Accor has trade mark registrations for “Cairns Harbour Lights” and “Harbour Lights”, which it used to promote accommodation at the Harbour Lights complex in, you guessed it, Cairns. It sued Liv Pty Ltd which was renting out apartments owned in the complex by others.
Amongst other things, the trial judge had held that “Cairns Harbour Lights” should be expunged from the Register and the registrations for “Harbour Lights” should be amended to remove some of the services including “accommodation rental services” and “rental of accommodation”. Liv, however was found to infringe through the use of:
(a) “Harbour Lights Cairns”; and
(b) “cairnsharbourlights.com.au”; and
(c) “harbourlightscairns.com.au”; and
The Full Court has now allowed Accor’s appeal, revoking the trial judge’s orders to expunge “Cairns Harbour Lights” and remove some of the services for which “Harbour Lights” is registered.
At 360+ paragraphs, more detailed consideration will have to wait.
One interesting aspect is that the Full Court confirmed that Liv’s use of “keywords” (really metatags) in the source code of its website was trade mark use and so infringing:
323 The title used in the source data is “Cairns Luxury Accommodation – Waterfront Apartments – Harbour Lights – Cairns Queensland”. The primary judge finds that the use of the words “Harbour Lights” in that title appears to be merely a description of the waterfront apartments referred to in the title: PJ at . As to the use of the keyword “Harbour Lights” (as described by the primary judge at  and quoted above), the primary judge regarded that use as also a reference to the apartments as those words appeared in the context of surrounding words such as “Cairns apartments”, “waterfront, luxury apartment” and “harbourside”. Thus, the words were not used as a badge of origin: PJ at .
324 The other words used in the source data as recited at  by the primary judge are these:
“content: = Harbour Lights Apartments in Cairns offer luxury private waterfront apartment accommodation for holiday letting and short-term rental”.
325 As to those words, the primary judge finds that the use of the words “Harbour Lights Apartments” in that phrase was, effectively, use as a business name for a business which offers “accommodation for letting and short?term rental” thus operating as a badge of origin to distinguish Liv’s services from others: PJ at . Such use is use of a mark substantially identical with and deceptively similar to each of the registered trade marks in suit. It is use in relation to each of the Class 36 and Class 43 services other than “commercial real estate agency services”, “agency services for the leasing of real estate properties” and “hotel services”: PJ at . (emphasis supplied)
Following last week’s post where Arnold J found Marks & Spencer liable for buying ads on the keyword INTERFLORA because of the initial interest confusion, the 10th Circuit Court of Appeals in the USA has heavily qualified when (perhaps that should be “if” or “if ever”) initial interest confusion can constitute trade mark infringement in the USA.
It involved Lens.com buying ads on the keyword, 1800contacts, for contact lenses.
The 10th Circuit upheld the trial judge’s exclusion of a consumer survey proferred by 1-800 Contacts to establish confusion. There were a number of reasons for its rejection including its flawed methodology. The 10th Circuit went on to conclude that the approximately 7% confusion shown by the survey would be insufficient to rise to trade mark infringement under US law. Accepting that each case depended on its own facts, the 10th Circuit endorsed the general proposition that:
The great weight of authority appears to be that “[w]hen the percentage results of a confusion survey dip below 10%, they can become evidence which will indicate that confusion is not likely.” 6 McCarthy § 32:189 at 32-440 (emphasis added by 10th Circuit).
One wonders whether an Australian court, which must ascertain whether a [substantial][OR a significant] number of the relevant audience might be caused to wonder, would be so robust as to conclude that 5% or 7% of the market was not a substantial (or significant) number?
Professors Goldman and Tushnet identify a range of reasons to regret the 10th Circuit’s decision not to drive a stake through the heart of initial interest confusion. Prof. Goldman does speculate, in particular, whether measuring “click-throughs” as a proxy for confusion could ever cross the 10% threshold.
Marks & Spencer has been found to have infringed Interflora’s trade mark in the UK by ‘buying’ ads triggered by Google searches for the keyword INTERFLORA.
An example of the ads Arnold J found infringing:
In Google France, the CJEU established that an advertiser would infringe a registered trade mark when its ads were triggered by a trade mark as a keyword where: 
“82 The essential function of a trade mark is to guarantee the identity of the origin of the marked goods or service to the consumer or end user by enabling him to distinguish the goods or service from others which have another origin (see, to that effect, Case C–39/97 Canon  ECR I–5507, paragraph 28, and Case C–120/04 Medion  ECR I–8551, paragraph 23).
83 The question whether that function of the trade mark is adversely affected when internet users are shown, on the basis of a keyword identical with a mark, a third party’s ad, such as that of a competitor of the proprietor of that mark, depends in particular on the manner in which that ad is presented.
84 The function of indicating the origin of the mark is adversely affected if the ad does not enable normally informed and reasonably attentive internet users, or enables them only with difficulty, to ascertain whether the goods or services referred to by the ad originate from the proprietor of the trade mark or an undertaking economically connected to it or, on the contrary, originate from a third party (see, to that effect, Céline, paragraph 27 and the case-law cited).
85 In such a situation, which is, moreover, characterised by the fact that the ad in question appears immediately after entry of the trade mark as a search term by the internet user concerned and is displayed at a point when the trade mark is, in its capacity as a search term, also displayed on the screen, the internet user may err as to the origin of the goods or services in question. In those circumstances, the use by the third party of the sign identical with the mark as a keyword triggering the display of that ad is liable to create the impression that there is a material link in the course of trade between the goods or services in question and the proprietor of the trade mark (see, by way of analogy, Arsenal Football Club, paragraph 56, and Case C–245/02 Anheuser-Busch  ECR I–10989, paragraph 60).
Arnold J found on the evidence that a significant section of the public were confused and so Marks & Spencer infringed.
It is not possible to do justice in a blog post to the full range of reasons contributing to his Lordship’s conclusion. Some that stand out follow.
Arnold J accepted (at ) that the majority of UK internet users appreciated the difference between paid ads and natural or organic search results, but there was still a significant proportion of internet users in the UK who did not. His Lordship also accepted that “nowadays” the majority of consumers appreciate .
nowadays the majority of consumers appreciate [they are being presented with ads by competitors to the brand they had searched for]. But I consider that a significant proportion do not. (emphasis supplied)
Secondly, the nature of the INTERFLORA brand appears to have been crucial. As you will no doubt be familiar, INTERFLORA is a network. It operates through a network of agents who are usually (always?) branded under their own names and trade marks. Those customers who were not buying online from “interflora.co.uk” or “interflora.com”, for example, would typically go into a retail outlet operating under its own name (and which may display the INTERFLORA name and logo).
The significance of this (at  and ) was that there was great potential for those customers who realised they were dealing with Marks & Spencer when they clicked on the ad mistakenly to think it was part of the INTERFLORA network.
That potential was in fact borne out by the evidence. In particular, there was evidence from “Hitwise data” that people who (1) searched on the keyword INTERFLORA and (2) as a result clicked on a Marks & Spencer ad generated in response to the search (3) were between 44 and 106 times more likely than the average visitor to the M & S flowers site to leave the M & S site without purchasing and instead go on to an INTERFLORA site.
At  – , his Lordship accepted the propositions that:
a significant number of consumers in Segment A decided after they had clicked through to the M & S website that it was not where they wanted to be and went to the Interflora website instead. The second is that the reason for this change of mind was that those consumers had clicked through from the M & S advertisement because they assumed from the appearance of the advertisement in response to their search that M & S was part of the Interflora network, but they realised that that was not the case when they clicked through to the M & S website and saw no reference to Interflora.
and such “initial interest” confusion was itself enough for trade mark infringement.
Thus, while Google doesn’t infringe by ‘selling’ keywords, the advertiser may and, in this case on Arnold J’s findings, did.
What, if anything, does it mean for us?
First off, the judgment is full of fascinating details about the “AdWords” and “search” market and the strategies that businesses deploy. For example, it appears that Google held around 90% of the global search (and paid advertising relating to search) market, with Bing and Yahoo! trailing out of site. There are also discussions of market research reports and Ofcom studies into what consumers understand when using the internet.
It is not clear whether Arnold J’s reasoning will provide us with much assistance here. First, while EU law does not appear to require use as a trade mark to infringe, the origin function referenced by the CJEU in Google France appears similar to our concept of use as a trade mark – as a badge of origin or to identify the trade source.
Secondly, it seems doubtful that the considerations identified in  of Google France would be relevant at all under our law. The idea of examining whether the “normally informed and reasonably attentive internet users” could ascertain the trade source from the ad, or do so “only with difficulty”, indicates that the content of the advertisment may make it clear that the trade mark owner is not the source of the advertised product. However, a registered trade mark is infringed in Australia even if the trade source is made clear, for example by a disclaimer or other identifying factor. The type of analysis being engaged in under EU law is rather more like what would take place in a passing off action or action for misleading or deceptive conduct.
Thirdly, our law does recognise the idea of “initial interest confusion”, but the number of people who apparently went to M & S’ website and “clicked away” does rather highlight the difficulties with the concept as applied to web searches: clicking the back button, or even doing another web search in the browser, is not so costly as walking out of the shop, hopping back in your car and going looking for the intended destination.
Arguably, the most significant point could be the starting point identified at  – the Court of Justice’s recognition that:
keyword advertising is not inherently or inevitably objectionable from a trade mark perspective. On the contrary, the case law of the CJEU in this field recognises that, as a general rule, keyword advertising promotes competition ….
On this view, it was only the very special nature of INTERFLORA as a ‘network’ that convicted Marks & Spencer.
For example Southern Cross v Toowoomba at  but, at least in the context of misleading or deceptive conduct / passing off (yes, I know this is a post about registered trade marks) some brake may be imposed on that in at least some cases. ?
The ACCC has successfully appealed the Google Adwords case for misleading and deceptive conduct.
So, for example, Alpha Dog Trainging has been operating a dogtraining business for 12 years. Dog Training Australia (Ausdog) bought ads on the keywords Alpha Dog Training through Google’s Adwords program. One ad generated was:
Alpha Dog Training
DogTrainingAustralia.com.au All Breeds. We come to you. No dog that can’t be trained.
Instead of being taken through through to Alpha Dog Training’s website, however, a user who clicked on the ad was taken through to Ausdog’s website.
A clear case of misleading or deceptive conduct by Ausdog.
Because of its role in “selecting” which ads got placed in what order, Google has also been found liable.
Prof. King, formerly an ACCC commissioner, highlights why and thinks the Court got it seriously wrong.
Google’s placement of advertisements, generated through its AdWords program, on search results pages is not misleading or deceptive conduct contrary to s 52 TPA / s 18 ACL (I’m afraid you have to scroll down). However, the advertiser’s use of another trader’s name in the headline for an advertisement which had nothing to do with that trade was.
So for example, the Trading Post used the AdWords program to generate an ad:
www.tradingpost.com.au New/Used Fords – Search 90,000 + auto ads online. Great finds daily!
The advertisements at the URL did not have anything to do with Kloster Ford or vehicles Kloster Ford was offering for sale.
The Trading Post therefore had engaged in misleading or deceptive conduct; Google did not.
Australian Competition and Consumer Commission v Trading Post Australia Pty Ltd  FCA 1086