Two recent cases tell us a lot about how arbitrators go about reaching their decisions and the considerations that influence those decisions. What are they influenced by and what are they... not influenced by?
THE ARM CASE
In the first case, ARM, the Complainant had a trademark for ARM, in fact a lot of them, so it got off to a good start. Also, the trademark was well known, another good point for the Complainant. And there were a lot of the ARM trademarks on the books before the domain name was registered, so that pesky problem for complainants, that the domain name had been registered before its trademark had been established, did not arise. The domain name was also identical to the trademark; both of them were simply ARM. So, some observers would say this was a copy, that the domain name registrant had simply copied the trademark and was showing signs of wanting to sell it, in fact for a very high price. But that is not the end of the story. Those points were all in favour of the complainant trademark owner, and the respondent was not represented by a lawyer, and yet the complainant still lost. How come?
The complainant had won on the first element, because the domain name was identical to the trademark; they were both ARM.
The complainant/ trademark owner also won round 2, because the panel found that the respondent had not shown a right or legitimate interest in the domain name. “Arm” was a common, or dictionary word, of course but that was not good enough for the panel, because it held that it was not enough merely to have a dictionary or generic word as your domain name. The panel readily acknowledged that “arm” was a known word with a dictionary meaning. Not enough, said the panel. You also had to show that you had genuinely used it within its normal meaning. We often wonder where that rule is supposed to have come from, as it is not in the Policy, and it seems to get its validity from the fact that it is frequently repeated, but there it is. It is, therefore, the general approach of panellists that a generic word will give not rise to an RLI unless it has been used within its usual meaning. This one was not. It had not been used within its usual meaning but, rather, had been used for the purpose of putting it up for sale on Sedo, the auction site, at a high price. As the distinguished panellist said, “there is no evidence that the Respondent has made demonstrable preparations to use the disputed domain name in connection with a bona fide offering of goods or services.”
So, all of these points moved inexorably in favour of the complainant. From that point on, all it had to do was show that the respondent had registered and used the domain name in bad faith. It was destined to win. But it did not win. It lost. How? Well, because the Policy is clear that the complainant has to prove its case and one of the things it must prove is that the respondent registered the domain name in bad faith. It had not done so. You might ask: what else could it have proved? That is probably the clue to why the complainant lost, i.e., that it might be easier to say what was not proved. Here are some of the things that did NOT happen.
The domain name could NOT be said to stand only for the trademark. The word “arm” is so common as an English word, and it could be an abbreviation, so you cannot assume that the registrant was having a “go” at the ARM trademark or its owner.
There was NO bad faith tone about what the respondent was doing. There was NO targeting. All the respondent was doing was trying to sell the domain name.
There was NO evidence to suggest the Respondent knew of the Complainant and its ARM trademark.
There was NO evidence that the respondent was trying to sell goods and services of the sort the complainant sold.
Internet users were NOT being mislead.
It is not enough to show the respondent/ registrant had a domain name it was trying to sell which was the same word as the trademark, no matter how suspicious that might seem, if the word in the domain name had a wider use than as a duplicate of the trademark, which it did in this case. If the respondent has not actually done anything inimical to the rights of the complainant, it still has a good chance of winning, which it did in this case. The lesson is that, to win, the trademark owner must not be handicapped or compromised by not filling in any of the above gaps i.e., by not proving the essential elements that it must prove.
THE CORNING AND CORMING CASE 
Corning is in materials science, but a lot of people would associate it with the ceramics of the dining room table and the kitchen. It must surely be a household name. The domain name was <corming.com> and some people have made the mistake of seeing CORNING and convincing themselves that this is the word they are seeing, when it is actually CORMING. So, the domain name was <corming.com>, not <corning.com>.
The Respondent was the prominent reseller of domain names and had put the disputed domain name, (<corming.com>), up for sale on its own website but had not used it for an active website or web page.
The Complainant proved its CORNING trademark, not surprisingly, and the domain name was held to be not identical, but confusingly similar to the trademark.
The complainant argued this was a classic typo case because CORMING would be read as CORNING. But, said the three-person panel, some people use CORMING as a surname, so the domain name could have been a deliberate invocation of CORMING and not a typo copy of CORNING. The Respondent could then establish that it had a right or legitimate interest in the domain name; if the word was used as a surname, this gave rise to a use not associated with the CORNING trademarked name. Moreover, the Complainant had not shown it had any trademark rights in CORMING.
As in the ARM case, there was also NO evidence that showed that the Respondent registered the disputed domain name to take advantage of Complainant’s claimed rights in CORNING. There was not even any evidence to show that the CORNING mark was famous, (which might have created an inference of fame) although one would have thought this would be easy to do if you understood what you had to prove and hence the evidence for which you had to go looking. Indeed, there was evidence from the respondent that CORNING was used for other goods and services and street addresses. Thus, the Respondent had not used the domain name for any use connected with the Complainant.
The typosquatting argument is beguiling at first appearance, but continually repeating without evidence to back it up would not turn a sow’s ear into a silk purse.
Also, the fact that the respondent was a reseller did not cause it any trouble, as “investing and reselling domain names is a legitimate business when such is not being done to take advantage of the trademark value associated with a particular term”. This what really got the Respondent home.
On the issue of bad faith, the Panel had no difficulty in holding that as the registration itself was legitimate, as the word was a surname in common use, there had been no targeting and as the offer to sell made no reference to the Complainant or its trademark, the Respondent had not acted in bad faith.
THE TWO DECISIONS TAKEN TOGETHER
These two decisions have a common thread. They both “looked” as if they were typosquatting cases, i.e., it looked on first impressions that the Respondent had just copied the Complainant and then tried to sell the domain name. In both cases the panels insisted on going further. Was the Respondent aiming at the Complainant? No, it was not, so there was no intention to commit a breach of the Policy. Moreover, there was no evidence that the Respondent had acted wrongly. In particular, selling a domain name is not wrong, except in the limited cases set out in the Policy, namely a sale to the complainant or a competitor. And there was no cross-over to a point where the Respondent was using the domain name to sell goods and services which were associated with the domain name.
Above all, the two cases show that what at first sight might be seen as open and shut decisions for the complainant turned out to be victories for the respondent. They also show there is no magic in the form in which cases come before a panel; the ARM case was before a single panelist, the Respondent was not represented by a lawyer and it was decided on the basis of what the Complainant had NOT proved; the CORMING case was a three person tribunal, the Respondent was self-represented and the case was vigorously defended with strong arguments supported by evidence adduced by the Respondent. The best advice to all parties is: do not take risks, but assess your case and decide what evidence you can bring forward and then do so.
 The Hon Neil Brown QC was a panellist in this case.