[2013] SGCA 26

Hai Tong Co (Pte) Ltd v Ventree Singapore Pte Ltd and another and another appeal
[2013] SGCA 26

Case Number	:	Civil Appeals Nos 1 and 2 of 2012
Decision Date	:	15 March 2013
Tribunal/Court	:	Court of Appeal
Coram	:	Andrew Phang Boon Leong JA; V K Rajah JA; Sundaresh Menon JA (as he then was)
Counsel Name(s)	:	Dedar Singh Gill, Yvonne Tang, Ong Sheng Li Gabriel and Chong Yi Ling Eileen (Drew & Napier LLC) for the appellant in Civil Appeal No 1 of 2012 and the respondent in Civil Appeal No 2 of 2012; Wong Siew Hong and Ong Zhenhui Wayne (Eldan Law LLP) for the respondents in Civil Appeal No 1 of 2012 and the appellants in Civil Appeal No 2 of 2012.
Parties	:	Hai Tong Co (Pte) Ltd — Ventree Singapore Pte Ltd and another

[LawNet Editorial Note: This was an appeal from the decision of the High Court in [2012] SGHC 104.]

15 March 2013

Judgment reserved.

Sundaresh Menon CJ (delivering the judgment of the court):

Introduction

1 There are two appeals before us arising out of Suit No 669 of 2010 (“Suit No 669”). Suit No 669 was brought by Hai Tong Co (Pte) Ltd (“Hai Tong”) against Ventree Singapore Pte Ltd (“Ventree”) and Sky (sued as a firm) (“Sky”) (collectively referred to as “the Defendants”) for trade mark infringement under s 27(2)(b) of the Trade Marks Act (Cap 332, 2005 Rev Ed) (“the Act”) and in the tort of passing off. The first appeal, Civil Appeal No 1 of 2012 (“CA 1 of 2012”), is Hai Tong’s appeal against the decision of the High Court judge (“the Judge”) dismissing its claim in passing off and awarding it only 90% of its costs of the proceedings. The second appeal, Civil Appeal No 2 of 2012 (“CA 2 of 2012”), is a cross-appeal brought by the Defendants against the Judge’s holding in favour of Hai Tong in its claim for trade mark infringement.

Background to the appeals

Parties to the dispute

2 Hai Tong is a company incorporated in Singapore engaged in the business of importing, exporting, distributing and selling cosmetics, sundries and beauty accessories, including a range of cosmetic products that has been marketed under the brand “Lady Rose”. Hai Tong is the registered proprietor of the following mark (“the Composite Mark”):

Trade Mark No	Trade mark	Specifications
T7048436B		Perfumery, essential oils, hair lotions, hair creams and cosmetics.

3 Ventree is a company incorporated in Singapore engaged in the business of importing, exporting and distributing cosmetic products. Ventree also distributes a line of cosmetic products. These products are sold under the brand name “Rose Lady” pursuant to a Sole Agent Agreement dated 15 July 2009 between Ventree and the manufacturer of these products, VOV Cosmetics Co Ltd (“VOV”). VOV, which is party neither to the proceedings below nor to the present appeals, is a company incorporated in South Korea and the registered proprietor of the “VOV” trade mark in Singapore. Sky is a partnership incorporated in Singapore. It operates a retail department store at People’s Park Complex selling, among other things, cosmetics and perfumes, including cosmetic products under the “Rose Lady” brand.

Background facts

4 The Composite Mark was registered by Hai Tong’s predecessor in 1970. It has been in use since 1960, even before it was registered. Hai Tong distributes “Lady Rose” products to various retailers across Singapore. These include department stores such as Robinsons as well as other retail cosmetics stores in the heartlands. Hai Tong also sells these products to walk-in customers at its shop in Sago Lane.

5 Since 15 July 2009, Ventree has been importing three of VOV’s “Rose Lady” products, namely, “VOV” Rose Lady Cover Foundation, “VOV” Rose Lady Skin Cover and “VOV” Rose Lady Two Way Cake (“the allegedly infringing products”). The allegedly infringing products contain an ingredient, rosa canina, or essence of dog rose. They are sold in peach-coloured packaging printed with the product name, which includes the words “Rose Lady” (“the Sign”) accompanied by sketches of flowers with five petals in various sizes in the background. The Sign was never registered under the Act. The allegedly infringing products were sold at various retail shops across Singapore, including department stores such as John Little and some retail cosmetics shops, including Sky’s retail store. Sky previously used to sell the allegedly infringing products until sometime in 2005 when it ceased doing so. It resumed selling one of the allegedly infringing products, namely, “VOV” Rose Lady Skin Cover, in January 2010.

6 On 1 August 2010, Hai Tong instituted Suit No 669 in the High Court against the Defendants, claiming that they had infringed the Composite Mark by importing and offering for sale the allegedly infringing products. Hai Tong also claimed that the Defendants had passed off the allegedly infringing products as Hai Tong’s goods.

7 Ventree claimed in its defence against Hai Tong’s claim for trade mark infringement that:

(a) the words “Rose” and “Lady”, as well as the flowers printed on the packaging of the allegedly infringing products, were not used as trade marks to indicate the origin or source of manufacture, but were instead used in a descriptive manner to describe the rosa canina ingredient used in these products, with the flowers serving a purely decorative purpose;

(b) its cosmetic products were in fact marketed and sold under the “VOV” brand;

(c) the words “Rose” and “Lady” and the flowers printed on the packaging of VOV’s cosmetic products were entirely different from the words and the floral device in the Composite Mark; and

(d) there was no likelihood of any confusion that the two lines of products might have a common origin.

8 Ventree also denied Hai Tong’s claim in the tort of passing off, and counterclaimed against Hai Tong under s 35(2) of the Act for groundless threats of trade mark infringement proceedings. Sky’s position was materially similar to that of Ventree. Indeed, the Defendants were represented by the same counsel in the proceedings below as well as in the present appeals.

Decision below

9 The Judge allowed Hai Tong’s claim for trade mark infringement under s 27(2)(b) of the Act, but dismissed its claim in the tort of passing off. The Judge’s grounds of decision can be found in Hai Tong Co (Pte) Ltd v Ventree Singapore Pte Ltd and another [2012] SGHC 104 (“the GD”).

10 As regards Hai Tong’s claim under s 27(2)(b) of the Act, the Judge found that Hai Tong had successfully proved all the elements necessary to constitute a valid claim of trade mark infringement. His analysis may be summarised as follows:

(a) The Sign was not merely a descriptor of the rosa canina ingredient, but was used in the course of trade, ie, as a badge of origin (even though it functioned as a “secondary identifier” to the “VOV” mark) (at [18] of the GD).

(b) The Sign was visually, aurally and conceptually similar to the Composite Mark and, in the overall comparison, similar to the Composite Mark (at [19] and [23]–[25] of the GD).

(c) The Sign was used on goods that were identical to those for which the Composite Mark had been registered, namely, cosmetic products comprising ladies’ skin cover and foundation products (at [27] and [30] of the GD).

(d) There was a likelihood of confusion on the part of the public that the allegedly infringing products came from the same source as “Lady Rose” products or from economically-linked sources. This was because: (i) there was a high degree of similarity between the Composite Mark and the Sign; (ii) both the Composite Mark and the Sign were applied to goods of an identical nature; (iii) these goods were marketed in similar retail locations such as department stores and cosmetics shops; and (iv) the goods targeted the same group of customers, namely, females who used cosmetics (at [30]–[32] of the GD).

11 As regards Hai Tong’s claim in the tort of passing off, the Judge dismissed it for the following reasons:

(a) Hai Tong only had “trivial” goodwill (at [39] of the GD). The Judge based this on Hai Tong’s sales figures, which, in his view, were “not particularly impressive” (at [36] of the GD), as well as on the lack of any effort by Hai Tong to advertise its products since 1979 (at [36] and [39] of the GD).

(b) Hai Tong failed to adduce any evidence that there was damage or a likelihood of damage to its trivial goodwill. The Judge noted that there was “no good commercial reason” (at [40] of the GD) for the Defendants, who, in his view, were importing or offering for sale products marketed under a better known brand than “Lady Rose”, to pass off their goods as those marketed under an inferior brand (at [38] and [40]–[41] of the GD).

12 In the light of his decision, the Judge granted the following orders, that:^{[note: 1]}

1. [Ventree] be restrained, whether acting by its directors, officers, employees, agents or any of them or otherwise howsoever, from infringing [the Composite Mark];

2. [Sky] and each of its partners be restrained, whether acting by himself, his employees or agents or any of them otherwise howsoever, from infringing [the Composite Mark];

3. The Defendants do deliver-up and/or destruct upon oath all products, printed or written matter, packaging and all other articles in the power, possession, custody and/or control of [the Defendants] and each of them, the use of which would be in breach of the foregoing injunctions;

4. The Defendants do pay [Hai Tong] damages, or at [Hai Tong’s] option, an account of profits to be assessed, [together with interest thereon;]

...

6. The Defendants’ counterclaim be dismissed; and

7. The Defendants do pay [Hai Tong] 90% of [Hai Tong’s] costs of the trial.

Issues before this court

13 The parties, between them, have appealed against the whole of the Judge’s decision. Accordingly, the issues before us are: (a) whether the Judge was correct in finding that the elements of Hai Tong’s trade mark infringement claim under s 27(2)(b) of the Act had been made out on the evidence; and (b) whether the Judge was correct in finding that Hai Tong only had “trivial” goodwill and that it had failed to prove that there was damage or a likelihood of damage to such goodwill.

Our decision in CA 2 of 2012

14 We first deal with the appeal in CA 2 of 2012, which is the Defendants’ appeal against the Judge’s holding that the Composite Mark had been infringed. The claim in this case was founded on s 27(2)(b) of the Act, but it bears noting that the Act prescribes a cascading regime for trade mark protection that varies according to the degree of similarity between the registered mark and the allegedly infringing sign and the degree of similarity between the goods or services for which the mark is registered and those in relation to which the allegedly infringing sign is used. At the simplest level, where the registered mark and the allegedly infringing sign are identical and the goods or services for which the mark is registered and those in relation to which the allegedly infringing sign is used are also identical, infringement is constituted without more under s 27(1) of the Act. On the other hand, where there is no similarity between the goods or services for which the mark is registered and those in relation to which the allegedly infringing sign is used in the course of trade, several further elements will have to be shown. These are set out in s 27(3) and include proof that the mark in question is well known, and that the use of the mark on the alleged infringer’s goods or services would indicate a connection between those goods or services and the proprietor of the mark; and further, that as a result, there is a likelihood of confusion on the part of the public and a likelihood that the interests of the trade mark proprietor will be damaged.

15 The trade mark infringement claim under s 27(2) of the Act occupies a space between the preceding two categories of infringement. Specifically, for its claim under s 27(2)(b) to succeed, Hai Tong has to prove the following elements:

(a) the Sign is similar to the Composite Mark;

(b) the Sign is used in relation to goods or services identical with or similar to those for which the Composite Mark is registered; and

(c) on account of the presence of the first two conditions, there exists a likelihood of confusion on the part of the public.

16 These requirements are cumulative, so that if either of the first two elements is not satisfied, an inquiry into the third will not arise and the claim under s 27(2)(b) of the Act will fail (see this court’s observations in Sarika Connoisseur Cafe Pte Ltd v Ferrero SpA [2013] 1 SLR 531 (“Sarika”) at [14] and The Polo/Lauren Co, LP v Shop In Department Store Pte Ltd [2006] 2 SLR(R) 690 (“Polo (CA)”) at [8]). In Polo (CA), this court observed (at [8]) that there was no infringement under s 27(2)(b) of the Trade Marks Act (Cap 332, 1999 Rev Ed) (“the 1999 TMA”), which is identical to s 27(2)(b) of the Act, if the likelihood of confusion was brought about by some factor other than the similarity between the registered mark and the allegedly infringing sign and the similarity between the goods or services for which the former was registered and those in relation to which the latter was used.

Threshold issue: use in the course of trade

17 But, before assessing Hai Tong’s claim against these requirements, there is a threshold issue. Hai Tong must first show that the Sign was used in the course of trade, ie, as a means of distinguishing the goods or services in question as to their origin (see City Chain Stores (S) Pte Ltd v Louis Vuitton Malletier [2010] 1 SLR 382 (“City Chain”) at [35]–[36]). This is provided for in the opening words of s 27(2) of the Act and may be contrasted with a sign that is used for purely decorative purposes, as was the case in City Chain (at [38]), or purely to describe the product or service in question or some element thereof.

18 We agree with the Judge that the Sign was used in the course of trade. When a sign consists of or incorporates common or generic words that may also describe the product or service in question, it may not be obvious that those words are being used in the course of trade. But here, we agree with the Judge’s observation that the combination of two common words can result in a sign that goes beyond merely describing the product in question and serves instead to distinguish it as to its origin (at [12] of the GD). The Judge found that the words “Rose” and “Lady”, when combined as “Rose Lady”, did indeed acquire such a character (at [13] of the GD). That this was in fact so in this case became even clearer when one of the partners of Sky, in the course of his testimony, conceded that the words “Rose Lady” were used as the name of a particular range of products rather than to describe any of the ingredients in the products (see [13] of the GD). We disagree with the Defendants’ submission that any distinctiveness in the combined use of the two words “Rose” and “Lady” had to be acquired through usage, advertising and promotion.^{[note: 2]} This submission is misplaced because at this threshold stage of the inquiry, the court, in considering the allegedly infringing sign, is only concerned with whether or not that sign is being used to distinguish the goods (or services) of a particular trader and not merely to describe or decorate the goods on which it has been applied. Given the concession referred to above, the contrary was not seriously arguable.

19 The Defendants also argued that the real badge of origin in this case was the registered mark “VOV”. But, there is no reason why the Sign could not function as a badge of origin alongside the “VOV” mark and so be said to be used in the course of trade. The Judge held (at [15] of the GD) that a product might have more than one trade mark, and this is correct. Secondary product identifiers and signs which are used to distinguish between different product lines of the same manufacturer can function as trade marks if they indicate to the consumer that the product in question originates from a particular undertaking (see Richemont International SA v Da Vinci Collections Pte Ltd [2006] 4 SLR(R) 369 at [14] and [39]–[40] as well as [15] of the GD).^{[note: 3]}

20 Moreover, even if the words “Rose Lady” were capable of being descriptive in nature, which we find not to be the case, we are satisfied that the Sign in this instance was in fact used as a badge of origin. In Super Coffeemix Manufacturing Ltd v Unico Trading Pte Ltd and another and another appeal [2000] 2 SLR(R) 214 at [28], in relation to whether the word “Coffeemix” in the defendants’ “INDOCAFE Coffeemix” line of products was used in the course of trade, this court said:

... [W]hether a word which is essentially descriptive, when used on a product, is used in the descriptive sense or in the trade mark sense would depend very much on the way it is presented on the product. We think the intention of the manufacturer is an important factor, though that is not necessarily decisive. Here, having regard to the fact that on the defendants’ product the word “Coffeemix” is presented most prominently, bigger in dimension than any other words on the package, and the admission by the defendants that they intended the style “INDOCAFE Coffeemix” to be used in the trade mark sense, we have to disagree with the learned trial judge that “Coffeemix” is not used by the defendants in the trade mark sense. We find that the word “Coffeemix” is used by the defendants in the trade mark sense.

Here, as the Judge correctly observed, the Sign was “prominently displayed” on the packaging of the allegedly infringing products (see [18] of the GD). Moreover, the inclusion of the words “Rose Lady” on the packaging would not, without more, have been understood as a mere description of the rosa canina ingredient.

21 We therefore agree with the Judge that the Sign was used in the course of trade as a badge of origin and accordingly turn to the substantive elements of a claim under s 27(2)(b) of the Act.

Similarity or identity of the goods or services

22 The Defendants have conceded that the Sign is used on goods that are identical in nature to those for which the Composite Mark is registered, ie, cosmetics. This element is therefore not in issue before us. However, the fact that the Composite Mark and the Sign are being used on identical goods will, as noted below (at [97]–[98]), have a bearing on the overall assessment of the likelihood of confusion. Also, once it has been established (or, as in this case, admitted) that the goods for which the Composite Mark has been registered are identical to those in respect of which the Sign is being used, then there is no longer any occasion to examine whether the two sets of goods are similar. Nor specifically, therefore, is there any occasion to consider those factors that are relevant to establishing the degree of such similarity, such as the trade channels through which the goods are distributed or marketed, or whether they are placed on the same shelves or are separated, or such other factors as were identified in British Sugar Plc v James Robertson & Sons Ltd [1996] RPC 281 (“British Sugar”) at 296–297 – see Susanna H S Leong, Intellectual Property Law of Singapore (Academy Publishing, 2013) (“IP Law of Singapore”) at paras 28.354–28.358, and Ng-Loy Wee Loon, Law of Intellectual Property of Singapore (Sweet & Maxwell, 2009 Rev Ed) at para 21.5.13.

Similarity between the Composite Mark and the Sign

23 The first issue that does arise is whether the Judge was right in finding that the Composite Mark and the Sign are similar. We set out below the Composite Mark and the Sign:

The Composite Mark	The Sign

24 The Composite Mark and the Sign are not identical, but the purposes of the law would not be met by a rule that only prohibits the later use of an identical mark. This would leave the owner of the earlier registered mark with protection that may prove too limited. On the other hand, it would be unfair to subsequent entrants to the market if they are prevented from using ordinary words or devices that may have some similarity to what has already been registered. Where the contesting marks are used on goods of an identical or sufficiently similar nature, the law seeks to strike a balance between these interests having regard to various considerations, one of which is the distinctiveness of the registered mark. This can be traced to the origins of statutory trade mark protection. When the English Parliament finally passed a law in 1875 to allow the registration of trade marks, it provided that only “special and distinctive” words could be registered as trade marks. Where the words were invented or had no direct reference to the quality or character of the goods in question, an action for infringement would lie and there would be no need to prove actual confusion on the part of customers or a fraudulent intent on the part of the defendant. In effect, the trade mark owner was recognised as having a property right in his mark. Where, however, ordinary descriptive words were used, an action would only lie in passing off and only to the extent that the plaintiff could show that the words had acquired distinctiveness through use and also prove deception – see Daniel M McClure, “Trademarks and Unfair Competition: A Critical History of Legal Thought” (1979) 69 Trademark Rep 305 at pp 313–314.

25 Of course, the law has developed since then, but the lack of any distinctive character remains an absolute ground for refusing to register a mark under ss 7(1)(b)–7(1)(d) of the Act unless the applicant can come within s 7(2) by showing that before the date of the application for registration, the mark has “in fact acquired a distinctive character as a result of the use made of it”. In relation to registered marks, the question of distinctiveness remains a relevant inquiry in assessing similarity between the registered mark and the allegedly infringing sign when considering a claim for trade mark infringement. In Sarika, this court noted (at [20]):

... [W]e would explain that the “distinctiveness” of the registered trade mark is a factor to be considered in the visual, aural and conceptual analysis to determine whether the allegedly infringing sign and the trade mark are similar. It stands to reason that the more distinctive the registered trade mark, the more it is necessary to show sufficient alterations to, or difference in the sign in order that it may not be held to be similar to the trade mark ...

26 Distinctiveness is considered within the assessment of similarity, as noted by this court in Sarika. However, for the purpose of elucidating the analytical process, we have highlighted it here as a separate step. Once that has been considered, the next consideration is the overall assessment of whether the registered mark and the allegedly infringing sign are similar having regard also to their distinctiveness; and if they are similar, whether such similarity, coupled with the identical nature of the goods in question here, is likely to cause confusion in all the circumstances. We address each of these elements below.

Distinctiveness

27 In the inquiry into distinctiveness, the focus is not on whether the defendant’s allegedly infringing sign is being used in the course of trade (see above at [17]–[20]). Rather, having crossed that threshold, it becomes relevant to examine the distinctiveness of the plaintiff’s registered mark in order to determine the extent of the latitude that will be allowed to a user of features that appear in that mark. But, what is the essence of distinctiveness? It is helpful to begin with the definition of a trade mark as provided for in s 2(1) of the Act:

“trade mark” means any sign capable of being represented graphically and which is capable of distinguishing goods or services dealt with or provided in the course of trade by a person from goods or services so dealt with or provided by any other person …

28 For present purposes, what is relevant is that a trade mark must be capable of distinguishing the origin of the goods or services in question. It has been observed in Nation Fittings (M) Sdn Bhd v Oystertec plc and another suit [2006] 1 SLR(R) 712 (“Nation Fittings”) at [136] that this in fact prescribes an initial, albeit low, threshold that all aspiring trade marks must clear. Distinctiveness thus connotes the sense that the trade mark is capable of saying something meaningful not about the goods or services in question, but rather, about whence these have come.

29 Where a trader chooses a word or mark that is purely inventive or fancy in the sense that it is made up and has no meaning at all, or that is arbitrary in the sense that it has nothing relevant to say about the goods in question themselves, then from the perspective of trade mark law, it will be regarded as distinctive. As we recently observed in The Singapore Professional Golfers’ Association v Chen Eng Waye and others [2013] SGCA 18 (“Chen Eng Waye”) at [29]:

The use of a name or a choice of words that is either meaningless or has no discernible correlation to the product or service in question or its key elements – in short, a fancy word – generally serves no purpose other than as a mark or badge of the origin or source of that product or service. … [emphasis added]

30 The position potentially gets more complicated where what is at issue are words or concepts that are in common use. Here, it will be necessary to consider whether the words mean something, and if they do, whether they serve in truth to describe the goods or services in question rather than as a badge of origin. The difficulty is compounded because traders registering a mark are often tempted to use words or symbols that describe their goods in glowing terms; but, if the words do no more than describe the goods, they may be found not to be distinctive at all. There is an additional point, in that the more fancy or inventive or arbitrary a mark, then in general, the greater the protection it will receive, in the sense that the defendant would have to demonstrate to a more compelling degree that his mark or sign is indeed dissimilar from the registered mark (see Sarika at [20], Polo (CA) at [10] and Ozone Community Corp v Advance Magazine Publishers Inc [2010] 2 SLR 459 (“Ozone Community”) at [44]). On the other hand, the more descriptive a mark, the greater the latitude afforded to the defendant in using words or ideas of a generic nature, even if these have some similarity to what is found in the registered mark. This follows as a matter of common sense from the fact that in order to balance the interests of honest traders who enter the market later, the rule that proscribes impermissible similarity cannot be applied in precisely the same manner to a mark that uses common descriptive words or devices as to one that does not. Otherwise, the law would prevent the legitimate use of an unacceptably large spectrum of common words and ideas. We therefore begin by examining the distinctiveness of the Composite Mark; then in that light, we turn to consider the similarity of the Composite Mark and the Sign (see Intuition Publishing Ltd v Intuition Consulting Pte Ltd [2012] SGHC 149 (“Intuition Publishing”) at [41]).

Distinctiveness of the Composite Mark

31 Some marks are inherently distinctive because they consist of inventive words without any meaning at all (eg, “Volvo” (see Polo (CA) at [23]) or “Nutella” (see Sarika at [36])). But, inherent distinctiveness can arise either because the words used in a mark are completely inventive and meaningless, or because even though they may have a meaning, they are used in a way that is completely arbitrary in relation to the goods or services in question. When used in this way, the words comprising the mark have no notional or allusive quality at all and say nothing about the goods or services in relation to which the mark is used.

32 Even a mark that is not inherently or obviously distinctive may become so by use. As to whether this is so, the matter may be assessed at the time of an application for registration of a trade mark pursuant to s 7(2) of the Act; or subsequently in an application for revocation under s 23 of the Act, where, if the application is founded on the assertion that the trade mark in question lacks distinctiveness on one or more of the grounds set out in s 7(1)(b), s 7(1)(c) or s 7(1)(d) of the Act, the application will fail if “in consequence of the use which has been made of [the trade mark], it has after registration acquired a distinctive character in relation to the goods or services for which it is registered” (see s 23(2) of the Act).

33 In this regard, it may be helpful to have regard to the following factors that were considered by the High Court in Doctor’s Associates Inc v Lim Eng Wah (trading as SUBWAY NICHE) [2012] 3 SLR 193 (“Subway Niche”) in determining whether a mark had acquired distinctiveness through use (see Subway Niche at [21]):

(a) the market share held by the registered mark;

(b) the nature of its use (whether or not it was intensive, widespread and long-standing);

(d) the proportion of the relevant class of individuals who identified goods sold under the mark as emanating from a particular source; and

(e) any statements from trade associations and chambers.

34 These are helpful though by no means exhaustive indicia in assessing whether or not the relevant segment of the public is likely to view common words or ideas used in a mark as a badge of origin rather than merely as a description of the goods in question or some element or feature thereof. The Judge, in discussing Hai Tong’s passing off claim, made certain findings and observations which formed the basis of his finding that Hai Tong had only “trivial” goodwill (see [39] of the GD). First, the Judge found that Hai Tong had not conducted any market survey which indicated its market share in the cosmetics industry. Secondly, the Judge noted that the annual sales figures provided by Hai Tong in relation to its “Lady Rose” products did not show how much of these sales related to its cosmetic products as opposed to its other non-cosmetic products, and even then, the Judge thought these sales figures “were not particularly impressive” (see [36] of the GD). Thirdly, the Judge noted that Hai Tong had not advertised its products since 1979, and that although Hai Tong claimed that it had an ongoing incentive promotion programme of providing free goods to retailers, no evidence of actual advertising and promotion to the public had been adduced. The Judge also noted that Hai Tong had not adduced evidence that the benefits of the alleged incentive promotion programme that it ran were in fact extended by its retailers to the public. Although the Judge’s observations were made in the context of his consideration of whether Hai Tong had the requisite goodwill to sustain its claim in the tort of passing off, the Defendants relied in CA 2 of 2012 on these findings to contend that apart from the fact that the Composite Mark had been used in trade for more than 40 years, there was little else to suggest that it had achieved a significant degree of distinctiveness through use.

35 In our judgment, this contention was ultimately misplaced. Although no application was brought for the revocation of the Composite Mark, the Defendants did bring a counterclaim founded on their contention that Hai Tong’s threats of trade mark infringement proceedings were groundless, and under s 35(4) of the Act, such a counterclaim would be well founded if the Defendants could show that the registration of the Composite Mark was liable to be revoked. However, the attempt to contend that this should be the case because of any lack of distinctiveness in the Composite Mark was unfounded. Of course, the words “Lady” and “Rose”, when used separately, are not inventive words or phrases. However, in our judgment, it would not be correct to analyse this by reference to each of the words standing alone. It is the juxtaposition of the words that engenders distinctiveness, in that when used together, the words do not convey a sensible meaning. One does not speak of a “lady” rose. Rather, the juxtaposed words convey the sense of a name or an appellation rather than having any meaning, and in that sense, the juxtaposed words are inventive and certainly capable of distinguishing the goods dealt with or provided in the course of trade by one person (in this case, Hai Tong) from those so dealt with or provided by others. The juxtaposed words were therefore not devoid of distinctive character; and in that light, it really did not fall for consideration whether these words had also or separately acquired a measure of distinctiveness through use.

36 The Defendants’ main argument rested on: (a) a consideration of the two words “Lady” and “Rose” separately, which is mistaken; and (b) their view that the words were not unusual or memorable for that matter. But, this too is incorrect. Distinctiveness can be found if words are connected and/or used in a meaningless way; and further, even assuming the words in question have a meaning, if, as is the case here, they are used in conjunction with goods in respect of which the words say nothing relevant. In this connection, in British Sugar, Jacob J observed as follows (at 306):

... What does devoid of any distinctive character mean? I think the phrase requires consideration of the mark on its own, assuming no use. Is it the sort of word (or other sign) which cannot do the job of distinguishing without first educating the public that it is a trade mark? A meaningless word or a word inappropriate for the goods concerned (“North Pole” for bananas) can clearly do. But a common laudatory word such as “Treat” is, absent use and recognition as a trade mark, in itself ... devoid of any distinctive inherently [sic] character. … [emphasis in original in italics; emphasis added in bold italics]

37 Regard may also be had to Sime Darby Edible Products Ltd v Ngo Chew Hong Edible Oil Pte Ltd [2000] 2 SLR(R) 604 (“Sime Darby”), where the court was concerned with an application to invalidate a trade mark consisting of two words “ROYAL SPOON” on the ground that this was devoid of distinctive character. The careful analysis of Lim Teong Qwee JC at [12]–[15] bears setting out at some length:

12 I come now to the Royal Spoon trade mark. Mr Gill [counsel for the applicant] submitted that the word “royal” was well-known and commonly used as a laudatory term to convey to the ordinary man’s mind the sense of pre-eminence or of superior quality. I think so. In relation to edible oils or vegetable ghee, I think it is a descriptive word which may serve to designate the quality of the goods. A trade mark which consisted exclusively of this word would not be registrable without proof of acquisition of a distinctive character as a result of use made of it. ...

13 Mr Gill said that the representation of a spoon or spoons in the packaging of edible oils and fats was very common in the industry as admitted by Ngo Chew Hong [the proprietor of the “ROYAL SPOON” trade mark]. ...

14 Is Spoon distinctive without having acquired a distinctive character as a result of use? The word denotes primarily a common utensil for domestic use but I think it is also a word of a descriptive nature in relation to items which are usually dispensed using such a utensil. It is a word which might naturally be used to describe the physical nature or characteristic of edible oils used in the kitchen. An illustration of a spoon dispensing a fluid or semi-solid substance might not mean much to anyone but once he sees that as a trade mark in relation to edible oils he would recognise the spoon as a utensil that is apt to describe the goods of a kind rather than those of a particular manufacturer. I agree that Spoon by itself is devoid of any distinctive character in relation to the goods in this case.

15 Ngo Chew Hong’s trade mark consists not of Royal or Spoon separately but of the two words used in combination. Royal is a laudatory word but by linking it with an object such as Spoon the “praise” is wholly directed towards it. It describes Spoon. It does not describe the goods. Any sense of pre-eminence or of superior quality conveyed to the ordinary man’s mind (I take this from Mr Gill’s argument) would be in relation to Spoon and not to the goods. I think it would be understood more in the sense of being kingly or majestic but it would in any sense in which the word is understood describe or attribute some quality to Spoon and not the goods. Again Spoon by itself may be associated with the goods but Royal Spoon in the sense the expression is likely to be understood by persons using the English language would be a wholly inappropriate phrase for edible oils or vegetable ghee (not unlike North Pole for bananas). In my judgment Royal Spoon is a trade mark which is not devoid of any distinctive character and if it had been registered under the new law [ie, the Trade Marks Act 1998 (Act 46 of 1998)] its registration would not have been in breach of s 7(1)(b).

[emphasis added]

38 To similar effect are the observations of the Federal Court of Australia in R & C Products Pty Limited v Hunters Products Pty Limited (1988) ATPR ¶40-839 (“R & C Products”) (at 49,054) when commenting on the brand name “blue & fresh”:

... The words “blue and fresh” are fancy words and ungrammatical as the name of a product ... They are quite unlike essentially descriptive words such as “Office Cleaning Services” ...

See also our observations in Chen Eng Waye at [29] and the passage that is cited there from Christopher Wadlow, The Law of Passing-Off: Unfair Competition by Misrepresentation (Sweet & Maxwell, 4th Ed, 2011) at para 8-075. It is thus evident that the Composite Mark is distinctive.

Similarity – general principles

39 In this light, we turn to consider the similarity between the Composite Mark and the Sign. In doing so, an assessment of their visual, aural and conceptual similarities is undertaken (see, eg, Sarika at [16]).

40 Further, the following principles are well established in Singapore trade mark jurisprudence (see generally Sarika at [16]–[18] and Subway Niche at [19] and the cases cited below):

(a) The thrust of the inquiry into similarity is directed at assessing substantive similarity. The three aspects of this evaluation, namely, visual, aural and conceptual similarities, aid the court’s evaluation by signposting its inquiry. They do not serve as a mechanistic formula of any sort. It follows that the law does not require all three similarities (visual, aural and conceptual) to be made out before the registered mark and the allegedly infringing mark may be found to be similar. Conversely, the fact that any one similarity is made out does not necessarily mandate a finding that the two marks are similar. Nor is each aspect of similarity of equal importance. Rather, the relative importance of each will depend on the circumstances, including the nature of the goods and the types of marks involved (see MediaCorp News Pte Ltd v Astro All Asia Networks plc [2009] 4 SLR(R) 496 (“MediaCorp”) at [32]), and a trade-off can be made between the three aspects of similarity (see Ozone Community at [40]).

(b) In assessing the similarity between two contesting marks, the court considers them “as a whole” (see City Chain at [47] and [50]), but does not take into account “any external added matter or circumstances” because the comparison is “mark for mark” (see MediaCorp at [33], citing Caterpillar Inc v Ong Eng Peng (formerly trading as Catplus International) [2006] 2 SLR(R) 669 (“Caterpillar”) at [55]). This is unlike the position when assessing whether there is a likelihood of confusion, where, subject to the further observations we have made below, it will generally be appropriate to have regard to other relevant circumstances – see below at [79]–[98].

(c) Although the focus of this part of the inquiry is on the marks in question, it is necessary to set out the viewpoint the court should assume. This viewpoint is that of the average consumer who would exercise some care and a measure of good sense in making his or her purchases, not that of an unthinking person in a hurry (see Polo (CA) at [34]). Would such a person consider the marks similar? We pause to note that despite the differing expressions that are sometimes used, this viewpoint is, for all intents and purposes, the same as that which is assumed when assessing the separate but related question of whether there is a sufficient likelihood of confusion. This is unsurprising given that there will be some inevitable overlap in the factual assessment of both elements.

(d) Furthermore, it is assumed that the average consumer has “imperfect recollection” (see MediaCorp at [33], citing Nautical Concept Pte Ltd v Jeffery Mark Richard and another [2007] 1 SLR(R) 1071 (“Nautical Concept”) at [30]). As such, the two contesting marks are not to be compared or assessed side by side and examined in detail for the sake of isolating particular points of difference. Instead, the court will consider the general impression that will likely be left by the essential or dominant features of the marks on the average consumer (see Saville Perfumery Ld v June Perfect Ld and F W Woolworth & Co Ld (1941) 58 RPC 147 (“June Perfect”) at 161–162).

Visual similarity

41 In dealing with the test for assessing the visual similarity of marks where one or more of the marks is a composite mark, there appears to have been some emphasis on the textual component of the composite mark. In so far as it has been advanced as a proposition of law that in general, words “talk” in composite marks, we do not agree. Our reasons shall become apparent upon our analysis of the legal position in Singapore, the UK, the European Union and Australia which follows. Rather, we consider that in assessing the visual similarity of two composite marks, the correct approach is to consider the marks in totality without placing undue emphasis on any particular component of such marks unless such emphasis is warranted on the facts.

(1) The position in Singapore

42 The genesis of the idea that words “talk” in composite marks appears to be a passage from the decision of the Registry of Trade Marks and Patents in Lee Cooper Group plc v Levi Strauss & Co [1995] AIPR 457 (“Lee Cooper”), which was then applied by the High Court in Chai Chyau Ling (doing business as Racetech Auto) v Racing Technology Pte Ltd [2009] SGHC 105 (“Racetech Auto”). In Racetech Auto, the respondent was the registered proprietor in Singapore of the trade mark “RACETECH”, which was of a unique typeface and registered without limitation of colour. The appellant’s logo (“the Logo”) comprised the words “RACE TECH” in a typeface which, at least to an untrained eye, was similar to that of the respondent’s word mark. An action was brought by the respondent against the appellant for trade mark infringement under ss 27(1) and 27(2) of the Act and for passing off. An assistant registrar awarded summary judgment to the respondent on both claims under O 14 of the Rules of Court (Cap 322, R 5, 2006 Rev Ed).

43 On appeal to the High Court, the court agreed with the assistant registrar and held that “it is most likely the word and not so much the device that leaves an impression on the minds of people and persuades them to buy from the trader owning that word mark” – see Racetech Auto at [37], relying (at [36]) on the following passage in Lee Cooper (at 501):

Words ‘talk’ in trade marks. Thus, if a mark consists of a device in combination with a distinctive word (ie a word of a trade mark nature), it is most likely that it is by the word that the public will normally order goods under the mark. Therefore, where two marks resemble each other to the extent that they each contain a similar device, the overall resemblance between them will generally be lessened if they bear words of a trade mark nature which are entirely different from each other. … [emphasis added by the court in Racetech Auto]

44 We make two observations on the court’s reliance on this passage from Lee Cooper. First, such reliance was unnecessary in Racetech Auto as the court had already found that the respondent’s mark and the Logo were identical under s 27(1) of the Act (see Racetech Auto at [28]); hence, the question whether they were also similar simply did not arise. Secondly, the passage in Lee Cooper was cited out of context, as an examination of the decision in that case will show. In Lee Cooper, the applicant applied to register the following composite mark in Class 25 of the Third Schedule to the Trade Marks Rules 1991 (S 86/1991)^{[note: 4]} (“Class 25”) for various clothing items such as shirts, trousers and jeans:

45 A notice of opposition was lodged by the opponent, Levi Strauss, on the ground that it had, in 1873, adopted as a trade mark a particular pattern of double lines of stitching (the Arcuate Stitching Design) which was applied to the back pockets of “Levi’s” pants and which was intended by Levi Strauss to distinguish the pants it produced from those produced by its competitors. Levi Strauss was also the registered proprietor in Singapore of the following trade marks in Class 25:

46 Before the Assistant Registrar for Trade Marks (“TMAR”), Levi Strauss argued that if registration and use of the applicant’s composite mark were allowed, the trade, or the public for that matter, would likely believe that the goods of the applicant originated from the same source as Levi Strauss’s goods, or, at least, that there was some connection in the course of trade between the applicant and Levi Strauss when in fact, there was no such connection at all. Levi Strauss argued that this would be in contravention of s 15(1) of the Trade Marks Act (Cap 332, 1992 Rev Ed) (“the 1992 TMA”). That section provided as follows:

Restriction on registration

15.—(1) It shall not be lawful to register as a trade mark or part of a trade mark any matter the use of which would, by reason of its being likely to deceive or cause confusion or otherwise, be disentitled to protection in a court of justice, or would be contrary to law or morality, or any scandalous design.

47 The TMAR applied the test formulated by the House of Lords in “Bali” Trade Mark [1969] RPC 472 in respect of s 15 of the 1992 TMA (see Lee Cooper at 499):

Having regard to the user of [Levi Strauss’s] mark, is the Registrar satisfied that the mark applied for, if used in a normal and fair manner in connection with any goods covered by the registrations proposed, will not be reasonably likely to cause deception and confusion amongst a substantial number of persons?

48 The TMAR then observed that Levi Strauss’s Arcuate Stitching Design was always used on the pockets of its pants in juxtaposition with the “Levi’s” label and/or the word “Levi’s” on a red tag. Therefore, under s 15 of the 1992 TMA, the TMAR found that she had to compare the “Levi’s” word mark and pocket arcuate design on the one hand, and the “Lee Cooper” word mark and pocket arcuate design shown on the applicant’s composite mark on the other. It was in this context that the TMAR observed (at 501):

Words ‘talk’ in trade marks, Thus, if a mark consists of a device in combination with a distinctive word (ie a word of a trade mark nature), it is most likely that it is by the word that the public will normally order goods under the mark. Therefore, where two marks resemble each other to the extent that they each contain a similar device, the overall resemblance between them will generally be lessened if they bear words of a trade mark nature which are entirely different from each other. (See UK Work Manual (Cap 10) ‘Searching and Comparison of Trade Marks’).

In Broadhead’s Application (1950) 67 RPC 209 at p 215 Evershed MR followed the observations of Lord Russell in Coca Cola Co (Canada) v Pepsi Cola Co (Canada) (1942) 59 RPC 127 where he stated:

Where you get a common denominator, you must in looking at the competing formulae pay much more regard to the parts of the formulae which are not common – although it does not flow from that that you must treat the words as though the common part was not there at all.

In Harrods’ Application (1935) 52 RPC 65 at 70, the Registrar stated:

It is a well-recognized principle that has to be taken into account in considering the possibility of confusion arising between any two trade marks, that, where those two marks contain a common element which is also contained in a number of other marks in use in the same market, such a common occurrence in the market tends to cause purchasers to pay more attention to the other features of the respective marks and to distinguish between them by those features. This principle clearly requires that the marks comprising the common element shall be in fairly extensive use and, as I have mentioned, in use in the markets in which the marks under consideration are being or will be used.

[emphasis added]

49 In our judgment, the real point the TMAR was making was that the textual components in that case stood out, or “talked”, because they were the dominant components of the marks in question. In that case, similar devices were deployed in both marks and, for that matter, were commonly used in the clothing industry, but generally in conjunction with distinct word marks. In our judgment, the TMAR’s focus on the textual components of the marks in Lee Cooper was not meant to fashion a general principle of trade mark law that the textual component of a composite mark is always and necessarily the dominant element. Rather, the textual components of the two marks in question in that case were in fact the dominant components of the respective marks because of the similarity of the arcuate stitching devices in both marks.

50 Nor is this court’s recent decision in Sarika an endorsement of a general proposition that words “talk” in composite trade marks. In Sarika at [71], in the context of discussing whether a claim under s 55(2) of the Act was made out on the facts, we said:

... [T]he Appellant highlights the fact that some of the [Respondent’s] “Nutella” marks are depicted together with other devices, such as a slice of bread, a knife, hazelnuts and milk glass (see [3] above), whereas the [Appellant’s] “Nutello” mark does not have any such depiction. However, as rightly pointed out by the Respondent, consumers would usually refer to the word or words in a trade mark and importantly, the distinctiveness of the “Nutella” word would have an impact on consumers. In this regard, it was observed in Racetech Auto … at [37] that where a distinctive word is used together with a device in a trade mark, it is very likely the word that would leave an impression on consumers’ minds. While we acknowledge that the applicability of this proposition depends very much on the circumstances of each case, we find this observation is appropriate here. A consumer of the Nutella spread would typically refer to it by name and give relatively less emphasis to the device printed along with it. [emphasis added]

51 Hai Tong relied on the words emphasised in the passage cited above and submitted that this stood for the general proposition that words “talk” in composite trade marks. But, this ignores the parts of the passage which have not been emphasised and which make it clear that all the court was saying was that the textual component of the composite mark in that case stood out on the specific facts in play there.

52 We note that in certain circumstances, words can indeed “talk” in trade marks even if they do not form the dominant component of a composite mark. One such instance is if the textual component is in itself already widely known (see Festina Lotus SA v Romanson Co Ltd [2010] 4 SLR 552 (“Festina”) at [31]–[34], citing with approval the decision of the European Court of Justice (“the ECJ”) in Medion AG v Thomson Multimedia Sales Germany & Austria GmbH (Case C-120/04) [2005] ECR I-8551 (“Medion”)). However, since Hai Tong has not submitted that the textual component of the Composite Mark at issue in this case has acquired such prominence, this is not directly relevant.

53 Another instance where the textual component of a composite mark would generally be the dominant component is where the goods or services bearing the mark are marketed and sold predominantly through online trade channels (see Festina at [55] and Intuition Publishing at [64]–[65]). This is because a customer who wishes to buy the goods or procure the services in question from the online store of the trade mark owner would have to enter the textual component of the mark into the address bar or search engine to get to the desired website. Thus, such a customer would need to recall the textual component as opposed to the device component of the composite mark. This might render the visual impact of the device component less significant as compared to the aural and visual impact of the textual component, even where the device component happens to be significantly larger than the textual component. But, this is not relevant in the present appeal as the parties’ goods are not predominantly marketed and sold through online channels.

(2) The position in the UK and the European Union

54 The law on assessing whether there is “confusing similarity” arising from the similarity of two marks or signs and the similarity of the goods on which these marks or signs are applied, under Art 9(1)(b) of the European Council Regulation 207/2009/EC of 26 February 2009 on the Community Trade Mark or Art 5(1)(b) of the First Council Directive 89/104/EEC of 21 December 1988 to Approximate the Laws of the Member States relating to Trade Marks, has been summarised recently by the English Court of Appeal in Specsavers International Healthcare Ltd v Asda Stores Ltd [2012] FSR 19 (“Specsavers”). The court in Specsavers considered a number of important European decisions, including Sabel BV v Puma AG, Rudolf Dassler Sport (Case C-251/95) [1998] RPC 199 (“Sabel”), Canon Kabushiki Kaisha v Metro-Goldwyn-Mayer Inc (formerly Pathé Communications Corporation) (Case C-39/97) [1999] RPC 117 (“Canon”) and Office for Harmonisation in the Internal Market (Trade Marks and Designs) (OHIM) v Shaker di L Laudato & C Sas (Case C-334/05) [2007] ECR I-4529, and said (see Specsavers at [52]):

…

(d) the visual, aural and conceptual similarities of the marks must normally be assessed by reference to the overall impressions created by the marks bearing in mind their distinctive and dominant components, but it is only when all other components of a complex mark are negligible that it is permissible to make the comparison solely on the basis of the dominant elements;

(e) nevertheless, the overall impression conveyed to the public by a composite trade mark may, in certain circumstances, be dominated by one or more of its components;

…

This suggests that there is no general legal proposition that any one component of a composite mark dominates.

55 We are mindful that some European decisions pre-dating Specsavers have particularly emphasised the textual (or verbal) component of a composite mark when it comes to retail goods. For example, in Honda Motor Europe Ltd v Office for Harmonisation in the Internal Market (Trade Marks and Designs) (Case T-363/06) [2009] ETMR 34 (“Honda”), the Court of First Instance of the European Communities (“the CFI”) observed as follows (at [30]):

... [I]t should be borne in mind that, according to settled case law, where a trade mark is composed of verbal and figurative elements, the former are, in principle, more distinctive than the latter, because the average consumer will more readily refer to the goods in question by quoting their name than by describing the figurative element of the trade mark (see, to that effect, Oberhauser v Office for Harmonisation in the Internal Market (Trade Marks and Designs) (OHIM) (Petit Liberto – Fifties) (T-104/01) [2002] E.C.R. II-4359 at [47], and Wassen International Ltd v Office for Harmonisation in the Internal Market (Trade Marks and Designs) (OHIM) (Stroschein Gesundkost – SELENIUM-ACE) [2005] E.C.R. II-2897 at [37]).

56 The same views are echoed in the decisions of the CFI in cases such as Claudia Oberhauser v Office for Harmonisation in the Internal Market (Trade Marks and Designs) (OHIM) (Case T-104/01) [2002] ECR II-4359 (“Oberhauser”) at [47] and Wassen International Ltd v Office for Harmonisation in the Internal Market (Trade Marks and Designs) (OHIM) (Case T-312/03) [2005] ECR II-2897 (“Wassen”) at [37], and in the decisions of the European General Court in Oyster Cosmetics SpA v Office for Harmonisation in the Internal Market (Trade Marks and Designs) (OHIM) (Case T-437/09) (2 February 2011) (“Oyster Cosmetics”) at [36] and Trubion Pharmaceuticals, Inc v Office for Harmonisation in the Internal Market (Trade Marks and Designs) (OHIM) (Case T-412/08) [2009] ECR II-239 (“Trubion”) at [45].

57 However, an examination of these decisions will reveal that these views should be understood in the specific context of considering and assessing the marks in question, rather than as reflecting some sort of a “general” principle. Thus:

(a) In Honda, the textual component of the composite mark

(a) was found to be the dominant component as the “S” device component “[did] not appear to be original or very elaborate” and was the first letter of the textual component “SEAT” (see Honda at [31]).

(b) In Oberhauser, the court observed that the average consumer looking at a jeans label would take in and remember the predominant textual component of the composite mark

(b) rather than the various figurative aspects of that mark because the latter aspects would be seen as “decorative features of a label that is commonplace for jeans in its form and components, and not as the most important element indicating the origin of the product” (see Oberhauser at [47]). In short, in this particular context, a complex device was seen to be decorative and not distinctive. But, that was factually the case with the composite mark in that case.

(c) dominant because the figurative element was placed below the verbal element, ie, in a “less visible position” (see Wassen at [37]). It may also be noted that the figurative element appeared essentially to be decorative.

(d) In Oyster Cosmetics, the court found the word “oystra” in the composite mark

(d) dominant because it was in the centre of the mark, and was written in black characters that were in bold and larger than those used for the other words in the mark. In contrast, the word “Kadus” was notable for its small size and was positioned at a clear distance from the other elements of the mark, such that it must be regarded by the relevant segment of the public as negligible compared with the word “oystra”. As for the three dots, these played a negligible role in the overall visual impression produced by the mark because of their smaller size and lack of inherent distinctiveness (see Oyster Cosmetics at [29]–[33]).

(e) In Trubion, the court found that the device component in the composite mark

(e) even if not negligible in terms of its size, was likely to be perceived by consumers essentially as a decorative element, and not as an element indicating the commercial origin of the goods (see Trubion at [45]).

58 The CFI in Quelle AG v Office for Harmonisation in the Internal Market (Case T-88/05) [2007] ETMR 62 (“Quelle AG”) made some observations which, although situated in a slightly different context, are rather more consistent with our view, namely, that the answer to the question whether words do “talk” in a composite trade mark depends in the end on whether the words are the dominant element in the particular case in question. We note in particular the CFI’s observations at [69]:

69 … [T]he degree of phonetic similarity between two marks is of less importance in the case of goods which are marketed in such a way that, when making a purchase, the relevant public usually perceives visually the mark designating those goods ... That is the case with respect to the goods at issue here. Although the applicant states that it is a mail order company, it does not submit that its goods are sold outside normal distribution channels for clothing and shoes (shops) or without a visual assessment of them by the relevant consumer. Moreover, while oral communication in respect of the product and the trade mark is not excluded, the choice of an item of clothing or a pair of shoes is generally made visually. Therefore, the visual perception of the marks in question will generally take place prior to purchase. Accordingly, the visual aspect plays a greater role in the global assessment of the likelihood of confusion ... The fact that those products may, in some circumstances, be the subject of discussion between consumers is therefore irrelevant, since, at the time of purchase, the goods in question and, therefore, the marks which are affixed to them are visually perceived by consumers. [emphasis added]

59 In Quelle AG, the composite mark in question was:

The CFI, in finding that the device component was the dominant component in the composite mark, noted (at [60]) that it was five times larger than the textual component, which appeared only as a caption, and even then, it seemingly referred to the device component.

(3) The position in Australia

60 In relation to the position in Australia, Hai Tong recognised that the courts there have cautioned against excessive emphasis on the textual component of a composite mark because this might result in the rights conferred on a composite mark being extended to include rights conferred on a pure word mark without an accompanying device. In Crazy Ron’s Communications Pty Limited v Mobileworld Communications Pty Limited [2004] FCAFC 196 (“Crazy Ron’s”), the Federal Court of Australia was concerned with an earlier composite mark with the words “Crazy John”

and the later allegedly infringing word marks “Crazy Ron’s” and “Crazy Ron”. In assessing the visual similarity of the earlier composite mark and the later word marks, the court reinforced the need to factually examine the marks in question (at [93]–[94]) and cautioned (at [100]):

... [A]lthough everything depends on the particular circumstances of the case, some caution needs to be exercised before characterising words in a complex composite registered trade mark as an ‘essential feature’ of that mark in assessing the question of deceptive similarity. If such a characterisation is made too readily, it effectively converts a composite mark into something quite different. … [emphasis added]

61 On the facts of Crazy Ron’s, the court found that the textual component in the earlier mark formed only a part of a composite mark that contained a number of elements. The words themselves were not especially prominent and, indeed, were subsidiary to the fantasy character, which occupied the dominant position in the overall image. If anything, the words seemed to identify the fantasy character, having regard to the deranged look on his face. The words also competed for attention with the letters “MW”, which themselves occupied a central position on the stylised globe (at [97]). The court thus concluded that there was no deceptive similarity between the earlier composite mark and the allegedly infringing marks because “[t]he additional features of the [earlier composite mark] are significant points of dissimilarity from the allegedly infringing mark[s] and need to be given due weight in determining the nature of a consumer’s ‘imperfect recollection’ of the [earlier composite mark]” (at [102]). The court indicated, however, that it would have been prepared to come to a different conclusion if there had been a finding, for example, that the words “Crazy John” had been “notoriously so ubiquitous and of such long standing that consumers must be taken to be familiar with it and its use in relation to particular goods” (at [99]), but there was no such finding, and in any case, the court had not been asked to make one (at [99]).

(4) Summary of the relevant principles in relation to the visual similarity of composite marks

62 In our judgment, and having surveyed the foregoing authorities, Racetech Auto and Lee Cooper should not be taken as standing for a general proposition of law that words “talk” in composite trade marks (see [42]–[51] above). We consider that the following non-exhaustive list of principles may be called in aid when assessing the visual similarity of composite marks :

(a) As we have noted above at [40(c)] and [40(d)], when assessing two contesting marks or signs, the court does so with the “imperfect recollection” of the average consumer (see MediaCorp at [33], citing Nautical Concept at [30]). The two marks or signs should not be compared side by side or examined in detail because “the person who is confused often makes comparison from memory removed in time and space from the marks” (see MediaCorp at [33], citing Caterpillar at [55]).

(b) The visual similarity of two contesting marks or signs must normally be assessed by reference to the overall impressions created by the marks or signs, bearing in mind their distinctive and dominant components. When the other components of a complex mark or sign are of negligible significance, it is permissible to make the comparison solely on the basis of any dominant element(s).

(c) The overall impression conveyed to the public by a composite trade mark may, in certain circumstances, be dominated by one or more of its components (see Specsavers at [52(e)]).

(d) The textual component of a composite mark or sign could (but will not necessarily or always) be the dominant component of the mark or sign. Some instances where this might be the case include where:

(i) The two marks or signs in question each contain a similar device. The overall resemblance between them may then be diminished if they bear words which are entirely different from each other (see Lee Cooper at 501).

(ii) The textual component is large, and is in a prominent location in relation to the other components or stands out from the background of the mark or sign (see the decision of the CFI in Shaker di L Laudato & C Sas v Office for Harmonisation in the Internal Market (Trade Marks and Designs) (Case T-7/04) [2009] ETMR 16 (“Shaker (CFI)”) at [41]–[43]).

(iii) The textual component is in itself already widely known (see Festina at [33], Medion at [34] and Crazy Ron’s at [99]).

(iv) The composite mark or sign is applied to goods or services marketed or sold primarily through online trade channels (see Festina at [55] and Intuition Publishing at [64]–[65]).

(e) The device component has been found to be an equally significant, if not the dominant, component of a composite mark or sign where:

(i) the device is significant and large (see, eg, the decision of the European General Court in New Yorker SHK Jeans GmbH & Co KG v Office for Harmonisation in the Internal Market (Trade Marks and Designs) (OHIM) (Case T-415/09) (29 September 2011));

(ii) the accompanying word(s) are devoid of any distinctive character, or are purely descriptive of the device component (see Quelle AG at [60]; see also Sime Darby at [18] and [20]–[21]) or of similar goods of a superior quality (see the decision of the CFI in Saiwa SpA v Office for Harmonisation in the Internal Market (Trade Marks and Designs) (OHIM) (Case T-344/03) [2006] ECR II-1097 at [41] and [45]); or

(iii) the device component is of a complicated nature (see, eg, Waseem Ghias t/as Griller v Mohammed Ikram t/as The Griller Original, Esmail Adia t/as Griller King, Shahzad Ahmad t/as Griller Hut, Griller Original Limited, Griller Hut Limited [2012] EWPCC 3).

(a) But usually not where:

(iv) the device is simple and will not evoke any particular concept for the average consumer (see, eg, the decision of the European General Court in Kavaklidere-Europe v Office for Harmonisation in the Internal Market (Trade Marks and Designs) (OHIM) and Yakult Honsha Kabushiki Kaisha (Case T-276/09) [2012] ETMR 45);

(v) the device component does not attract the attention of the average consumer of the goods in question because such a consumer is regularly confronted with similar images in relation to those goods (see Shaker (CFI) at [42]); or

(vi) the device component is more likely to be perceived as a decorative element rather than as an element indicating commercial origin (see Trubion at [45] and Oberhauser at [47]; see also Wassen).

(5) Application to the facts

63 We turn to the Judge’s decision as regards the visual similarity of the Composite Mark and the Sign in this case. We agree with the Judge that the length, structure (ie, the number of words) and letters used in the textual component of the Composite Mark and the Sign are identical (see [22] of the GD; see also Ozone Community at [49]). We also agree with the Judge that marks consisting of inverted words can be visually similar (see [21] of the GD). This has been found to be the case in a number of previous cases including:

(a) The decision of the CFI in Hedgefund Intelligence Ltd v Office for Harmonisation in the Internal Market (OHIM) (Case T-67/08) (11 June 2009) (“Hedgefund Intelligence”), where the contest was between the words “InvestHedge” (in the mark applied for) and “HEDGE INVEST” (in the earlier registered mark). There were differences in the colours, fonts and casing of the letters used in the two contesting marks; there was also the use of an adornment in the earlier registered mark, but not in the mark applied for. Nonetheless, the CFI held (at [35]) that “[t]he mere inversion of the elements of a mark cannot allow the conclusion to be drawn that there is no visual similarity”. The CFI further noted (at [36]) that the figurative elements of the two marks, including the colour of the mark applied for and the adornment in the earlier registered mark, were insignificant, and that “the characters of the signs in question” (at [36]), despite the use of lower case for the mark applied for and upper case for the earlier registered mark, were not sufficient for the purposes of distinguishing them. We also mention in passing in this connection the observations of the Federal Court of Australia in Renaud Cointreau & Cie v Cordon Bleu International Ltee (2001) 52 IPR 382 (at [68]) that it is artificial to place undue emphasis on the script and style of a word mark without regard to the words, which are what “give it substance”.

(b) The decision of the Federal Court of Australia in R & C Products, where, in an action for misleading or deceptive conduct under the Trade Practices Act 1974 (Cth), the use of the words “Fresh & Blue” by the respondent was thought to be deceptive conduct in relation to the prior use by the applicant of the words “blue & fresh” for its product, both products being similar in nature and function.

(c) The decision of the hearing officer in Automobile Club de L’Ouest v Gardiakos (2005) 66 IPR 191, where two applications to register trade marks, each comprising the words “Adelaide Le Mans”, were disallowed on the opposition of the applicant having earlier priority, who had applied to register a trade mark comprising the words “Le Mans Adelaide”. The hearing officer held (at [32]) that having considered the applicable principles and tests, the “Adelaide Le Mans” marks were “deceptively similar” to the “Le Mans Adelaide” mark.

(d) The decision of the hearing officer in Macquarie Bank Limited v Great Southern Loan (SP) Pty Ltd (2007) 73 IPR 573, where the hearing officer (at [18]) had “no hesitation in finding that the compared trade marks [viz, “FORWARD THINKING” and “THINKING FORWARD”] are deceptively similar”.

64 Of course, the assessment of similarity is ultimately one of fact arrived at upon the application of well-established principles. We have set out some of the key principles at [39]–[41] and [62] above, and we do not foreclose the possibility that even allowing for the imperfect recollection of the average consumer who is not looking at the contesting marks one alongside the other, the mere reversal of the elements comprising the marks and the use of distinct and differentiated fonts or styles might conceivably be enough to displace the similarity between the marks. But, as a matter of common sense and having regard to the overall impressions that are likely to be formed, this would not normally be the case, and on the facts before us, we do not think these differences alone displace the visual similarity between the Composite Mark and the Sign.

65 The Judge was also of the view that the textual component of the Composite Mark was the dominant component in the mark. We agree, though for slightly different reasons, on the facts of this case. In our judgment, the device component of the Composite Mark, which consisted of a single simple stylised rose, was relatively insignificant for several reasons:

(a) As noted above, the textual component of the Composite Mark was distinctive as the words “Lady Rose” conveyed no meaning and had no notional or allusive quality. The device component, which was a simple stylised rose, if anything, underscored the textual component by emphasising the “Rose” portion of “Lady Rose”. Reference may be had in this regard to the decision of the Fourth Board of Appeal in the English case of Dragonfly Teas Limited v Dawn Boreham, Richard Simon Jeremy Boreham (Case R 271/2010-4) (7 April 2011), where, in a contest between an earlier word mark consisting of the word “DRAGONFLY” and a later composite mark that had the same word in normal type face accompanied by a device of a dragonfly, it was observed (at [30]):

... The device element will immediately be understood to represent a dragonfly, which merely reinforces the meaning of the word element. The word element ... remains the dominant or at least co-dominant element of the [later composite mark]. ... [T]he device of a dragonfly only acts to reinforce the word element ‘Dragonfly’ of the [later composite mark] in the perception of the relevant public as it is reasonable to assume that the word ‘Dragonfly’ is accompanied by the device of a dragonfly.

(b) The simple stylised rose in the Composite Mark was not a feature that drew any significant attention away from the words.

(c) A customer trying to recollect the Composite Mark would be far more likely to recall the textual component rather than the device component.

(d) If the textual component of the Composite Mark and the Sign were found to be substantially similar, it was unlikely that the simple stylised rose in the Composite Mark would serve as a point of distinction.

(e) In Trubion as well as Oberhauser, which each featured a more complex device component, the respective device components were thought, if anything, to be likely to be perceived as decorative elements and not as elements indicating the commercial origin of the goods in question. In Honda (at [31]), the “S” device component (reproduced at [57(a)] above) was found not to be significant because it was neither “original [n]or very elaborate”, and because it was the first letter of the textual component “SEAT” and, in that sense, emphasised the latter.

(f) It was also noteworthy that there were at least some products that Hai Tong sold where the words “Lady Rose” without the floral device (as opposed to the Composite Mark in its entirety) were used (especially on smaller items).^{[note: 5]}

66 Hai Tong also submitted that flower devices were common to the cosmetics trade and pointed to the use of a flower device on the packaging of the allegedly infringing products. Hai Tong relied on a passage from Mark Davison, Tracey Berger & Annette Freeman, Shanahan’s Australian Law of Trade Marks and Passing Off (Lawbook Co, 4th Ed, 2008) at para 30.1545, where the learned authors state:

Where an element which is common to [the] trade marks under comparison is ... in common use in the trade, its presence must be to some extent discounted in considering whether the trade marks are deceptively similar. …

67 We note that a similar argument was accepted by the CFI in Shaker (CFI) as one of the reasons why the textual component of the composite mark in that case stood out over the device component (see Shaker (CFI) at [42]). This was also the case in Oberhauser (see Oberhauser at [47]–[48]). To the extent that it was submitted on this basis that the flower device in the Composite Mark was not distinctive or significant, we agree. This is not to say that the “Rose Lady” get-up on the allegedly infringing products was similar because it too featured a flower. The floral prints on the packaging of these products (and “Rose Lady” products in general) seem clearly decorative in nature and, in any case, are quite different from the flower device in the Composite Mark. The floral prints in the “Rose Lady” get-up feature a graphical design of a flower with five petals viewed aerially from above, while the flower device in the Composite Mark is a side-cut view of a rose. That said, this does serve to highlight the relative insignificance of the device component of the Composite Mark and to support the conclusion that on the facts of this case, the textual component of the Composite Mark is dominant. In all the circumstances, we are satisfied that the Sign and the Composite Mark are visually similar.

Aural similarity

68 We turn to the assessment of aural similarity between the Composite Mark and the Sign. We agree with the Judge’s finding that the Composite Mark and the Sign are aurally similar (see [24] of the GD). When comparing marks or signs aurally, a quantitative assessment is carried out to ascertain if the marks or signs in question have more syllables in common than not (see Sarika at [28] and Ozone Community at [55], citing with approval L Bently & B Sherman, Intellectual Property Law (Oxford University Press, 3rd Ed, 2009) (“Bently & Sherman”) at p 865). The syntax of the marks or signs being compared is also considered. Hence, it is relevant to consider the total number of syllables in each mark or sign concerned and ascertain whether there is a pause in the reading of the mark or sign, as when a mark or sign is made up of two discrete words (see Ozone Community at [14] and Intuition Publishing at [54]).

69 In the present case, there is no doubt that both the Composite Mark and the Sign consist of exactly the same syllables. The only difference is in their syntax in that they are read in inverted order. However, we think that the fact that the syllables are articulated in an inverted order does not prevent the Composite Mark and the Sign from being aurally similar (see Hedgefund Intelligence at [39]; this also accords with our earlier observation (at [63] above) that marks consisting of inverted words can be visually similar). Hence, we agree with the Judge that the Composite Mark and the Sign are aurally similar.

Conceptual similarity

70 Finally, we turn to the assessment of conceptual similarity. We also agree with the Judge’s finding that the Composite Mark and the Sign are conceptually similar. In considering whether there is conceptual similarity, the inquiry is directed at the ideas that lie behind or inform the marks or signs in question (see Sarika at [34] and Intuition Publishing at [55]). In the present case, we agree with the Judge’s finding that the Composite Mark and the Sign are conceptually similar as both evoke the idea of “something floral and feminine” (see [25] of the GD).

71 In the overall comparison of the similarity of the Composite Mark and the Sign, we find them sufficiently similar for the purposes of s 27(2)(b) of the Act and so move to the final stage of the inquiry, namely, the inquiry into the likelihood of confusion on the part of the public as to the origin of, respectively, Hai Tong’s products and the allegedly infringing products.

Likelihood of confusion

How is this different from the inquiry into similarity?

72 It is important first to recall the proper province of trade mark protection, which concerns protecting the right of a trade mark owner to use a species of property – namely, his mark – as an exclusive badge of origin to denote that goods bearing that mark emanate from him. The point has been well explained by the ECJ in SA CNL-SUCAL NV v HAG GF AG (Case C-10/89) [1990] ECR I-3752 (“SA CNL-SUCAL”) at [13]–[14]:

13 Trade mark rights are, it should be noted, an essential element in the system of undistorted competition which the Treaty [ie, the Treaty establishing the European Economic Community] seeks to establish and maintain. Under such a system, an undertaking must be in a position to keep its customers by virtue of the quality of its products and services, something which is possible only if there are distinctive marks which enable customers to identify those products and services. For the trade mark to be able to fulfil this role, it must offer a guarantee that all goods bearing it have been produced under the control of a single undertaking which is accountable for their quality.

14 Consequently, as the Court has ruled on numerous occasions, the specific subject-matter of trade marks is in particular to guarantee to the proprietor of the trade mark that he has the right to use that trade mark for the purpose of putting a product into circulation for the first time and therefore to protect him against competitors wishing to take advantage of the status and reputation of the trade mark by selling products illegally bearing that mark. In order to determine the exact scope of this right exclusively conferred on the owner of the trade mark, regard must be had to the essential function of the trade mark, which is to guarantee the identity of the origin of the marked product to the consumer or ultimate user by enabling him without any possibility of confusion to distinguish that product from products which have another origin (see, in particular, the judgments in Case 102/77 Hoffmann-La Roche v Centrafarm [1978] ECR 1139, paragraph 7, and in Case 3/78 Centrafarm v American Home Products Corporation [1978] ECR 1823, paragraphs 11 and 12).

[emphasis added]

(See also the observations of the ECJ in Hoffmann-La Roche & Co AG, Basel and Hoffmann-La Roche AG, Grenzach-Wyhlen v Centrafarm Vertriebsgesellschaft Pharmazeutischer Erzeugnisse mbH (Case 102/77) [1978] ECR 1139 at 1152 and Canon at [28]–[29].) Without such a regime, “consumers would no longer be able to identify for certain the origin of the marked goods and the proprietor of the trade mark could be held responsible for the poor quality of goods for which he was in no way accountable” (see SA CNL-SUCAL at [16]).

73 In a claim for trade mark infringement, the plaintiff must show that there is a likelihood of confusion as to the origin of, respectively, his goods and the defendant’s goods by virtue of the identity or similarity between his mark and the defendant’s sign as well as the identity or similarity between the goods to which they are applied. Confusion in this context extends to the situation where the consumer mistakenly assumes some kind of economic link or connection between goods bearing the registered mark and those bearing the allegedly infringing sign.

74 Indeed, the authorities are clear that there are at least two specific aspects to the element of confusion. The first is mistaking one mark for another, so that the perception of the relevant segment of the public is that the allegedly infringing mark is in fact the same as the registered mark. In short, in this context, the relevant segment of the public may not even perceive that there are two distinct marks. The second is where the relevant segment of the public may well perceive that the contesting marks are different, but may yet remain confused as to the origin which each mark signifies and may perceive that goods bearing the earlier registered mark and those bearing the later allegedly infringing mark emanate from the same source that has chosen to diversify its range of marks for any of a multitude of marketing or promotional reasons or that the two sets of goods emanate from sources that are economically linked or associated.

75 However, the “mere association” by the relevant segment of the public of the earlier registered mark with the later allegedly infringing sign based on their similar use is not in itself a sufficient basis for concluding that there is a likelihood of confusion in the absence of any possibility of misapprehension as to the origin of the goods or services concerned (see City Chain at [58]). Put another way, it is not sufficient that the relevant segment of the public would recognise or recollect similarities between the contesting marks if there is no likelihood of confusion as to origin ensuing.

76 Unlike the inquiry into similarity, which is a somewhat objective inquiry directed at the features of the marks in question, the inquiry into the likelihood of confusion focuses on the effect that such similarity is likely to have on the relevant segment of the public. Specifically, the question is how they are likely to perceive the origin of goods or services bearing the contesting marks, having regard to the similarity (or, as the case may be, the identity) between the contesting marks as well as the similarity (or identity) between the goods or services in relation to which each is used.

77 It is evident that the assessment of the likelihood of confusion has the potential to become a more fact-sensitive inquiry than the one directed at the element of similarity between contesting marks. It is useful here to set out s 27(2)(b) of the Act, which, in its material part, reads as follows:

A person infringes a registered trade mark if, without the consent of the proprietor of the trade mark, he uses in the course of trade a sign where because —

...

(b) the sign is similar to the trade mark and is used in relation to goods or services identical with or similar to those for which the trade mark is registered,

there exists a likelihood of confusion on the part of the public.

78 This can be broken into several elements as follows:

(a) the infringer must be acting without the consent of the trade mark owner;

(b) the infringer must use his sign in the course of trade, and as we have noted above, this means he must use it as a badge or identifier of origin and not for purely decorative or descriptive purposes;

(d) the trade mark and the allegedly infringing sign must be used on identical goods or services (as is the case here), or at least on goods or services that are similar to one another; and

(e) because of elements (c) and (d) above, there exists a likelihood of confusion on the part of the public. The likelihood of confusion required is that which is to be expected amongst a substantial portion of the relevant segment of the public. While not amounting to a majority, the relevant segment of the public must go beyond a “de minimis” level (see Sarika at [57] and Mobil Petroleum Co, Inc v Hyundai Mobis [2010] 1 SLR 512 at [77]–[78]).

What factors may be considered?

79 The last element of s 27(2)(b) of the Act is potentially unclear. The words of the statute could be construed as suggesting that the likelihood of confusion is a conclusion that follows necessarily and, indeed, only upon the existence of similarity between the registered mark and the allegedly infringing sign as well as similarity (or identity) between the goods in relation to which each is used. On the other hand, if the inquiry is to be determined solely on the basis of such similarity, why was it necessary to stipulate in the Act that the likelihood of confusion must ensue? It is convenient to frame the broad parameters of the issue by setting out at some length an extract from the decision of Lai Kew Chai J in The Polo/Lauren Co, LP v Shop In Department Store Pte Ltd [2005] 4 SLR(R) 816 (“Polo (HC)”) at [11]–[12]:

11 The plaintiff argued that this court should follow the classic formulation as stated in a string of cases including British Sugar plc v James Robertson & Sons Ltd [1996] RPC 281 (“British Sugar”), Saville Perfumery Ld v June Perfect Ld and F W Woolworth & Co Ld (1941) 58 RPC 147 (“June Perfect”) and Wagamama Ltd v City Centre Restaurants Plc [1995] FSR 713 (“Wagamama”). The legislation considered in those cases are in pari materia with s 27(2)(b) of the [1999] TMA. Under this approach, it is necessary to show step-by-step that there is a likelihood of confusion because the marks are similar and the goods for which the mark is used are identical or similar to those covered by the registration. The plaintiff further contended that the test by which confusion is assessed is extremely strict such that the court cannot take into account added matter or circumstances in assessing whether there is a likelihood of confusion in relation to the way the defendant uses its sign.

12 The defendant, predictably, argued for a more global approach, one which has its roots in a European effort to harmonise their trade marks regime. This is seen in the tenth recital in the preamble to the First Directive 89/104/EEC of the Council of European Communities, of 21 December 1998, to Approximate the Laws of the Member States Relating to Trade Marks (OJ EC No L40 of 11.2.1989) (“the Directive”), where it is stated that “it is indispensable to give an interpretation of the concept of similarity in relation to the likelihood of confusion” and that an appreciation of the likelihood of confusion depends on numerous elements, in particular, the recognition of the trade mark in the market, any association which can be made with the used or registered sign, and the degree of similarity between the trade mark and the sign and between the goods and services identified. This so-called “global assessment test” was endorsed in the leading European Court of Justice (“ECJ”) decision in Sabel BV v Puma AG, Rudolf Dassler Sport [1998] RPC 199 (“Sabel”) and followed by the English Court of Appeal in The European Limited v The Economist Newspaper Limited [1998] FSR 283 (“The European”).

[emphasis in original]

80 Lai J went on to make it clear (at [15]–[16] of Polo (HC)) that he rejected the global assessment test that had been set out in Sabel because he considered that:

(a) the global assessment test was not compatible with our legislation, including the then equivalent of s 27(2)(b) of the Act; and

(b) under the global assessment approach, the ultimate test for infringement was the likelihood of confusion, and as a result, other factors aside from the similarity between the contesting marks and the similarity between the goods on which these marks were applied, such as the reputation of the earlier registered mark, could decide the issue.

81 The following further extracts from Lai J’s judgment (at [15]–[16] of Polo (HC)) bear setting out:

15 ... It is clear from the plain words of s 27(2)(b) that the likelihood of confusion, if any, must be the direct result of the similarity between the registered mark and the disputed sign and between the goods and services used in relation thereto. The provision does not make it an infringement if the likelihood of confusion is caused by some other factor. Neither does it make similarity merely one among other factors that may be considered. This is obviously sound because the [1999] TMA protects the proprietary rights of the owner’s trade mark and nothing else. This conceptual clarity is lost if we adopt the global assessment test which takes as its ultimate test the question of whether there is a likelihood of confusion ...

16 ... If infringement is to be made out, it should be on the basis of the identity or similarity of the sign to the registered mark and the goods applied thereto. I should add that while the global assessment test should be rejected, the cases that have followed this approach contain useful dicta particularly in respect of how the court should approach the question of whether two marks are similar.

[original emphasis omitted; emphasis added in italics]

82 However, Lai J also eschewed an unduly restrictive approach. He thought that there were two ways in which “factors extraneous to the mark” (see Polo (HC) at [22]) might be relevant:

(a) to establish the degree of similarity required in order to decide the question of whether there existed a likelihood of confusion, Lai J thought that the court might need to be apprised of such things as the trading circumstances of the goods in question as this might have a bearing on how much care a customer would bring to his or her prospective purchase, and this in turn could affect how similar the marks needed to be to give rise to a sufficient likelihood of confusion; and

(b) to establish that the likelihood of confusion was real and not merely hypothetical or speculative.

83 Subsequent cases have not always (or consistently) applied this analysis or maintained the conceptual distinction between the global assessment approach that is applied in European jurisprudence by reason of the relevant legislative and treaty provisions that apply there and the approach suggested by Lai J, which contemplates the use of some extraneous material but within a more structured analytical framework: see the analysis of the line of case law after Polo (HC) that is set out in IP Law of Singapore at paras 28.374–28.385.

84 In our judgment, Lai J was correct in insisting upon regard being had to the terms of our legislation and demanding that conceptual clarity be maintained in this assessment. It is helpful to begin with the Act itself:

(a) It bears emphasising that under s 4(1) of the Act, a registered trade mark is a property right which gives the proprietor the rights and remedies provided by the Act.

(b) Under s 8(2) of the Act, one of the relative grounds for refusing registration of a mark is the existence of an earlier mark that is identical or similar to the one for which registration is sought and where both are used in relation to similar or identical goods or services. This is the mirror provision to s 27(2), which concerns infringement being established on these grounds, save that whereas s 8(2) concerns the likelihood of confusion arising due to the intended registration of a later aspiring mark, s 27(2) concerns that arising from its use. But, there is nothing to suggest that the test that is to be applied at the time of registration to determine registrability should be materially different from that applied to determine whether use of a later sign constitutes infringement.

(c) Under s 26(1) of the Act, the proprietor is conferred the exclusive rights “to use the trade mark” and to authorise others to do so in relation to the good or services for which it has been registered.

(d) Under s 22 of the Act, a registered trade mark may be revoked if it has not been used for a period of five years following its registration, or if its use has been suspended for an uninterrupted period of five years without any proper reason. A trade mark may also be revoked if, because of the acts or omissions of the proprietor, the mark has become generic so that it no longer serves the function of being a badge of origin, or if it has become deceptive in a material respect.

(e) Further, under s 23 of the Act, a trade mark may be declared invalid if, among other things, there is an earlier trade mark in relation to which s 8 applies.

(f) Save as expressly provided in the Act, there is no basis for diluting or extinguishing the exclusive rights and remedies that accrue to the proprietor of a registered trade mark.

85 We have set out these provisions because they are instructive in construing the proper approach to be taken by the court in assessing the likelihood of confusion. In our judgment, the following principles can be deduced by construing these provisions as a whole and harmoniously:

(a) Given the correspondence between s 8(2) on the one hand and s 27(2) on the other, we do not think the sort of considerations that can be taken into account to assess the likelihood of confusion under s 27(2) should be materially different from those that apply under s 8(2). Indeed, it seems implausible that an application for registration by a party using a sign that is contested under s 8(2) would yield a different outcome than would be the case in an action for infringement initiated by the owner of a registered mark under s 27(2).

(b) The proprietor of a trade mark is conferred the exclusive right to use the trade mark in relation to the class of goods or services for which it has been registered. Further, the proprietor has the right to obtain relief under the Act for infringement of his trade mark. The Act should be construed in a manner that does not dilute these rights save in accordance with the statutory framework, for instance, either (under s 23) on showing that in fact, the requirements of s 8(2) (or such other relevant provisions as the case may be) do apply and ought to have warranted a refusal of registration, or (under s 22) for non-use of the trade mark or use leading to its loss of utility as a badge of origin altogether.

(c) Having regard to the express terms of s 27(2), there are three specific elements that plainly must be considered. These are: (i) the similarity between the registered mark and the allegedly infringing mark; (ii) the similarity or identity between the goods or services in relation to which the marks are used; and (iii) the relevant segment of the public in relation to whom the court must consider the likelihood of confusion. Each of these elements can vary. The marks may be identical or similar, and if the latter, they can vary in their degree of similarity. In the same way, the goods or services in relation to which the marks are used may be identical or similar, and again, if the latter, they may vary in the degree or extent to which they are similar. In assessing the degree of similarity, it may be appropriate to have regard to steps that have been taken to differentiate the goods or services. And as to the relevant segment of the public, there may be characteristics that are particular to the group in question. Each of these factors will have a bearing on the likelihood of confusion. As an illustrative proposition, the likelihood of confusion must be greater where, say, the contesting marks are similar to a high degree, the goods are identical in nature and the segment of the public in question is undistinguished in its attention than would be the case if the marks and the goods are somewhat similar but not exceedingly so, and the relevant segment of the public happens to be highly knowledgeable and very fastidious. We earlier observed at [40(b)] above that in comparing the similarity between two contesting marks (ie, element (i) above), this is done without regard to “external added matter or circumstances” (see MediaCorp at [33], citing [55] of Caterpillar). But, this is not the case when assessing the extent to which the goods bearing the contesting marks are similar and the likelihood of confusion among the relevant segment of the public (ie, elements (ii) and (iii) above). To prevent the assessment of these elements from becoming an exercise consisting largely of guesswork, it must be legitimate to have regard to extraneous factors that affect the degree and extent of the weight to be accorded to each of these elements and how they interplay with one another. Thus, for instance:

(i) It would be legitimate to have regard to the types of goods in question. So, other things being equal, if these are medicinal or highly technical products (and these are offered as examples only), it would be legitimate for the court to have regard to the fact that because of their nature, a higher degree of similarity between the contesting marks might have to be shown to warrant a finding of likely confusion than when dealing with products that demand less fastidiousness or care on the part of the relevant segment of the public.

(ii) If the registered mark in question is not yet being used, it would be legitimate for the court to assume that it is or will be used in a normal and fair manner and assess the likelihood of confusion on that basis, having regard to the way in which the defendant is using his allegedly infringing sign – see Origins Natural Resources Inc v Origin Clothing Limited [1995] FSR 280 (“Origins Natural Resources”) at 284.

(iii) It would be legitimate to consider the contesting marks and the degree of similarity that is found to exist. As a self-evident proposition, the greater the similarity, the more likely that confusion may ensue. A showing of some similarity between the contesting marks may not suffice to cross the threshold when the nature of and the differences between the products bearing the marks as well as the circumstances and characteristics of the relevant segment of the public are taken into account.

(iv) On the other hand, it will not be appropriate to consider factors that effectively and impermissibly dilute or diminish the exclusive rights that are given to the proprietor of a trade mark. For instance, it would not generally be permissible in the context of trade mark infringement, in contrast, conceivably, to passing off, to take into account differences in the colour of the packaging that happens to have been used. This is so for several reasons. First, a proprietor’s exclusive right to use his trade mark does not usually depend on the type or colour of packaging on which it is applied. If a later sign were to be regarded as dissimilar because it has been applied on a product that bears a different colour of packaging, then this could curtail the right of the proprietor of the registered mark to use his mark with packaging of that colour, and in our judgment, there is no basis for such a curtailment. Further, it may be noted that consistent with s 3(2), under s 27(4) of the Act, a sign is used, among other things, when it is applied on goods or their packaging. Moreover, a trade mark itself is not registered in general by reference to any particular packaging or colour of packaging. The position may of course be different where some aspect of the packaging is part of the trade mark, or where the packaging is part of the product (as may be the case with some types of medication), or if the difference in packaging between the trade mark proprietor’s goods and the competing goods is of such a nature that it in fact differentiates the goods themselves or the segment of the public that is targeted (see further our observations below at [94(e)]–[94(f)]), or if it is shown that the packaging is part of the trade mark owner’s badge of origin or trade get-up. But, in the usual case, this should not factor in the analysis.

86 In In the Matter of an Application by the Pianotist Company Ld for the Registration of a Trade Mark (1906) 23 RPC 774 (“Pianotist”), where an application to register “Neola” as a word mark for pianos was opposed by the proprietor of the “Pianola” trade mark, Parker J observed as follows (at 777):

… You must take the two words. You must judge of them, both by their look and by their sound. You must consider the goods to which they are to be applied. You must consider the nature and kind of customer who would be likely to buy those goods. In fact, you must consider all the surrounding circumstances; and you must further consider what is likely to happen if each of those trade marks is used in a normal way as a trade mark for the goods of the respective owners of the marks. If, considering all those circumstances, you come to the conclusion that there will be a confusion – that is to say, not necessarily that one man will be injured and the other will gain illicit benefit, but that there will be a confusion in the mind of the public which will lead to confusion in the goods – then you may refuse the registration, or rather you must refuse the registration in that case. …

87 It will be evident from this that Parker J was adopting an approach that focused on the contesting marks, the goods to which they were to be applied and the customer who would likely buy the goods in question. This is precisely in line with what we have set out above. In Pianotist, in concluding that there was no likelihood of confusion in relation to the marks in question in that case, Parker J had regard to the likely price at which the articles would be sold as well as the fact that those who were likely to purchase the articles would generally be educated (at 777–778). Pianotist has been followed locally (see Nation Fittings at [99] and the cases cited there). In our judgment, the passage from Parker J’s judgment that we have cited at [86] above offers a pithy and adequate description of the approach to be taken, having regard to the express words of our statute. Of course, in Pianotist, Parker J also mentioned the need to examine all the surrounding circumstances, but this must be understood in the context of the specific considerations that we have mentioned. Thus, while it is permissible to have regard to “extraneous factors”, this should not be taken as a reference to such factors “at large” and still less to those that are inconsistent with the legislative framework because they impermissibly curtail the rights granted to the registered proprietor of a trade mark. Rather, s 27(2) of the Act contemplates the interplay among:

(a) the degree of similarity between the contesting marks;

(b) the degree of similarity between the goods or services in relation to which these are to be used; and

in order to assess the likelihood of confusion. Therefore, those factors that bear upon these elements in combination or upon the similarity of the goods or services in question and the characteristics, nature and likely responses of the relevant segment of the public can and should be considered. But, where the goods or services in question are identical, then extraneous factors relating to that element will not arise for consideration in assessing the likelihood of confusion – see [22] and [85(c)] above as well as [103(a)] below. This follows from the simple fact that once an element is found to be identical, it ceases to be material to inquire whether it is similar, and consequently, the degree to and the ways in which it may be differentiated.

88 In our view, this is also consistent with the approach taken in the earlier English cases where contesting marks were compared without reference to extraneous factors, but where the courts did have regard to such considerations to the extent that they might inform the inquiry into the nature and the extent of the differences between the goods or services bearing the contesting marks or the particular characteristics of the relevant segment of the public among whom the likelihood of confusion was to be assessed. In June Perfect, the court at first instance applied a visual side-by-side comparison of the contesting marks, both of which featured the word “June” (at 156). This was rejected by the English Court of Appeal, which reversed the first instance decision on the ground, essentially, that one had to have regard to the recollection of a consumer of the essential or distinguishing feature of the contesting marks, rather than look at points of difference upon a visual examination of the marks (at 161–162). In his judgment, Sir Wilfrid Greene MR stated (at 161):

… [T]he statutory protection [conferred by the Trade Marks Act 1938 (c 22) (UK)] is absolute in the sense that once a mark is shown to offend, the user of it cannot escape by showing that by something outside the actual mark itself he has distinguished his goods from those of the registered proprietor. …

89 This might be read as suggesting an absolute bar on having regard to extraneous factors, but that would be incorrect in our view. That sentence is properly understood as directed at the proposition that in comparing the similarity between two contesting marks, it is impermissible to look at external factors – see [40(b)] above. This is evident from the sentences that immediately follow in the judgment of the Master of the Rolls, where he stated that “[a]ccordingly” (see June Perfect at 161), infringement could occur if there was sufficient resemblance between the contesting marks. The focus in this part of the judgment was on the similarity between the contesting marks alone. Consistent with this view, in the House of Lords, Viscount Maugham observed that if the defendants’ use of the word “June” constituted an infringement, then this would not be saved by a disclaimer that the defendants’ goods were nonetheless not of the same origin as the plaintiff’s goods or by the use of the word as part of a fuller name such as “June Hair Curler Co” or “June Perfect Ld” (at 174). Significantly, the House of Lords disallowed the third injunction which had been granted by the English Court of Appeal, which was directed at the defendants carrying on business under any name that included the word “June”, with Viscount Maugham observing (at 176) that “[the defendants] may be able by proper precautions to sell the [infringing] articles in connection with their name of June Perfect Ld., while clearly distinguishing those goods from the [plaintiff’s] goods”.

90 This is also how the dictum of Greene MR has been understood in subsequent cases: see Decon Laboratories Ltd v Fred Baker Scientific Ltd [2001] RPC 17 at [8(5)]; see also BP Amoco plc v John Kelly Ltd [2002] FSR 5, where, applying the said dictum from June Perfect and following the decision of Jacob J in Origins Natural Resources (see [85(c)(ii)] above), the Court of Appeal of Northern Ireland held that where the plaintiff had registered a particular shade of green as a trade mark, the defendant’s use of a similar shade but with a different logo was nonetheless an infringement. The court considered a number of authorities before it said as follows:

42 Under the 1938 Act [ie, the Trade Marks Act 1938 (c 22) (UK)] it was settled law that once a mark is shown to offend, the user of it cannot escape by showing that by something outside the actual mark itself he has distinguished his goods from those of the registered proprietor: see Saville Perfumery Ltd v. June Perfect Ltd (1941) 58 R.P.C. 147 at p. 161, per Greene MR. The same conclusion has been reached in a number of cases under the 1994 Act [ie, the Trade Marks Act 1994 (c 26) (UK)]. In Origins Natural Resources Inc v. Origin Clothing Ltd [1995] F.S.R. 280 at 284 Jacob J. said that section 10 of the 1994 Act:

requires the court to assume the mark of a plaintiff is used in a normal and fair manner in relation to the goods for which it is registered and then to assess the likelihood of confusion in relation to the way the defendant uses its mark, discounting external added matter or circumstances. The comparison is mark for mark ...

or, as Robert Walker J. observed in United Biscuits (UK) Ltd v. Asda Stores Ltd [1997] R.P.C. 513 at 535, mark for sign. Several other judges experienced in this field sitting at first instance have reached the same conclusion: we were referred to the decisions of Laddie J. in Wagamama Ltd v. City Centre Restaurants plc [1995] F.S.R. 713 at pp. 731–732, Neuberger J. in Premier Brand UK Ltd v. Typhoon Europe Ltd [2000] F.S.R. 767 at p. 778 and Pumfrey J. in Decon Laboratories Ltd v. Fred Baker Scientific Ltd ([[2001] RPC 17]).

43 There is no contrary authority at any level in the United Kingdom ... If the decision in Levi Strauss & Co v. Shah [1985] R.P.C. 371 remains good law, as the judge concluded, then it would follow that use of a registered mark with additions, whether by the proprietor or the offender, is to be regarded as use of the mark. We can see no compelling reason in principle why it should not be, and a conclusion to the contrary would deprive the proprietor of a registered mark of much of the protection which he should receive in consequence of registration. …

[emphasis added]

91 In British Sugar, Jacob J, taking the same approach, applied June Perfect and considered that “Robertson’s Toffee Treat” was identical with “Treat” because “the first two words are added matter” (at 294). But, he then went on (at 294–295):

The questions arising under section 10(2)(a) [of the Trade Marks Act 1994 (c 26) (UK)] are:

(1) Is the mark used in the course of trade?

(2) Are the goods for which it is used similar to those covered by the registration?

(3) Is there a likelihood of confusion because of that similarity?

The first of these questions causes no difficulty here. The problems arise under the second and third questions. [The plaintiffs] seek to elide the questions of confusion and similarity. Their skeleton argument contends that there is “use in relation to a product so similar to a dessert sauce that there exists a likelihood of confusion because the product may or will be used for identical purposes.” I do not think it is legitimate to elide the question in this way. The sub-section does not merely ask “will there be confusion?”: it asks “is there similarity of goods?”, if so, “is there a likelihood of confusion?” The point is important. ...

…

… [Q]uestions of similarity are independent of a particular likelihood of confusion.

[emphasis added]

92 It is thus evident that: (a) extraneous factors may not be considered in assessing the similarity between the contesting marks; and (b) the inquiry into the likelihood of confusion is a separate inquiry from that into the similarity between the marks or between the goods or services to which they are applied. What is significant is that in British Sugar, although Jacob J applied the dictum of Greene MR in June Perfect, he did so only in relation to the comparison of the contesting marks; in relation to the comparison of the similarity between the goods in question, he was perfectly content to examine extraneous factors. In fact, at 296–297 of British Sugar, he set out the sort of criteria that might be considered in assessing whether the goods were similar, as follows:

… I think the following factors must be relevant in considering whether there is or is not similarity:

(a) The respective uses of the respective goods or services;

(b) The respective users of the respective goods or services;

(d) The respective trade channels through which the goods or services reach the market;

(e) In the case of self-serve consumer items, where in practice they are respectively found or likely to be found in supermarkets and in particular whether they are, or are likely to be, found on the same or different shelves;

(f) The extent to which the respective goods or services are competitive. This inquiry may take into account how those in trade classify goods, for instance whether market research companies, who of course act for industry, put the goods or services in the same or different sectors.

This is rather an elaboration on the old judicial test for goods of the same description. It seeks to take account of present day marketing methods. … I do not pretend that this list can provide other than general guidance. …

[emphasis in original]

93 On the facts before him, Jacob J observed that although the goods in question (viz, ice-cream syrup and sweet spread) to some extent had the same use, they were not in direct competition, would be found in different places in supermarkets and had a somewhat different physical nature. Of course, the determination in each case will turn on its own facts, but the principles and the analytical process are not dissimilar. Ultimately, what underlies such an approach is that in trade mark law, the courts, in applying the applicable trade mark legislation (in our case, the Act), are seeking to strike a balance between, on the one hand, the interests of the proprietor of the registered trade mark whose property rights in the trade mark give him a monopoly of sorts, something which is needed to incentivise investment in and development of his business; and on the other hand, the interests of other traders who should be free to trade without unnecessary or undue restraints.

94 In Polo (CA), this court identified several factors which may aid in the assessment of whether there is a likelihood of confusion. Subject to the forgoing observations, we consider that a number of these (and others) remain relevant, including:

(a) the disparity in the prices of the goods or services concerned (see Polo (CA) at [27] and [34]), where this goes towards drawing out the real and substantive differences between them or towards highlighting the different degrees of care likely to be applied by the prospective customer in each instance. As observed by Jacob LJ in Reed Executive Plc v Reed Business Information Ltd [2004] RPC 40 at [78], “a fifty pence purchase in the station kiosk will involve different considerations from a once-in-a-lifetime expenditure of £50,000”;

(b) the nature of the customer and of the article in question (see Polo (CA) at [27] and also our observations at [85(c)(i)] and [87] above);

(c) the degree of similarity between the contesting marks and the overall impression given by them (see Polo (CA) at [28]; see also our observations at [85(c)] generally and, in particular, at [85(c)(iii)] above);

(d) the imperfect recollection of the average consumer (see Polo (CA) at [28]), because this is directly related to the likelihood of confusion in this group;

(e) the locations at which the goods or services concerned are sold or offered (see Polo (CA) at [34]), because this might aid in assessing the real similarities or differences between the goods or services as well as between the respective groups of customers that each category of goods or services might attract;

(f) the steps taken by the defendant to differentiate his goods or services from those of the registered trade mark proprietor (see Polo (CA) at [28]), because this goes directly to the degree of similarity or difference between the goods or services in question; and

(g) the importance of the goods or services concerned to the consumer. As observed by the European General Court in Trubion, medicinal products, whether or not acquired on prescription, and dietetic products in general may be regarded as products to which consumers who are reasonably well-informed and reasonably observant and circumspect will pay a high degree of attention. It goes without saying that this category of products is not closed.

95 It will undoubtedly be significant to have regard to the relevant segment of the public and any peculiar characteristics associated with this group. We affirmed in Sarika that where a product is commonly available and purchased by the general public, the relevant segment of the public would be the general public, though the court will try to ascertain who the target consumers really are (see Sarika at [59]; see also the extremely careful and helpful analysis applied by Andrew Phang Boon Leong J (as he then was) in Nation Fittings at [96]–[103]).

96 It will be evident that all such factors do come within the framework of s 27(2)(b) of the Act, being related to the contesting marks, the goods or services concerned and the relevant segment of the public amongst whom the likelihood of confusion is being assessed. It may well be, in the final analysis, that the approach set out above, when applied in real cases, might often prove to be a distinction without a difference when compared to the global assessment approach. This was thought to be the case by this court in Polo (CA) at [9] and by Phang J in Nation Fittings at [95], but we consider it important to retain the conceptual clarity that Lai J spoke of in Polo (HC) at [15] and to ensure that the analysis is carried out in a manner that is faithful to the statutory provisions that are applicable here. However, we consider that this conceptual clarity is best achieved by admitting such extraneous factors as are expressly contemplated by the terms of s 27(2) and in the manner explained at [85] and [87] above, rather than in the precise manner suggested by Lai J in Polo (HC) at [22] (see [82] above).

97 Lastly, we also consider that the factors relating to the similarity between the goods or services concerned, the similarity between the contesting marks and the characteristics of the relevant segment of the public must be assessed in an interdependent way. Thus, in Sarika at [65]–[66], this court had regard to the high degree of similarity between the contesting marks and the proximity in the nature of the goods concerned as consumable foodstuff in assessing the likelihood of confusion. Similarly, in Hedgefund Intelligence, the CFI (albeit in making a global assessment) observed (at [30]) that there would be a degree of interdependence between the factors to be considered:

... [G]lobal assessment implies a degree of interdependence between the factors taken into account and, in particular, between the similarity of the trade marks and that of the goods or services concerned. Accordingly, a low degree of similarity between the goods or services covered may be offset by a high degree of similarity between the marks, and vice versa ...

98 Reference may also be made to James Mellor et al, Kerly’s Law of Trade Marks and Trade Names (Sweet & Maxwell, 15th Ed, 2011) at para 9-056, where the same point is made.

Evidence of actual confusion

99 The question whether there is likely to be confusion is ultimately one for the court rather than for witnesses. This is also the position taken by the learned authors of Bently & Sherman (at p 874). In Tan Tee Jim SC, Law of Trade Marks and Passing Off in Singapore (Sweet & Maxwell Asia, 2nd Ed, 2005) (at para 6.38), the learned author states:

... [W]here [both the mark and the sign being compared] have been in use, it would be useful to adduce evidence to prove likelihood of confusion, such as:

(1)evidence of witnesses purporting to testify to actual confusion;

(2)evidence of trade witnesses who gave their opinion of the likelihood of confusion; and

(3)evidence of witnesses who had responded to a survey carried out on behalf of [the plaintiff].

100 But, finding and producing actual evidence of confusion may be difficult as members of the public are generally reluctant to attend proceedings in court. Moreover, those who are confused and deceived may remain so, and there is no reason to think that they would be moved from this state so as to bring their past experience to the attention of the parties. Furthermore, even those members of the public who have come to appreciate that they were misled will frequently have no reason to draw this to the attention of the parties, especially if the realisation comes some time after any relevant dealings with the parties. Finally, persons who discover that they have been deceived and decide not to do business with the offending party as a result may then have no further reason to contact the parties and inform them of their past confusion. For these reasons, while evidence of actual confusion can be very helpful, its absence should not be accorded undue significance (see the observations of this court in Novelty Pte Ltd v Amanresorts Ltd and another [2009] 3 SLR(R) 216 (“Amanresorts”) at [80]).

101 It also bears emphasising that unlike the action in passing off, a claim for trade mark infringement under s 27(2)(b) of the Act does not require proof of damage or the likelihood of damage. Although proof that there has in fact been a diversion of custom may be probative on the assessment of the likelihood of confusion, it is not a necessary part of the claim. Hence, while the inquiry into confusion needs to be contextualised appropriately as we have set out above, that does not extend to having to prove that the plaintiff’s customers have changed or are likely to change their purchasing behaviour as a result of any such confusion. This is unlike the position with passing off (see Chen Eng Waye at [63]).

Application to the facts

102 The Judge found that there was a likelihood of confusion on the part of the public. It is perhaps useful to set out in full the Judge’s reasoning in this regard (at [30]–[32] of the GD):

30 Here, I found that there was a likelihood of confusion on the part of the public. First, as discussed above, there was a high degree of similarity in the marks, and the goods were identical. As the court in [Ferrero SPA v Sarika Connoisseur Cafe Pte Ltd [2011] SGHC 176] held at [98], “accordingly, it stands to reason that where there is greater similarity in the marks and the goods, it would be easier to find a likelihood of confusion.”

31 Second, the goods were also being marketed in similar retail locations such as departmental stores and cosmetics shops, and targeted the same group of customers, females who used cosmetics. It thus followed that there was a risk that the public might believe that the “Rose Lady” goods came from the same source as the “Lady Rose” goods, or from economically-linked sources. Additionally, Mr Park In Pil, director of [Ventree], had conceded under cross-examination that it was possible that consumers could be confused by the similarity of the two marks:

Q:	Right, right, so that is actually my example, the mother just says to the daughter, “Please buy a Lady Rose foundation for me.” The daughter goes to the shop, she sees “Rose Lady”. He’s already accepted that “Lady Rose,” “Rose Lady” are similar. So it must follow that the daughter can come into the shop and be led to believe that they are the same product. You agree, right? …
Interpreter:	He agrees about the point.

32 As such, I found there was likelihood that a substantial portion of the relevant public, viz consumers of cosmetic products, would be confused.

[emphasis in original]

While we do not find these observations of the Judge to be objectionable (save that we do not think the hypothetical that was put to the witness in the course of cross-examination has any relevance in assessing the likelihood of confusion), there are in fact a number of other factors that further inform the analysis.

103 In particular, the following factors are significant:

(a) In relation to the various elements to be made out in this case (see [15] above), it should be noted that the alleged infringement consisted of the application of the Sign on goods of an identical nature to those for which the Composite Mark was registered. This much was conceded by the Defendants. It followed from this that it was therefore not relevant to inquire into minor differences in pricing or between the specific shades of cosmetics on offer, or whether or not they were placed on the same shelves, because such differences are immaterial where the goods in question are identical. It was therefore also not necessary or appropriate to consider the extent of similarity by reference to whether the goods had been differentiated (see [22] above). Nor were the differences in pricing or the nature of the goods in this case such as to entail a substantially different level of care being applied by prospective customers.

(b) For the reasons set out at [63]–[71] above, we are dealing with marks that had a significant degree of visual, aural and conceptual similarity.

(c) There was no material difference in the type of packaging used such as to differentiate the goods concerned. Nor, in any event, did that arise for consideration here where the goods were admittedly identical. It is true that Hai Tong’s products were mostly packaged in blue while the Defendants’ products were mostly packaged in beige. But, if differences as to the colour of the packaging that was used by Hai Tong and the Defendants respectively are ignored, as, in our view, they should be in this case for the reasons outlined at [85(c)(iv)] above, the similarity between the Composite Mark and the Sign is further accentuated. Hai Tong also observed that its products sometimes used packaging of a colour that was much closer in appearance to that used on the Defendants’ products.^{[note: 6]}

(d) The differences between the Composite Mark and the Sign consisted of the reversal of two identical words, but the weight of the authorities show that this has seldom, if ever, been accorded any significance. As for differences in font or typesetting, the same can be said. In relation to the rose device that was part of the Composite Mark, this, in our view, was not a dominant element in this case. Moreover, there were even products where Hai Tong affixed the textual component of the Composite Mark alone without the device component. Aside from this, not only are floral elements common place in female cosmetic products, the Defendants’ own products also featured a floral element, albeit in a decorative way.

(e) It is therefore evident that there really was not much that was material to distinguish between the Composite Mark and the Sign, and these were being applied, as we have already noted, on goods of an identical nature.

(f) Much was made of the contention that the words “Lady Rose” were bland. But, as we have observed at [36]–[38] above, inspiration is not the yardstick for distinctiveness. On the facts of this case, the Composite Mark was distinctive because it lacked any notional or allusive quality, and was in any event meaningless in relation to the goods for which it was registered.

(g) In terms of the prospective customers, both sets of products appeared to target the same segment of the market, namely, ladies who purchase cosmetic products that are not notoriously exclusive. The Defendants argued that the relevant segment of the market targeted by “Rose Lady” products was different from that targeted by “Lady Rose” products because “Rose Lady” products were targeted mainly at younger women, while “Lady Rose” products were targeted mainly at women in their 40s. The lack of supporting evidence aside, we found this too artificial and narrow a distinction.

(h) The products in question did not involve the sorts of expenditure that would mandate significant prior due diligence on the part of the relevant segment of the public.

104 If due allowance were then made for the fact that the Composite Mark and the Sign were not to be compared side by side, and the question was posed whether the prospective customer of the product in question who carried in her mind an imperfect recollection of the overall impression she had formed of the Composite Mark might be confused on seeing the Sign and think that it was either the same or at least related, the answer would plainly be yes, in our view. This, in the final analysis, is the true nature of the inquiry.

105 We note the Defendants’ submission that ladies, when selecting cosmetic products to be applied to their faces, would be more discerning as compared to when purchasing other items, because the selection of cosmetics involves assessing and choosing the suitability of the product for their complexion and skin type. However, there was nothing to suggest that there was either such a difference in the products in question or their relative quality as would have displaced the likelihood of confusion.

106 The Defendants submitted that Hai Tong did not adduce any evidence of actual confusion as to the origin of, respectively, “Lady Rose” products and “Rose Lady” products despite the Composite Mark having been registered and in use for at least 35 years, while Ventree had been importing “Rose Lady” products only for two years as at the time of the hearing below. However, for the reasons stated above at [99]–[101], we do not consider this fatal to the claim. Moreover, as the Judge found at [5] of the GD, the proceedings were initiated because of some confusion in the marketplace which became evident when Hai Tong received complaints from its customers of problems with the quality of its products that were sold in peach-coloured packaging, when it was “Rose Lady” products rather than “Lady Rose” products that used such packaging.

107 Having regard to all the factors and viewing them from the perspective of the average consumer with imperfect recollection, we were satisfied that there was a sufficient likelihood of confusion. This was so from the perspective that the average consumer in these circumstances might have mistaken one product line for the other; and also, that such a consumer, had she appreciated that the two product lines were not identical, might yet have been left with the impression that they originated from the same source or from related sources.

108 In these circumstances, we dismiss the appeal in CA 2 of 2012 with costs. In the premises, the Defendants’ counterclaim for groundless threats of trade mark infringement proceedings also fails.

Our decision in CA 1 of 2012

109 We now turn to the appeal in CA 1 of 2012, which is Hai Tong’s appeal against the Judge’s decision dismissing its claim against the Defendants for passing off. It is trite that in an action in the tort of passing off, the claimant must prove the “classic trinity” of goodwill, misrepresentation and damage (see Chen Eng Waye at [19]–[20]). Although the action for passing off and that for trade mark infringement overlap in some respects, there are a number of key differences. For one thing, the former is a common law action, whereas the latter is rooted in statute. Passing off is also to be distinguished from the usual action for trade mark infringement because it will be necessary for the plaintiff in the former, but not in the latter, to prove that:

(a) he has goodwill in his business;

(b) his goodwill is associated with his mark or name or get-up; and

110 Moreover, even in relation to the element of misrepresentation, it has been observed in Kellogg Co v Pacific Food Products Sdn Bhd [1998] 3 SLR(R) 904 (“Kellogg”) that at least in relation to the corresponding provision of the 1992 TMA, the test in the tort of passing off is probably a more demanding one than the corresponding inquiry in a trade mark infringement action (see Kellogg at [32]). In the former, it is necessary to show that the defendant’s actions amount to a misrepresentation that is likely to deceive the relevant segment of the public, whereas in a trade mark infringement action, the question is whether, because of the similarities between the registered mark and the allegedly infringing mark and between the goods or services on which these marks are applied, there is likely to be confusion. Moreover, as observed in Polo (HC) at [23] and endorsed by this court in Polo (CA) at [33], in an action in passing off, the likelihood of deception is to be assessed having regard to all the circumstances, whereas in a trade mark infringement action under s 27(2)(b) of the Act, the likelihood of confusion must stem from the similarity between the contesting marks and the similarity between the goods or services to which they are applied. Lastly, whereas passing off is concerned with protecting the plaintiff’s goodwill in his business, the action for infringement is geared towards protecting the trade mark owner’s proprietary rights in respect of his trade mark. With these observations, we turn to consider Hai Tong’s claim for passing off.

Goodwill

111 The bedrock of the tort of passing off is the plaintiff’s goodwill in his business. This has been famously described as the attractive force which brings in customers (see The Commissioners of Inland Revenue v Muller & Co’s Margarine, Limited [1901] AC 217 at 224).

112 In the present appeal, there is no doubt that customers have been exposed to the Composite Mark since it was first used in the 1960s. Over the years, Hai Tong has undoubtedly acquired goodwill through trading. The Judge found that this goodwill was “trivial” (at [39] of the GD) because he found the sales figures of Hai Tong’s “Lady Rose” line of goods in Singapore to be “not particularly impressive” (at [36] of the GD). The annual sales figures were $235,150 for the “Lady Rose” range in the year 2009, when the acts of the Defendants that are complained of first commenced.

113 In Amanresorts, we said (at [44]):

... [G]oodwill is not an all-or-nothing attribute in that it is not the case that the plaintiff either has goodwill in Singapore or, otherwise, has no goodwill at all. Clearly, goodwill can be limited to particular sections of the public. These sections of the public can be small so long as they are not negligible. ...

114 We do not think the Judge was correct to describe Hai Tong’s goodwill as either trivial or negligible. Hai Tong has developed a sustainable business over a period of more than four decades. We therefore find that Hai Tong’s goodwill is sufficient to warrant invoking the protection of the courts through an action in passing off.

Misrepresentation

115 The second element of the tort of passing off is misrepresentation. To bring an action in passing off, the plaintiff must show that his goodwill is associated with his get-up (ie, his mark, name and/or labelling) (see Amanresorts at [38]–[39]), and that by the use of a similar get-up, the alleged tortfeasor is passing off his goods as those of the plaintiff or as emanating from a source related to the plaintiff. In this connection, we observed in Chen Eng Waye at [28]–[29] that a threshold question which arises is whether the plaintiff’s name or mark is distinctive of his goods, and that distinctiveness is more easily found with fancy names that have no apparent relationship to the goods. That hurdle is crossed here for the reasons set out at [35]–[38] above. Further, we have found in respect of the claim under s 27(2)(b) of the Act that there was a sufficient degree of similarity between the Composite Mark and the Sign, and further, that there was a likelihood of confusion arising in the circumstances. We reiterate the factors set out at [103]–[104] above. Although the relevant tests for the tort of passing off and a claim for trade mark infringement are not identical, and although in an action for passing off, the court is not constrained in the same way that it would be in a trade mark infringement action in identifying the factors it may take into account, we are satisfied here that for those same reasons, the element of misrepresentation is made out.

Damage

116 Because the tort of passing off is directed at protecting the plaintiff’s goodwill, the plaintiff must show the real potential for damage to his goodwill.

117 In the GD, the Judge found that there was no likelihood of damage to Hai Tong’s goodwill (see [11(b)] above). The Judge reached this conclusion on the basis that “VOV”, with its far wider range of cosmetic products, was a better known brand than “Lady Rose”, and that there was hardly any evidence of damage to Hai Tong’s goodwill placed before him. We disagree with the Judge’s finding on this point because there was no evidence to justify a finding that “Lady Rose” products are of an inferior quality to VOV’s “Rose Lady” products, or that the former is an inferior brand as compared to the latter.

118 In the course of arguments, Hai Tong’s counsel submitted that on the facts of this case, given the identical field of business activity giving rise to direct competition, the similarity between the Sign and the Composite Mark as well as the likelihood of confusion which we found in relation to the trade mark infringement claim, there was a real likelihood of damage to Hai Tong’s goodwill arising from the diversion of sales of “Lady Rose” products if consumers of the relevant segment of the public were led to believe that “Rose Lady” cosmetic products were the goods of Hai Tong or were otherwise associated with or connected with Hai Tong. In our view, this is correct, and we further have regard to the matters which we alluded to at [103]–[107] above. The claim in passing off is therefore made out on these facts. Hai Tong’s appeal is therefore allowed.

Conclusion

119 In all the circumstances, we allow Hai Tong’s appeal in CA 1 of 2012 and dismiss the Defendants’ appeal in CA 2 of 2012. In the result, Hai Tong will have its costs here and below. However, although there were two appeals before us, the issues overlapped to a great extent and Hai Tong should have only one set of costs. The usual consequential orders will follow. We give liberty to the parties to seek our directions within 21 days of the date of this judgment as to any ancilliary orders that they may require.

^{[note: 1]}Appellants’ Core Bundle for CA 2 of 2012 at vol 1, pp 27–28.

^{[note: 2]}Para 55 of the Appellants’ Case dated 16 July 2012 in CA 2 of 2012.

^{[note: 3]}Para 30 of the Appellants’ Case dated 16 July 2012 in CA 2 of 2012.

^{[note: 4]}It should be noted that the Assistant Registrar for Trade Marks in Lee Cooper referred to the Trade Marks Rules 1968 (S 46/1968) (see Lee Cooper at 462), but it was actually the Trade Marks Rules 1991 which were applicable, the Trade Marks Rules 1968 having been revoked on 1 March 1991.

^{[note: 5]}See Hai Tong’s Further Submissions filed on 30 October 2012 at para 35.

^{[note: 6]}Respondent’s Skeletal Arguments dated 8 October 2012 filed in CA 2 of 2012 at para 49.

Hai Tong Co (Pte) Ltd v Ventree Singapore Pte Ltd and another and another appeal [2013] SGCA 26

Hai Tong Co (Pte) Ltd v Ventree Singapore Pte Ltd and another and another appeal
[2013] SGCA 26