By definition “[a]ny word” or “any combination [of words]” can function as trademarks, but whether alone or combined for that purpose no use can overrule their ordinary meanings, support their removal from the public domain, or prevent speculation or use of identical or confusingly similar words by businesses other than rights holders. The fact is, all words that are not coined found in dictionaries, word lists, and circulating in world cultures are generic. It is the manner in which words are used and assembled that elevate them to higher categories of distinctiveness. Anyone can hold <internationalbusiness.com> or <wallstreet.com> but only the rights holders can own INTERNATIONAL BUSINESS MACHINES and WALL STREET JOURNAL. Any mark consisting of words alone or combined as phrases categorized as descriptive (even those carrying suggestive connotations) that can be used in their generic senses, will be.

The consequence of selecting and composing words is a recurrent theme in UDRP decisions. In recent cases, dictionary words in English and other languages include “dialoga.” “karma,” “wallabies,” “legally,” “citadel,” and “telephon”; descriptive phrases include “cell gene,” “power agent,” “cloud insure,” and “secure lock.” The lower the classifications are the heavier the demand for proof of cybersquatting. In determining the strength of a mark, U.S. courts focus on two factors: (1) the mark’s inherent distinctiveness and (2) its acquired distinctiveness in the marketplace; constituting high to low in strength. Complainants whose marks are inherently distinctive will almost certainly prevail; those with marks that qualify through acquired distinctiveness are vulnerable to having their complaints of cybersquatting dismissed, unless they satisfy UDRP’s evidentiary demands as illustrated in the Celgene and Power Agent cases discussed below. It is this certainty that supports the secondary market in domain names.

In the universe of trademarks, dictionary words and common phrases far outstrip uncommon combinations and coined (neologisms) as marks. So, for example, domain names incorporating uncommon combinations of words such as LEADER IN ME, ROOMS TO GO, and MONSTER ENERGY (all recent transfers) will be canceled or forfeited to rights holders. This rests on the unlikelihood of others inventing the same uncommon, unexpected, even surprising combinations as rights holders, although where the domain name predates the mark no such inference can be drawn, as in SECRET LAB (<secretlab.com>) registered 17 years before the registration of the mark.

In contrast, where words and phrases as marks are common (neither unexpected nor surprising in composition), they will be seen as interchangeable with domain names, a result likely to favor respondent, although the outcome may depend on other factors. Typing errors (<trueevalue.com>) are infringing but deliberate misspelling (<cedit.com>) are not. Where respondent defaults, and there is no explanation for its registration and use of an uncommon phrasing, it is more likely Panels will infer bad faith. GOLD FIELDS, for example, not exactly uncommon in the mining industry perhaps but less so as a common phrase: and in DATA PATH, common in the data analysis field, but suspicious when used out of context. For <datapath.net> the Panel suggested the decision could have favored Respondent had it offered affirmative proof of rights or legitimate interests, but its proof was unpersuasive. The Gold Fields case has been criticized for drawing inferences from Respondent’s silence but complainants too (although not in this case) are also vulnerable to negative inferences as illustrated in the Airtron case (incompetently presented; complainant failed to establish a prima facie case that Respondent lacked rights or legitimate interests. Not further discussed in this essay).

The procedure for drawing inferences of bad faith (a form of extrapolating from known facts to probable imports) rather than resting entirely on concrete evidence is a normal product of logical reasoning. What goes into a record or is missing from it has consequences (the Airtron result). Decision makers are interested in the reliability of facts. Assertions supported by documentary evidence are more reliable on their face, and where there are inconsistencies between supported and unsupported (alleged) facts on any given issue, it is natural to draw inferences favoring that which is either more probable than not or regarded as more reliable. The issue is examined also in an earlier essay, Drawing Inferences from the Record

Setting aside what some commentators condemn as decisions favoring complainants in undefended cases, Panels have recognized (even in undefended cases) that respondents acquiring domain names composed of common terms for resale in the “open market” have rights or legitimate interests in them. This has been a principled position from the beginning of the UDRP. The passage of time and multiple decisions has consolidated this view. To illustrate with two cases from the early UDRP docket, Meredith Corp. vs. CityHome, Inc., D2000-0223 (WIPO May 18, 2000) (<countryhome>) and Audiopoint, Inc. v. eCorp, D2001-0509 (WIPO June 14, 2001) (<audiopoint>),

In Meredith, the Panel held that “seeking substantial money for what [Respondent] believes to be a valuable asset is not tantamount to bad faith.” This view shines through in Karma International, LLC v. David Malaxos, FA1812001822198 (Forum February 15, 2019) in which the Panel held that a reseller is “free to place whatever market value it chose on the name,” and that

[n]owhere in the Policy is there a requirement that a respondent is under a positive obligation to use (or surrender an unused) domain name [or that] [f]ailure to use a domain name is ... per se evidence that its owner has no right or legitimate interest in [it].”

On the contrary “[a re-seller of generic domain names . . . has long been held to conduct a legitimate business, whether the names have been used in connection with their generic meaning, or not used at all.” Further, “[r]espondent was under no positive obligation to use the name or to sell it at any particular price.” But, even more important, Respondent acquired the domain name earlier (1994) than the Complainant its mark (2018).

The three-member Panel in Audiopoint offers a different facet of the same view. It held that “speculation in domain names when done without any intent to profit from others’ trademark rights may itself constitute a bona fide activity [as affirmative proof] under paragraph 4(c)(i)” (emphasis added). The qualification of intent goes to the heart of assessing unlawful conduct.

WIPO Overview 3.0 memorializes this consensus view implicitly by limiting the “speculation argument” to “distinctive or famous mark[s].” Panels tend to “view [a respondent defense of legitimate speculation] with a degree of skepticism,” Paragraph 3.1.1. (Since marks are by definition “distinctive” WIPO is obviously referring to arbitrary and fanciful (inherently distinctive) rather than descriptive (acquired distinctiveness) and suggestive marks (acquired or inherently distinctive depending on the choice of words or phrases). Marks composed from common terms tend to suggest or are likely to have multiple associative possibilities which are precisely why Panels recognize a “defense of legitimate speculation.”

From the earliest pronouncements, then, a consensus has formed that word and phrase choices determine rights. This is consistent, incidentally, with corresponding jurisprudence under trademark/cybersquatting law in the U.S. and most likely true in other jurisdictions also. When the matter is examined historically it can be seen that there has been a steady advance in rejecting arguments that attempt to weaken domain resellers’ rights or legitimate interests, how they use their domain names (actively or passively), and their valuations of domain assets.

While seemingly impenetrable to some rights holders whose marks postdate registration of domain names, the fact they have the standing to maintain a UDRP proceeding is not evidence for an actionable claim (nor would it be, incidentally, in federal court under the Anticybersquatting Consumer Act, as illustrated in the UDRP and ACPA filings for <beautiful people.com>, Beautiful People Magazine, Inc. v. Domain Manager / PeopleNetwork ApS / Kofod Nicolai / People Network Aps / Nicolai Kofod / People Network, FA1502001606976 (Forum May 4, 2015) (<beautifulpeople.com>. complaint denied and in federal court dismissed with attorney’s fees Joshua Domond and Harold Hunter, Jr v. PeopleNetwork APS d/b/a Beautifulpeople.Com, Beautiful People, LLC, Greg Hodge, and Genevieve Maylam, 16-24026-civ (S.D. FL. Miami Div. 11/9/17).

The Panel in Karma International rejected Complainant’s argument that the case presented a “unique situation wherein the Respondent may have acted in a manner where bad faith registration can be inferred [since it] was willfully blind to any current or future trademark rights” (emphasis added). The Panel properly held that “this submission [is] fanciful since it lacks any reasonable basis.”

The notion that a rights holder could ever be “a more appropriate owner of [a] disputed domain name [composed of generic elements] than the Respondent ... is misguided,” Dialoga Servicios Interactivos, S.A. v. Finlead AG, D2018-2768 (WIPO February 8, 2019). Why? Because such a notion presumes a right greater than any granted under trademark law. In Dialoga Servicios Interactivos, Respondent (a reseller of domain names) acquired <dialoga.com> after a prior registrant allowed its registration to lapse but years Complainant’s first use of the mark in commerce. While priority is always a key factor, it does not displace the principle that domain names are registered on a first-come-first-served basis.

The Panel in Dialoga Servicios Interactivos accepted Respondent’s rebuttal evidence with the following explanation.

[T]he disputed domain name has value [Respondent attributes to it] because it consists of a common word in no less than three languages which are spoken by many millions of people. The word itself is particularly attractive as a domain name, implying as it does both communication and dialogue.

The Panel concluded that “[t]his is the nature of [the domain name reseller] business.”

Consistent in all these cases are underlying deficiencies of proof. In Dialoga Servicios Interactivos, there is the added censure of omitting evidence. For some, if facts do not fit the wished-for-narrative they fabricate ones that do, which is the basis in both Karma International and Dialoga Servicios Interactivos for revere domain name hijacking. As the Panel noted in Karma International “The explicit claims to bad faith registration and use made in the Complaint are largely specious and the accusations leveled at Respondent are groundless and malicious.”

Moving to combinations of dictionary words, in Celgene Corporation v. PRIVACYDOTLINK Customer 3639471 / Domain Administrator, Portmedia Holdings Ltd, D2018-2673 (WIPO February 11, 2019) Complainant challenged as confusingly similar to CELGENE. However, the Panel found that “Respondent established that these terms are “not infrequently used in conjunction with each other in the context of the related concepts of ‘cell therapy’ and ‘gene therapy.’” Also, “while the Complainant refers to the disputed domain name as a ‘misspelling’ of the Complainant’s mark this is not one of those cases where a domain name is explicable only as a misspelling of a complainant’s trademark.” There was no proof that Respondent ever engaged in typosquatting to support an inference that it was doing that in this case.

Similarly in Darryl Davis Seminars, Inc. v. Privacydotlink Customer 656889 / Domain Admin, Abstract Holdings International Ltd, D2018-2238 (WIPO January 21, 2019) for the two words “power” and “agent”—<poweragent.com>—that could conceivably fit any number of market circumstances. The Panel found

in view of the circumstances of the case, in particular the common and descriptive nature of the term incorporated in the disputed domain name, the Respondent’s offer for sale of the disputed domain name does not demonstrate that it registered and is using the disputed domain name in bad faith, i.e., with a view to selling it to the Complainant.

While Complainant offered only conclusory allegations, “Respondent has satisfied the Panel that it registered the disputed domain name for its inherent value as a domain name incorporating a common descriptive term, as part of its business as an investor in such domain names.” Rights holders should take seriously the Panel’s analysis of what it takes to prevail in a UDRP proceeding: “Complainant has not provided any evidence of use of its trademark and has failed to demonstrate that the Respondent was likely to have been aware of the Complainant and its trademark at the time of its registration of the disputed domain name based on the Complainant’s alleged reputation.”

There has been sufficient development in the jurisprudence that (as Panels have repeated said) parties are expected to be knowledgeable of UDRP’s demands. Even if dictionary words and common phrases are shown to have secondary meaning sufficient to qualify for trademark registration, complainants are nevertheless hard-pressed to establish them as exclusive identifiers of source for their specific goods or services without having concrete proof the domain names were registered with them particularly in mind.

It has become clear too that rights holders with marks composed of dictionary words and common combinations fail, and are exposed to sanctions for reverse domain name hijacking, by misunderstanding the limits to their statutory rights. It is also clear that the jurisprudence has enabled the creation of a vigorous secondary market capable of supporting investors holding multimillion numbers of domain names composed of generic elements to prosper.

The federal docket supports this conclusion. The Court in Advertise.com, Inc. v. AOL Advertising, Inc., 616 F.3d 974 (9^th Cir. 2010), for example, concluded that ADVERTISING.COM (albeit, a registered mark!) was nevertheless common, which means it could not wall off plaintiff from lawfully using <advertise.com>. Similarly, with ENTREPRENEUR and <entrepreneurpr.com>, Entrepreneur Media, Inc. v. Smith, 279 F.3d 1135, 1147 (9th Cir. 2002), and no doubt this will extend to Booking.com which was recently successful in challenging the USPTO’s refusal to register BOOKING.COM,, Booking.com B.V. v. USPTO, et al., 17-2458 (4^th Cir. 2/4/2019) when in its turn it challenges another business or domain name it perceives as cybersquatting on its mark; we can look forward to an ACPA action or UDRP proceeding, and how interesting that will be!