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. ?