Minds + Machines

A PDF version of this survey and analysis, with full data tables, is available as a downloadable PDF here.

Summary: A survey of the domain registration behavior of Fortune 100 companies reveals that they have not registered many of their trademarks in recently created generic top-level domains (gTLDs). A sample of 1043 brands were registered in less than 30% of the eight new open gTLDs created after 2001. If historical registration data is a guide, brands are unlikely to undertake many defensive domain name registrations in the proposed new gTLDs, and furthermore are unlikely to be the victims of cybersquatting.

With the help of DomainTools, we surveyed 1043 brands owned by the Fortune 100.

Our purpose was to discover to what extent large companies – which have been the loudest critics of ICANN’s new gTLD program – have actually registered their brands defensively in the already existing new gTLDs.

By counting which domain names had been registered, reserved, or otherwise made unavailable for new registration, we are able to see which gTLDs are either registered in defensively (by brands) or cybersquatted by malefactors. Where many names are registered (e.g., .com), we can deduce whether defensive registrations and/or cybersquatting is prevalent.

The data shows that brand names are registered as domain names 87% of the time in .com, .net, and .org; just 67% of the time in .info and .biz; and only 29% in .mobi, .asia, .cat, .jobs, .name, .pro, .tel and .travel – the newer ICANN-created gTLDs where it is easy to register a domain name, and may be attractive to cybersquatters.

Overall, the claims of brand owners that they will be forced to spend significant amounts of money performing defensive registrations in the proposed new gTLDs are not supported by the historical data, which shows that they largely do not undertake defensive registrations in new gTLDs, nor is there any extensive cybersquatting in new gTLDs.

Will new gTLDs force brand owners to register names defensively?

In a previous study, we found that the cost of enforcement of trademark rights in new gTLDs is likely to be small – on the order of $.10 per registered trademark, per year. The current study looks at the likely cost of defensive registrations to trademark holders.

A vocal group of brand owners has repeatedly claimed that the cost of defensive registrations would be “astronomical” and a “major burden on U.S. businesses.”

The following correspondence to ICANN provides a flavor of these claims:

• “…astronomical registration fees required to protect trademarks across new gTLDs.” (Overstock.com)
• “The new gTLD program raises substantial concern about costs to businesses, such as a need to file many defensive registrations, making this program like a nuisance tax which many companies cannot afford.” (Time-Warner)
• “[EOIs] would seem to be a premature invitation for mischief that will end up costing trademark owners big bucks one way or another.” (Cowan, Liebowitz, and Latman)
• “A more likely scenario would be for a business just to register site Web addresses [sic] pairing their brand name with any new extensions, such as fios.telephone or gillette.razor. But even that defense could cost marketers up to $1.5 billion, estimates the not-for-profit Coalition Against Domain Name Abuse.”
• “The scale on which owners of famous marks such as Microsoft will have to secure defensive registrations and pursue cybersquatters is staggering.” (Microsoft)
• “Brand owners are also registrants and have thousands of domains of which we use less than one hundred – all the rest are for protection. We spend a fortune to protect our brand.” (Unidentified brand owner)

Brand owners use words like “astronomical,” “staggering,” “a fortune,” and “billions.” But is it really true that new gTLDs will force trademark owners into registering new domain names that they don’t really want?

A recent study by Paul Stahura definitively concluded that the answer was a definite “no”, by examining which trademarks (in general) were registered across com, org, net, mobi, biz, info and us. Stahura concludes:

The vast majority of trademark holders are not registering their trademark in all the current generic TLDs, let alone all the TLDs.

Bhavin Turakhia of Directi agreed, and provided an explanation:

Trademark holders have no reason to register a domain name in a new TLD if the domain name is not going to get any traffic. Speculators have no reason to register a domain name in a new TLD if the domain name is not going to get any traffic, since they will be unable to generate revenue from it or sell it to the trademark holder.

Finally, Dennis Carlton, an ICANN-appointed economist, noted:

… claims that the introduction of new gTLDs will necessitate widespread defensive registrations appear to be exaggerated and are inconsistent with the oft-noted observation that there have been a limited number of registrations on gTLDs introduced in recent years.

These two sets of voices are completely at odds. On the one hand, the adamant insistence by brand holders that defensive registrations in new gTLDs are necessary and costly. On the other, equally insistent dissenting voices saying that brands neither register defensively in new gTLDs, nor do they need to. We attempted to look systematically at the data to try to determine who is correct, or if the truth is somewhere in the middle.

I. Methodology

In attempt to shed more light on the question, we took a look at the new gTLDs launched since 2000, and analyzed registration patterns of major brands across all open gTLDs.

We took Fortune Magazine’s list of the top 100 companies in the United States. Then we found the major brands for each of these companies. Some companies had just a few brands, some had many. In the case where the brand was hyphenated, we also examined the non-hyphenated version of the brand (e.g., Wal-Mart and Walmart). Overall, we looked at 1043 brands and unhyphenated brand variants. On average, this works out to about 10 brands per company.

Then, with the help of DomainTools, we examined the whois records for each brand and brand variant across 13 different gTLDs, listed below. We looked at whether the brand name was available for registration, or if it had been registered, reserved, or was otherwise unavailable for registration. (We did not examine who the owner was, because our study is looking not at the frequency of cybersquatting or fair use registrations, but rather the interest in the domain name by anyone at all, whether by the brand owner or other registrant.)

Specifically, we looked at the following gTLDs, because anyone (either the brand owner or a cybersquatter) can register a name in them with minimal effort or cost:

• .com, .net, and .org. These pre-ICANN gTLDs are open for registration by anyone on a first-come, first-served basis.
• .info – Launched in 2001. Open to anyone.
• .biz – Launched in 2001. Technically only for businesses, but in practice available to anyone.
• .name – Launched in 2002, originally intended for personal names, but now open to general registrations.
• .pro – Assigned in 2002, launched in 2006. Initially designed for certified professionals, now easily registered by anyone through a number of ICANN-accredited registrars.
• .mobi – Launched in 2005 by mTLD, targeting the mobile internet. The restriction that .mobi sites follow mTLD-sponsored standards for mobile-device viewing was never enforced. Open for general registration.
• .travel – a TLD launched in 2006 targeting the travel industry, originally quite restrictive, but now essentially open. In the past 18-24 months, the criteria of qualification has widened and is easily obtained.
• .cat – Launched in 2006. The only requirement for a .cat is that there is some content somewhere in the Catalan language, a requirement that is trivial to comply with.
• .tel – Launched in 2008. Open to all registrants.
• .asia – Launched in 2006. Registrant must be based in Asia. Most F100 companies are eligible because they have Asian operations, but even if not, proxy services are easily available.
• .jobs – Launched in 2005. Specifically targeted at companies and brand holders.

We then tabulated the results, available as Appendix 1 of this study (download the PDF version of this study for the tabulated results).

II. Results and Analysis

Confirming the insights of Stahura, Turakhia and Carlton, we found that brands are not consistently registered across open gTLDs, and that most brands owners have not registered their names in most of the newer gTLDs.

Some specific results include:

1. Brand owners register their names religiously in .com.

Defensive registrations are a real phenomenon in .com. 100% of the 1043 brands and brand variations are registered in .com. Our earlier study on UDRP filings suggests that this is where the vast majority of cybersquatting also takes place.

2. Outside of .com, even in large and established gTLDs, registration patterns are haphazard. Brand owners do not register defensively here with any regularity.

After .com, the drop-off in registrations is steep. Of the larger, older open gTLDs, we found these percentages:

• .net – 878 of 1043 (84%)
• .org – 790 of 1043 (76%)
• .info – 723 of 1043 (69%)
• .biz – 676 of 1043 (65%)

3. In smaller open gTLDs only 29% of trademarks are registered. Brand owners largely ignore new gTLDs.

For all the rest the combined coverage was a mere 29%, or less than 1/3 of the combined coverage in .com, .net, and .org. In only one of the newer open gTLDs was coverage above 50%. The results in rank order:

• .mobi – 592 of 1043 (57%)
• .asia – 461 of 1043 (44%)
• .tel – 477 of 1043 (46%)
• .name – 341 of 1043 (33%)
• .pro – 182 of 1043 (17%)
• .jobs – 180 of 1043 (17%)
• .travel – 108 of 1043 (10%)
• .cat – 59 of 1043 (6%)

4. Brand names are more likely to be registered in heavily marketed TLDs.

Among the newer gTLDs, .mobi, .tel, and .asia, which have been the most heavily marketed of recent gTLDs, had by far the most registrations, beating out even the long-established .name.

5. The cost of defensive registrations to brands results in part from the 300% markup over street price that they pay for their domain names.

Brand owners, for reasons best known to themselves, pay a significant premium over retail to register domain names through brand management companies, and this is a large component of their costs, which would be much lower if they paid street prices. Overstock.com, for instance, estimates “a conservative annual registration fee of $30 per domain name.” However, as of this writing, the retail price for a .info domain name is $.89,.biz $5.99, .net $5.99, while a .com domain name is $7.29 per year.

IV. Examples of brand registration patterns.

For the purposes of illustration, we examined the largest brands of some of the more vehement opponents to new gTLDs. We found that they do not practice blanket defensive registrations.

Verizon has vocally opposed any expansion of the new gTLD space and is a clear opponent of new top-level domains. One the key brands of Verizon is “FIOS.” Fios.com, fios.net and fios.org are all registered to Verizon. But Verizon has not registered fios.name. For under $15.00 per year, anyone can, today, register that name:

Microsoft is also a leading opponent of new TLDs and EOIs. It has, however, not registered “win32,” one of its key brands, in .tel:

Disney does not protect its theme-park brand. Disneyparks.com is registered, but disneyparks.biz is available for $8.99, as is disneyparks.asia:

Philip Morris has pre-registered its key cigarette brands across all gTLDs. But many of its non-cigarette brands are readily available, such as “triscuit”:

There are many such examples, as our study shows.

Large brands, and their brand protection agencies such as Mark Monitor, heavily concentrate defensive registrations in .com, and are far less interested in other TLDs, and when it comes to the newest TLDs, they have shown very little interest at all.

V. Conclusion

If past behavior is predictive of future actions, we will see a minimal amount of defensive registration activity in new gTLDs by brand owners, and new gTLDs will be largely ignored by cybersquatters.

The complete manual for Espresso, our registry platform, is now available online. The online manual features each and every screen that a new top-level domain registry operator will use in day-to-day administration, along with helpful explanatory appendices and diagrams. As Espresso evolves and improves, the manual will be updated as well.

Please, have a look at the Espresso manual. Comments and questions welcome.

We’ve added a new ICANN timeline for new gTLDs — colorful, graphical, suitable for framing. It’s somewhat approximate during/after the ICANN approval stage, because we don’t know how long that will take (it will depend on how many apps they get), and because your Sunrise and Landrush periods may be different (more or less lengthy) than what we have here.

Our latest timeline is always available on the timeline page.

Comments/corrections welcome.

Preview below: