When I installed WordPress, I promised myself I won’t write to simply fill space. I wasn’t going to force posts without a worthy trigger, even if I missed writing. The big reveal of Top Level Domains (TLDs) by ICANN was just what I needed.
I took a long, hard look at the applicants and submissions. I then took the whole thing into excel and had some fun with it.
The trigger this time was pretty simple: ICANN is laughing all the way to the bank, and that’s wrong.
The first image in my mind once I did the math was that of Scrooge McDuck diving head first into gold coins (which I had thought was a horrible and painful idea since childhood) and a fat, mustachioed walrus in a top hat and monocle, fanning himself with wads of cash. But after I stopped giggling, I thought seriously about how this should be evaluated, and what will probably happen instead. I have a lot to say, so get comfortable.
Let’s start with some fun numbers.
1,930 overall applications, 1,409 of which were for unique TLDs.
13 applicants went for .APP, the most requested TLD, among them Amazon and Google, but not Apple.
4 proposals had more than 10 solicitations, 36 had more than 5, 230 more than one.
503 individual applicants.
307 applications by start-up Donuts.co (who apparently raised $100M in VC for this exercise. There goes $57M on application fees alone! But there is no bubble, no siree!)
101 applications by Google, 76 by Amazon, 11 by Microsoft, 1 by Apple.
66 were filed as Geographic TLD, 84 as Community.
Upwards of 400 applications for brand names (I’ve counted over 375 well known brands, but with non-Latin character application and smaller brands I am not familiar with, I’ll give this number some wiggle room).
$357M to ICANN for application fees alone.
-$20K is the loss ICANN projected in their budgets for 500 applications (if that means they’re expecting a ~$100K loss, there needs to be some serious inquiry into this not-for-profit’s inefficiencies).
Between the existing generic and country-specific TLDs there are currently about 300 Top Level Domains. If only half of the new applications are approved, we’ll have a thousand “.something” in play. Is this a good thing? Is it even needed?
Personally, I think the answer is no.
Overall, the expansion of generic TLDs did not work as planned, and as a business model, I would argue the experiment had, so far, failed. The gTLDs approved over the past decade have not gotten the traction hoped, despite a legitimate business case that can be made on their behalf. The true “innovation”–if you can call it that–with TLDs has been domain hacking, or the use of unconventional domains to creatively spell words or create associations. Most of the domain hacking has been accomplished using country-specific TLDs such as .ly (Libya), .tv (Tuvalu), .it (Italy), .is (Iceland), .me (Montenegro) and others.
But for each successful hack, there’s a gTLD that isn’t nearly as successful with a legitimate purpose. Only one .info site is accessed somewhat regularly by me (mta.info). The trusted .com continues to be preferred over .biz, unless you’re in Turkey and you use it as a hack (Biz means “we”). The .museum registry managed a commendable global collection of registrants, but with the majority continue using their existing domains, this appears to be more about protecting brand than an earnest exploitation of the new gTLD’s potential. The MoMA and the Guggenheim among others use .org, the Louvre and Musée d’Orsay celebrate their “Frenchness” by using .fr and the Getty goes for a .edu. Other notables like the Uffizi gallery, Tate Modern or the State Hermitage did not apply. .mobi was a misguided attempt from the start, and with responsive web design slowly becoming standard, I hope it is retired. Even .xxx is facing opposition from the adult entertainment industry who do not wish to be segregated and potentially overly regulated, and it is unlikely to be adopted as the industry standard.
I won’t even mention .coop, .pro, .int, .aero and others. I don’t doubt there are valuable sites out there using the expanded gTLDs, but not enough, to me, to justify the business case. The internet still predominantly uses the original gTLDs, and the viable alternatives since have seen their success born of unintended usages.
So how can anybody approach an approval process for such a program? While I do not have business models for any applicant, below is my proposed guidelines, based on the evaluation criteria set in ICANN’s Applicant Guidebook, and why I believe it will eventually be ignored in favor of money. Yes, I’m a cynic. (*Note, I will not be touching on non-Latin characters applications, because I can’t decipher or know enough about them).
The first evaluation criterion is similarity to existing TLDs. With no exact duplicates, no applicant threw away $185K. Other similarities are very limited, so I doubt any will fail here. There is a possibility, however, of a stricter interpretation, which may reject proposals such as .eus and .frl for being too confusing with the existing .eu and .fr; or Google’s .gbiz and the multiple .mobile applications for similarity with existing .biz and .mobi.
The second criterion addresses confusion with other applied-for TLDs, and here there is a potential for rejections. There are three types of “confusions”, in my opinion: singular vs. plural applications, string synonyms, and base vs. compound words.
When singular and plural forms of the same word have both been applied for, just one should be approved. Examples include: .coupon/s, .deal/s and .hotel/s. While this will be a case-by-case decision, my gut says the singular will be awarded in most cases.
With competing synonyms, only one should be approved as well. Notable examples which will likely be approved include .film vs. .movie, .auto vs. .car, .audio vs. .music vs. .song, and the mega-fight of the century between .site, .web, .webs and .website.
For base vs. compound words, approving the base word only will make it is easier and faster for users. The highest profile example is .tech vs. .technology, but there are others, including .now vs. .nowtv.
Unfortunately, that’s the evaluation criteria extent set by ICANN’s Guidebook (a third one addresses similarity with Internationalized Domain Names, and is out of this post’s purview). With the above the only criteria for rejection, my estimate returns over 1,000 likely approvals. I would not be surprised to see all of them approved as each new TLD carries a yearly maintenance fee of at least $25K, which means a guaranteed $25M in ICANN’s coffers. And that, sadly, is reason enough to approve them all. Moreover, for IP protection purposes, brands are likely to purchase additional secondary domains, especially if a gTLD like .suck is approved. That means additional money lubricating the wheels of the system, but no clear benefits for users.
There are several additional criteria I would like to see incorporated into the evaluation process to truly foster the user innovation they desire. My extended evaluation criteria cover: scope and usage potential, overarching TLD user experience, trust through regulation and transparency, streamlining to reduce user confusion, embracing domain hacking, and guidelines for treating branded applications. There are natural overlaps here, so the following will not necessarily map 1:1 to these topics.
Many gTLDs’ potential seem to be exhausted with generic directory-like applications. Some have a small possible universe: .nba might be a great solution for each team, but there are only thirty of them. It is a big enough brand to potentially merit it, but maybe the generic .basketball should be preferred. Others have an infinitely larger universe, but to me lack value unless handled by a single company with a unified experience throughout: .forsale has been requested by two applicants and could compete nicely against a Craigslist for example, but unless one body orchestrates the experiences throughout all secondary domains, I don’t see it succeeding. .bbb (Better Business Bureau) or .ieee (Institute of Electrical and Electronics Engineers) are similar propositions. Extended directories and reviews can only benefit consumers and facilitate information gathering, but unless they control the entirety of the experience, there is no added value to a simple search on their current domain.
My take: I would like to see priority given to non-branded, larger-scoped TLDs. I think there can be benefits for both single-source and decentralized secondary domains, to be evaluated on a case-by-case basis.
In theory, I like the idea of approving more semantic directory-TLDs like .author (from Amazon) or .cpa which could be used as regulated personal/professional pages, but where is the line drawn? What occupations get a TLD?
A similar argument can be made for .pharmacy, what business is worthy of a TLDs? A unique feature for .pharmacy is creating trust through transparency and regulation. If consumers know the application process for a .pharmacy domain, and understand that only verified pharmacies can secure one, e-commerce potential may increase as a result of established authority and authenticity.
My take: Despite liking the idea of semantic TLDs, I would concentrate exclusively on regulated TLDs. Ensuring quality sites through registration requirements is valuable. .cpa applicants will be required to prove credentials to secure a domain, as will .pharmacy, and as a result consumers will be able to rest assured they are transacting with trusted entities.
Confusion with other submissions was covered as an official criterion, but I think it is prudent to go beyond ICANN’s definition. Similarity is defined as the probability of user confusion: so how should .education, .college, .university, and the handful of universities applying for gTLDs (.mit, .latrobe and .monash, .rmit) be evaluated when .edu exists? Similarly, how are .army, .airforce and .navy evaluated with .mil in use, and with most branches using a .com for their user-facing official sites? Other examples include potential confusion between .phone and .tel or .sex, .porn and .adult with .xxx.
A similar issue arises when considering word sets. Is it necessary to offer .corp, .inc, .llc, .llp, .gmbh, .sarl for businesses when .com has served them well, and .biz has not caught on? What of .ngo with .org already extensively used?
My take: I would reject all. Simple.
Another type of confusion is exemplified with .dwg, .mov: the formats for AutoCAD’s drawings and the popular movie extension. Confusion could also translate to servers interpreting the page: should it initiate a file download or render a site?
My take: Deny all applications matching file extensions or potentially triggering server conflicts.
The best source of TLD innovation came from domain hacking, embrace it. There is an application for .ing by Google, and given the complexity of securing .ng (Nigeria) domains, the most active verb has yet to be hacked, .ing can change that completely. Another application is by the city of Istanbul for .ist and that could be fun.
My take: More of these. Please. The fine folks making sites and start ups that advance humanity and technology (or just provide yet another way to share a photo on Facebook) have said their piece and approve of domain hacking. I’m disappointed there were not more three-letter nouns or propositions. I can make a business case for a .out TLD, which could be appropriate for outdoorsy or active pursuits, for reviews, for fun hacks and even for the gay community. And can there please be a new country with a .up country code?
Lastly, brands. Lastly, because they are obviously the thorniest. New operators stand to make a tidy sum from hundreds of brands looking to protect their IP and prevent anybody unauthorized from saying bad things. This is likely to happen on TLDs that are both generic- and targeted-enough. Looking at the applications, there are hundreds clear cut brands, product names, brand extensions, and service marks and taglines. Is there really a need for branded TLDs? What are the determining factors in approving one?
My take: Bad idea. This might be the most controversial statement in this post, but I think brands are not TLD material. If they don’t exist, neither company nor user will be any poorer for it, and creating them will not necessarily add value, might end up hurting innovation, or create orphaned TLDs that will be decommissioned like microsites currently are.
Often, brands are rather common, generic words. Among the applications there are a good number which could catch on with consumers, but not if they’re held by a company for brand use only. Examples include .off, .live, .open, .mini, .express, .office. If these are not available to users I see branded applications as too narrow to be useful. An interesting application by Google for .plus also falls under this category. I guess the intention will be for each Google+ user to become a secondary domain, securing a foothold in personal pages, fully trackable and controlled, as gateway to the social web Google struggle to make a serious play in. I would still rather see .plus as truly a generic, non-Google affiliated TLD, if approved.
Service marks and taglines, for their part, should be denied categorically. I spotted .onyourside (Nationwide), .afamilycompany and .rightathome (SC Johnson). These are too focused on potentially transient campaigns, are long and awkward, and add no value to users.
I can see massive corporations with many internal systems, intranets and protected employee systems getting branded TLDs for internal purposes. There’s value for IBM employees in timesheets.ibm or directory.ibm. For users, though, ibm.com just makes more sense than website.ibm. In a way, this is the inverse to single-source directory-TLDs. For them to make sense, directory-TLDs need one company orchestrating a unified experience across all secondary domains; here, standardization of secondary domains across branded-TLDs are needed for users to count on always having say, shop.ibm available. An easier, more intuitive implementation will still be for a .shop to be approved and each brand can house their e-commerce efforts in a ibm.shop
Even with this extended evaluation criteria, there are still potentially hundreds of approved TLDs, but at least it will be a bit more manageable. Phew, that’s all, I think.
Oh, one more question, am I really missing something here? Can there really be 4 separate individuals that thought up a business model for .pizza? Does anybody know what?