique applications per batch. They won’t: there are perhaps 15 qualified registry service providers in the world, and therefore ICANN will not be faced with more than that number of unique tech sections—the great majority of the application—in the submissions they receive. Even though ICANN is headquartered at what has become known as “Marina Delay,” it is difficult to imagine that with resources at its disposal, ICANN will not be able evaluate 2000 applications in 12 months.

Summary

Even though information is spotty, there’s enough to start understanding what the post-application landscape will look like. The analysis above is our broad prediction.