Skip to content
laws.sgby aturio
Grounds publishedDecided 16 Oct 2012
Contents

Court of Appeal

Sarika Connoisseur Cafe Pte Ltd v Ferrero SpA

Feedback & support

[2012] SGCA 56Case no. Civil Appeal No 102 of 2011

What happened, in plain English

There is no plain-English summary of this judgment yet.

Ask about this judgment

Grounds of decision

Official text from eLitigation

  1. Chao Hick Tin JA (delivering the judgment of the court):

  2. Introduction

  3. [1] This is an appeal against the decision of a High Court judge (“the Judge”) given in Suit No. 9 of 2010 which was instituted by Ferrero S.P.A (“the Respondent”) against Sarika Connoisseur Cafe Pte Ltd (“the Appellant”) for trade mark infringement and passing off.

  4. [2] The Appellant, a company incorporated in Singapore, is a retailer in the food and beverage business and a wholesaler of coffee, cocoa and tea. It owns and operates a chain of cafe outlets (“cafes”) presently known as “tcc – the connoisseur concerto” (“TCC”). In its cafes, it offers customers, amongst others, a variety of gourmet coffee beverages and coffee concoctions.

  5. [3] The Respondent is a company incorporated in Italy. It manufactures and retails confectionery, including a cocoa-based hazelnut spread (“Nutella spread”) marketed under the brand name “Nutella”, which is the product concerned in the present case. The spread is sold at various retail outlets in Singapore, including supermarkets. The Respondent is the registered proprietor of the following marks (shown below) in Class 30 of the International Classification of Goods and Services in Singapore (“ICGS”). To avoid confusion, we will, as did the Judge, refer to the first mark as the “Nutella” word mark, while all three marks will be referred to collectively as the “Nutella” trademarks:

  6. Background to the dispute

  7. [4] The facts of this case are neither complex nor in serious dispute. On 1 August 2007, the Appellant introduced a new gourmet hot coffee beverage served in a shot glass under the “Nutello” sign in its cafes. As the Judge rightly noted, the term “Nutello” is more appropriately described as a “sign” under the definition in section 2(1) of the Trade Marks Act (Cap 332, 2005 Rev Ed) (“TMA”), as distinguished from a “trade mark” defined in section 2(1) of the TMA. The ingredients in the “Nutello” drink (“Nutello beverage”) comprised espresso, milk foam, cocoa powder and Nutella spread, amongst others. This drink was listed under the “Espresso Specialties” section of the TCC Drinks Gallery Menu (the “menu”). The menu describes the Nutello beverage as: “Espresso with lashings of nutella – perfect for cocoa lovers!” (“menu description”).

  8. [5] To effect sale of the Nutello beverage, the appellant had used the term “Nutello” in various forms in its promotional material such as the menu, booklet and website. They are set out below:

  9. [6] On 3 December 2009, the Respondent issued a cease-and-desist letter to the Appellant, objecting to the latter’s use of the “Nutello” sign and the menu description of the Nutello beverage. Damages amounting to S$50,000 were demanded.

  10. [7] On 15 December 2009, the Appellant requested an extension of one month till 17 January 2010 to respond to the Respondent’s letter, but in a letter dated 16 December 2009, the Respondent refused to accede to this request. On 6 January 2010, the latter instituted Suit No 9 of 2010 (“the action”) in the High Court.

  11. [8] In July 2010, sales of the Nutello beverage were discontinued by the Appellants, ostensibly due to an annual menu overhaul.

  12. The decision below

  13. [9] In the court below, the Judge held that:

  14. (a) the Appellant had infringed the Respondent’s “Nutella” word mark under s 27(2)(b) TMA given that there was a similarity of marks and goods resulting in a likelihood of confusion by the relevant public;

  15. (b) the Respondent’s “Nutella” marks are well known trademarks both to the relevant sector of the public as well as to the public at large in Singapore. The Appellant had infringed the Respondent’s “Nutella” marks under s 55(2) and s 55(3)(a) TMA, and had diluted the Respondent’s marks under s 55(3)(b)(i) TMA. However, he found that the Appellant had not taken unfair advantage of the distinctive character of the Respondent’s “Nutella” marks under s 55(3)(b)(ii) TMA; and

  16. (c) the Appellant had committed the tort of passing vis- à-vis the Respondent’s “Nutella” product when the former promoted and offered for sale its “Nutello” drink.

  17. [10] Consequently, the Judge granted injunctions to restrain the Appellant from infringing the Respondent’s “Nutella” trademarks and from passing off its “Nutello” products as products of, or connected to, the Respondent. He also ordered that damages be assessed.

  18. Issues in the appeal

  19. [11] The appellant raised the following issues in this appeal (with accompanying sub-issues):

  20. (a) Whether there is trademark infringement under s 27(2)(b) TMA? (“The First Issue”)

  21. (i) Whether the “Nutello” sign is similar to the “Nutella” word mark;

  22. (ii) Whether the “Nutello” sign was used in goods similar to those which the “Nutella” word mark is registered for;

  23. (iii) Whether there is a likelihood of confusion by the public?

  24. (b) Whether there is trademark infringement under s 55(2) TMA? (“The Second Issue”)

  25. (c) Whether there is infringement by damaging connection of well known trademark under s 55(3)(a) TMA? (“The Third Issue”)

  26. (d) Whether there is dilution by blurring under s 55(3)(b)(i) TMA? (“The Fourth Issue”)

  27. (i) Whether it is conceptually incongruous to find both infringement and dilution;

  28. (ii) What is required to show dilution in an “unfair manner”;

  29. (iii) Whether dilution by blurring is made out on the facts?

  30. (e) Whether the damage element needed to establish the tort of passing off is made out? (“The Fifth Issue”)

  31. (i) Whether actual damage has to be shown to establish the said damage element;

  32. (ii) Whether the Respondent’s expansion into the Singapore drinks industry would be restricted?

  33. [12] We note that the Appellant is not appealing against the finding of the Judge that the “Nutella” trademarks are well known marks, and neither has the Respondent taken issue with the Judge’s finding that there is no unfair advantage taken by the Appellant of the “Nutella” trademarks’ distinctive character under s 55(3)(b)(ii) TMA.

  34. [13] For ease of reference, we set out here the legislative provisions in the TMA, viz, s 27(2)(b), s 55(2), s 55(3)(a) and s 55(3)(b)(i), that will be referred to in the discussions below:

  35. The decision of this court

  36. The First Issue: Whether there is trade mark infringement under section 27(2)(b) TMA?

  37. [14] The step-by-step approach propounded in British Sugar plc v James Robertson & Sons Ltd [1996] RPC 281 (“British Sugar”) in relation to the English equivalent of s 27(2)(b) TMA was approved and adopted by this Court in The Polo/Lauren Co, LP Shop In Department Store Pte Ltd [2006] SGCA 14 (“Polo (CA)”) at [8]. This court in Polo (CA) was of the view that this approach was conceptually more appropriate and in line with the structure of the provision, as opposed to the global assessment test enunciated by the European Court of Justice in Sabel BV v Puma AG, Rudolf Dassler Sport [1998] RPC 1999 and Canon Kabushiki Kaisha v Metro-Goldwyn-Mayer Inc [1999] RPC 117. Under s 27(2)(b) TMA, three conditions need to be satisfied in turn for infringement to be established. First, the alleged offending sign, “Nutello”, must be shown to be similar to the registered mark, “Nutella”. Second, both the sign and the mark must be used in relation to identical or similar goods or services. Third, because of the presence of the first two conditions, there must exist a likelihood of confusion on the part of the public.

  38. (i) Similarity between the “Nutello” sign and the “Nutella” mark

  39. [15] The Judge had found that the “Nutello” sign and the “Nutella” word mark were not identical, and this finding is not challenged by the Respondent. In this appeal, the Appellant has only raised arguments on dissimilarity between the sign and the word mark and sought to dispute the individual findings of the Judge relating to visual, aural and conceptual similarity of the two.

  40. General approach to the similarity analysis

  41. [16] Before we determine whether the sign and the trademark are similar, we will first set out briefly the relevant principles of this analysis. The decided cases have established that the court will consider three aspects of similarity, viz, visual, aural and conceptual similarity: Polo (CA) at [24]; Mobil Petroleum Co, Inc v Hyundai Mobis (“Hyundai Mobis (CA)”) at [17]. However, it is not a pre-requisite that all three aspects of similarity must be made out before there can be a finding of similarity between the sign and the mark: Mediacorp News Pte Ltd v Astro All Asia Networks plc [2009] 4 SLR(R) 496 at [32] (“Mediacorp”). The relative importance of each aspect of similarity varies with the circumstances, in particular, with the goods and types of marks: Mediacorp at [32], citing Bently and Sherman, Intellectual Property Law (Oxford University Press, 3 Ed, 2009) (“Bently & Sherman”) at p 864. Simply put, a trade-off between the three aspects of similarity can be made, and each case ought to be viewed in its own context: Ozone Community Corp v Advance Magazine Publishers Inc [2010] 2 SLR 459 (“Ozone Community”) at [40]. Whether there is similarity between the sign and the mark is a question of fact and degree for the court to determine: City Chain Stores (S) Pte Ltd v Louis Vuitton Malletier [2010] 1 SLR 382 (“City Chain”) at [47]; Johnson & Johnson v Uni-Charm Kabushiki Kaisha (Uni-Charm Corporation) [2007] 1 SLR 1082 at [9] (“Johnson & Johnson”).

  42. [17] The comparison to be made is one of “mark for mark”. Each mark must be considered as a whole without any external matter being taken into account: City Chain at [50]; Caterpillar Inc v Ong Eng Peng (formerly trading as Catplus International) [2006] 2 SLR(R) 669 (“Caterpillar”) at [55].

  43. [18] The relevant perception is that of an “average consumer” who is “reasonably well-informed and reasonably observant and circumspect”: Chai Chyau Ling (doing business as Racetech Auto) v Racing Technology Pte Ltd [2009] SGHC 105 (“Racetech Auto”) at [30], citing Love & Co Pte Ltd v Carat Club Pte Ltd [2008] SGHC 158 and the European Court of Justice’s decision of Lloyd Schuhfabrik Meyer & Co GmbH v Klijsen Handel BV [2000] FSR 77. That said, to determine similarity, the sign and the mark ought not to be examined side by side in close proximity as the “imperfect recollection” of the consumer must be taken into account, given that comparisons are usually made from memory, removed in time and space from the actual sign and mark in question: Caterpillar at [55], citing M Ravindran, Butterworths’ Singapore Statutes, Intellectual Property, Trade Marks Act – A Commentary (LexisNexis, 2002 Ed) at p 42); Johnson & Johnson at [11].

  44. [19] It appears that the distinctiveness of a trademark is an important factor to bear in mind: Polo (CA) at [17] and [23]; Ng-Loy Wee Loon, Intellectual Property Law in Singapore (2009, Rev ed, Sweet & Maxwell) (“Ng-Loy”) at para 21.5.12. The question of whether a sign and a mark are similar will often be dependant on the inherent or acquired distinctiveness of the trademark: Bently & Sherman at p 866.

  45. [20] On a related note, we take this opportunity to clarify the general approach to be adopted in the similarity analysis. The two-step approach articulated by the Judge (see GD at [47]), following that set out in Ozone Community (at [40]-[44]), may inadvertently give rise to the misconception that the distinctiveness of the registered trademark is either another aspect of, or an element in, the determination of similarity, separate from the three aspects of similarity, viz, visual, aural and conceptual. Rather, we would explain that the “distinctiveness” of the registered trademark is a factor to be considered in the visual, aural and conceptual analysis to determine whether the allegedly infringing sign and the trademark are similar. It stands to reason that the more distinctive the registered trademark, 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 trademark (see below at [35]).

  46. Whether there is visual, aural and conceptual similarity of the “Nutello” sign and the “Nutella” mark

  47. (a) Visual similarity

  48. [21] First, the Appellant submits that the “Nutello” sign is dissimilar due to the different font, typeface, design and also the differences in the letters “a” and “o” used. It also argues that the visual differences are more pronounced if a comparison is made between how the sign and the word mark are actually used in practice, in contrast to the registered form of the word mark.

  49. [22] On the other hand, the Respondent avers that the Judge was correct in his approach that in determining visual similarity of word marks, the court ought to look at the length of the marks, structure of the marks (viz, whether there are the same number of words), and whether the same letters are used: Ozone Community at [49], citing Bently & Sherman at p 865.

  50. [23] In the present case, the length of the “Nutello” sign and the “Nutella” word mark (comprising seven letters each) and the structure of the sign and the mark (comprising one word each) are the same. The letters used in both words are also the same save for the last letter, the former ending with “o” while the latter ending with “a”. We agree with the Judge that the case in point is that of Hyundai Mobis v Mobil Petroleum Company Inc [2007] SGIPOS 12 (“Hyundai Mobis (IPOS)”) which involved marks of similar characteristics. The marks considered there were “Mobis” and “Mobil”. Both marks had the same length, structure and had identical letters save for the last letter. The Presiding Assistant Registrar (“PAR”) there held (which the judge in that case did not disagree with) that when viewed as a whole, while the ending parts of the marks were dissimilar, a large part of the marks was in fact similar. He therefore found visual similarity. Similarly, this analysis can also apply to the present case given that the letters used are substantially identical and the identical length and structure of the “Nutello” sign and the “Nutella” mark. Further, the effect of imperfect recollection of consumers cannot be discounted, particularly since the difference between the words here is merely a single ending letter. As stated in Ozone Community (at [50]) citing Bently & Sherman at p 865, “where at first glance the sign and the mark may appear dissimilar, when the possibility of imperfect recollection is taken into account, they may in fact appear similar.” Therefore it would not be difficult to conclude that the “Nutello” sign and “Nutella” mark would likely appear visually similar to an average consumer, given a difference of only the final letter.

  51. [24] Additionally, we think as the Judge did, that the argument that the cursive font, typeface and design used in the sign render it visually different from the mark is misconceived. The Judge followed the reasoning in Richemont International SA v Da Vinci Collections Pte Ltd [2006] 4 SLR(R) 369 (“Richemont”) (at [20]-[24]), which relied on Morny Ld’s Trade Marks, in the Matter of (1951) 68 RPC 131 (“Morny”), to hold that for registered word marks, use of the word in any font or stylisation would be infringing. In Morny, the “Morny” mark was registered in plain block capital letters. Jenkins LJ in Morny (at 149-150) observed that the registration of a word in block capitals ought to cover use of that word in any clearly legible form of lettering and must not be confined to its representation in block capital letters. The reason was that if it were otherwise, protection of a word registered in block capitals would be illusory “for anyone could use the same word in some other form or type or lettering in imitation handwriting or some other form of that kind”, and it would also be “obviously ... impossible for anyone seeking to register a particular word … to cover in his application ... every conceivable form of lettering in which it might be represented.” The “Nutella” word mark here is registered in bold block capital letters and not in any fancy or stylised form (see [3] above). Following the reasoning in Morny, regardless of the font, typeface or design of the “Nutello” sign used by the Appellant (here, it is presented in stylised form and in sentence case), the mark and the sign are still similar.

  52. [25] We would further add that the word “Nutella”, being an invented word, is in itself distinctive. Where the distinctiveness resides in the word itself, differences in presentation of the word mark, e.g., in font type and colour, are immaterial: Racetech Auto at [39]; Richemont citing Kerly, Law of Trade Marks and Trade Names (Sweet & Maxwell, 14 Ed, 2005) at para 2-036. Therefore, given the inherent distinctiveness of the word “Nutella”, the Appellant’s argument that the “Nutello” sign is dissimilar to the Respondent’s word mark because of differences in stylisation and font type cannot stand.

  53. [26] In light of the preceding analysis, we hold that the “Nutello” sign and the “Nutella” mark are visually similar.

  54. (B) Aural similarity

  55. [27] Secondly, the Appellant submits that the “Nutello” sign and the “Nutella” mark are aurally dissimilar. It argues that in contrast to the clear pronunciation of “Nutello”, the pronunciation of “Nutella” is inconsistent given that the pronunciations would in fact be Noot-ella and Nut-ella. Also it argues that the average customer would focus on the differences between the sign and the mark, viz, the ending syllables –ella and –ello, and that aural dissimilarity would be enhanced with the accent on the different vowel sounds at the end of the respective words. Against this, the Respondent submits that the sign and the mark are aurally similar because the average Singaporean would pronounce “Nutella” as Nut-ella and not Noot-ella, and that this pronunciation is similar to how “Nutello” is pronounced. It also emphasises the importance of the first syllable in pronouncing the words, and that the vowel sound in the last syllable (i.e., “o” and “a”) in the respective words is insufficient to conclude aural dissimilarity.

  56. [28] A quantitative assessment of the relative number of syllables which the two marks have in common is usually conducted in ascertaining aural similarity in word marks: Ozone Community at [55], citing Bently & Sherman at p 865. Here, there are three syllables in both the words, “Nutello” and “Nutella”, with the majority (viz, the first two out of three) in common.

  57. [29] The Judge held that it was likely that it would be the first syllable of the words “Nutella” and “Nutello” that was emphasised. The Appellant’s argument that customers will focus on the ending syllables of both words (viz, –ella and –ello) can hardly be correct. The Appellant relied on the decision in Hyundai Mobis (IPOS) to dispute the significance of the first syllable of word marks. However, in that case, despite the PAR finding that the second syllable would be emphasised in pronouncing “Mobil” and “Mobis”, he still found the second syllable in both marks to be aurally similar (see Hyundai Mobis (IPOS) at [40]). Therefore, even if the emphasis is placed on different ending syllable(s) of a multi-syllabic word, it does not mean that this would necessarily point to aural dissimilarity. For the present matter, we agree with the Judge that the first syllable would likely be emphasised in pronouncing the words “Nutella” and “Nutello”. It would do violence to the natural pronunciation of both words to say, as the Appellant does, that the ending vowels would be accented. We therefore find no reason to disagree with the Judge’s observation that there is greater aural similarity between the two words which end with vowels instead of consonants (see GD at [63]).

  58. [30] It is also not incorrect for the Judge to have considered how an average Singaporean consumer would pronounce the respective words. No new arguments are raised by the Appellant as to why pronunciation of the words in question by Singaporeans ought to be anything other than “nut-ella” and “nut-ello” as found by the Judge (see GD at [62]). In Hyundai Mobis (IPOS), the PAR noted that there was a risk of the ending consonant of the marks there being slurred in speech in Singapore. Although the respective words “Nutello” and “Nutella” here end with vowels, it is difficult to see how the conclusion can be any different, as far as slurring is concerned.

  59. [31] The Judge (at [64] of his GD) also rightly made allowances for imperfect recollection and careless pronunciation and speech: Future Enterprises Pte Ltd v McDonald’s Corp [2006] 4 SLR(R) 629 at [12], citing Aristoc, Ld v Rysta, Ld (1945) 62 RPC 65 at 72-73 (“Aristoc”). It was stated in Aristoc, Ld v Rysta, Ld (1943) 60 RPC 87, and affirmed by the House of Lords in Aristoc at 72, that first impressions mattered in determining aural similarity between words. Indeed, it is plausible for a consumer, who has an imperfect recollection, to regard the word “Nutello” to be similar to “Nutella” in their ears.

  60. [32] Ultimately, it is the pronunciation of the relevant words as a whole that must be critical in ascertaining aural similarity: Festina Lotus SA v Romanson Co Ltd [2010] 4 SLR 552 (“Festina”) at [48]. For the reasons above, we therefore agree with the Judge that the “Nutello” sign and the “Nutella” mark are aurally similar.

  61. (C) Conceptual similarity

  62. [33] Third, the Appellant argues that conceptual similarity is absent between the “Nutello” sign and the “Nutella” mark because both are invented words conveying no meaning. It also avers that the Judge had not identified the concept underlying the sign and the mark in question. The Respondent agrees with the Judge that both words are meaningless as they are invented ones, and since “Nutello” was derived from the word “Nutella”, there is conceptual similarity between them.

  63. [34] As was rightly stated by the Judge, conceptual similarity involves the consideration of the “ideas that lie behind or inform the earlier mark”: Festina at [39], citing Bently & Sherman at p 866. However we have difficulties with his finding that there is conceptual similarity between the “Nutello” sign and the “Nutella” mark. His reliance on the authority of Festina to arrive at his conclusion may not be wholly appropriate. In Festina, the “Festina” and “J.Estina” marks comprised invented words and had a pictorial element of a crest and crown respectively. The judge found conceptual similarity between the marks because the respective crest and crown devices indicated a sense of “class and “status” (see Festina at [45]). In the present case, however, although the words “Nutello” and “Nutella” are invented and meaningless like those words in Festina, the “Nutello” sign and “Nutella” mark here do not each have a device element which evokes similar ideas. Instead, we think that the case of Hyundai Mobis (IPOS) is of greater relevance to the present. There, the marks in question were word marks and comprised invented words, similar to the situation in the present case. The words “Mobil” and “Mobis” were invented. The PAR observed that while the applicants there argued that “Mobis” was derived from the words “mobile” and “system”, it was not obvious when one sees the word “Mobis” that it meant so. He therefore concluded that no finding of conceptual similarity could be made because both the “Mobil” and the “Mobis” marks were meaningless. Similarly, following this reasoning, since the words “Nutello” and “Nutella” are invented and meaningless with no particular idea underlying each of them (and nothing has been shown in that regard) it is difficult to say that they are conceptually similar. The Judge seemed to think (at [69]) that just because “Nutello” was derived from “Nutella” with only a change to the last alphabet, as the Appellant’s witness admitted, this suggests that there was conceptual similarity. We do not see how that follows. While these circumstances certainly justify a finding that there is similarity in the visual and aural sense, they do not necessarily suggest a conceptual similarity. The Judge did not indicate what the concept was. They are just meaningless words and do not evoke any ideas. With respect, on this issue, we hold that the Judge is in error.

  64. Distinctiveness of the “Nutella” word mark

  65. [35] As we have earlier mentioned (see [20] above), the distinctiveness of the earlier mark is a factor to be considered in the similarity analysis. On this point, the Respondent argues in support of the Judge’s finding that as the “Nutella” mark is inherently distinctive, being an invented word, merely changing the last letter from “a” to “o” as per the “Nutello” sign would not render the latter sufficiently dissimilar.

  66. [36] In Bently & Sherman at p 866-867, the learned authors state that the question of whether marks are similar will oftentimes depend on the inherent or acquired distinctiveness of the mark for the goods for which it has been registered. Therefore, if the trademark is highly distinctive, it follows that a sign which has been substantially modified may possibly still be regarded as similar. For such trademarks, then, there is a high threshold to be met in creating a sign or a mark sufficiently dissimilar to it. Inherently distinctive marks include those which comprise inventive words without any notional or allusive quality, an example being the word “Volvo”: Polo (CA) at [23]. Here, the “Nutella” word mark is an invented word with no particular meaning to it; it possesses a considerable degree of inherent distinctiveness. This distinctive quality is a factor to be considered in determining similarity between the mark and the “Nutello” sign. As such, we agree with the Judge that merely changing the last letter of the “Nutella” word mark from “a” to “o” to form the “Nutello” sign would not suffice to meet this high threshold to render the “Nutello” sign dissimilar from the “Nutella” word mark.

  67. Overall similarity between the “Nutella” word mark and “Nutello” sign

  68. [37] We have found that the “Nutella” word mark is inherently distinctive and that it is visually and aurally, but not conceptually, similar to the “Nutello” sign. As stated above at [16], a finding of similarity of marks can be made even though not all three aspects of similarity are established. It is the overall picture that emerges to the court which would be decisive. The wisdom of this “trade-off” approach was noted by the High court in Festina (at [53]) as offering “flexibility and latitude to the courts in coming to a conclusion (with regard to the overall similarity of marks) against a contextual background by adapting to the ever evolving commercial reality”.

  69. [38] For the present case, we are of the view that the visual and aural aspects of similarity are relatively more important than conceptual similarity. This is because the “Nutella” mark and “Nutello” sign are word-only and effectively can only be engaged visually or verbally. Moreover, considering the mode of purchase of products sold under the respective mark and sign, viz, the “Nutella” cream spread being bought from supermarkets and the “Nutello” drink being ordered in TCC cafes, visual reading and recognition of the words in the mark and sign are the only ways of selecting or ordering these products. For the “Nutello” sign in particular, the verbalisation of it is needed when one orders the “Nutello” drink in the TCC cafe.

  70. [39] As such, in light of the significance of visual and aural similarity of the mark and sign and the inherent distinctiveness of the former, we find that similarity is established as between them. The first element of s 27(2)(b) TMA is thus made out.

  71. (ii) Similarity of goods

  72. [40] Section 27(2)(b) TMA contemplates that with regards to goods, identity or similarity must be shown. For the purposes of this appeal, we need only consider the question of similarity of the goods. As the Respondent has not appealed against the Judge’s finding that the goods are not identical, this aspect of the Judge’s finding is no longer in dispute.

  73. [41] In finding that the Nutello beverage was similar to the “chocolate products” specification for which the “Nutella” word mark was registered, the Judge explained that the relevant comparison to be made was between the Nutello beverage on the one hand and the actual product in use (i.e., the “Nutella” spread), as well as the types of products in respect of which the respondent has registered its word mark “Nutella” on the other. He held that the comparison ought to be made with the possible range of “chocolate products” as used in the trade besides the respondent’s “Nutella” spread. In finding similarity, he took cognisance of the fact that the main ingredient in the Nutello beverage was Nutella spread and that the Nutello beverage was marketed to attract chocolate lovers, in addition to the manner in which the Nutello beverage was served and consumed. He further noted that a market survey showed that consumers considered the Nutello beverage as being similar to chocolate products, and that the drink would fall within the “beverages” category (e.g. “Beverages (chocolate-based)”; “chocolate-beverages with milk”; “chocolate-based beverages”) which is also within Class 30 ICGS.

  74. [42] The Appellant argues that the parties’ goods are in fact dissimilar. Firstl, it submits that the Judge was precluded from finding similarity because he had already found that the Nutello beverage was not identical to a chocolate product. Secondl, it submits that it is wrong to find similarity of goods simply because the Nutello beverage falls within the same class, viz, Class 30 ICGS, as the “chocolate products” for which the Nutella spread is registered. In support of this, the following reasons are proffered: Class 30 ICGS comprises various different products; Jacob J in British Sugar did not state that the factors enumerated in his judgment (see British Sugar at 296-297 and below at [48]) (“British Sugar factors”) are irrelevant if the goods are of the same class; the term “chocolate products” is too vague and wide, extending even to products which the Respondent has not produced and applied the mark to or which fall into other ICGS classes; and that the respondent intended to refer only to chocolate cream spread when it registered its “Nutella” mark.

  75. [43] However, the Respondent makes the following arguments to the contrary. First, that the correct comparison ought to be between the Nutello beverage and the goods in respect of which the “Nutella” mark is registered for, and that a finding that the goods were non-identical would not preclude a finding that goods were similar. Second, it avers that the Judge did not discount the British Sugar factors in reaching his decision that the goods were similar. It argues that there is similarity of goods since both appeal to customers who enjoy the “cocoa taste”, with the Nutello beverage being made with Nutella spread as an ingredient, and the “Nutella” word mark has been used by the Respondent in similar Nutella-based beverages of its own, although not in Singapore.

  76. [44] At this point we would clarify that, contrary to the Appellant’s first argument, a finding that the goods were non-identical would not preclude a finding that goods were similar. We do not understand why a finding made under the identity limb of s 27(2)(b) TMA should be determinative or conclusive of a finding under the similarity limb. Indeed, the fact that s 27(2)(b) refers to the criteria of “identical” and “similar” disjunctively shows that they could mean different things, and the literal meanings of the words point to that as well.

  77. [45] Section 27(2)(b) TMA requires similarity between the goods and services for which the offending sign is used and “those for which the trade mark is registered”. In Bently & Sherman (at p 869), the learned authors state that in an infringement action, courts usually compare the defendant’s goods (here, the Appellant’s goods) as they have been used with the goods in the claimant’s specification (here, the Respondent’s specification). In doing this, the court has to interpret the specification and then characterise the defendant’s goods or services (here, the Appellant’s goods) to see if they fall within the specification. The Judge held (at [86] of the GD) that the comparison should be made between the Appellant’s Nutello beverage product, and both the actual Nutella product in use (i.e., the “Nutella spread”) as well as the type of products for which the “Nutella” word mark was registered (i.e., the possible range of “chocolate products” as used in trade). This approach is also consistent with the decision of the European Court of Justice in Vedial v OHIM (Hubert) (Case C – 106/03P [2005] ETMR 23), which was a case concerning a proposed registration of a later mark, and where the court held that one should look at the proposed specification of the new mark against the specification of the existing registered mark in determining identity or similarity. While it is true that the judge in British Sugar (at 297) made a comparison between the “Robertson product” and the “British Sugar product”, it seemed to be in relation to ascertaining infringement. It is not clear whether, when making the comparison between the actual goods, the court there was doing so for the purposes of determining similarity of goods or for determining confusion, as the court there ultimately held that there was no likelihood of confusion and in turn there was no infringement.

  78. [46] We agree that the concept of similarity of goods should not be over extended or be interpreted too broadly. In British Sugar (at 296) it was stated that the purpose of the words “similar goods” in s 10(2) of the Trade Marks Act 1994 (c 26) (“UK TMA 1994”) and the Directive on which it is based (namely the First Council Directive 89/104/EEC of 21 December 1988 to approximate the laws of the Member States relating to trade marks) is to prevent marks from conflicting, both for their respective actual goods and also for a penumbra. However, we need to point out that under the express terms of our s 27(2)(b) TMA, which is substantially similar to s 10(2) of UK TMA 1994, the comparison to be made is not between the alleged infringing goods and the actual goods of the trademark owner but between the alleged infringing goods and the products in respect of which the trademark is registered for. This is also the view of Bently and Sherman (see above at [45]). To say that in determining similarity one should only look at the alleged infringing goods and the actual goods of the trademark owner and not at the specifications would be to take a view inconsistent with the express terms of the provision. If that were to be the correct view, what protection is there left for the trademark owner in respect of goods which are within the specifications of the registered mark, but in respect of which the owner has yet to embark upon actual production? Having said that, a late entrepreneur is not without a remedy. He could apply to the Registrar of Trade Marks or to the court for revocation of the registration of the mark in respect of goods which the owner has yet to produce after five years of registration pursuant to s 22(1)(a) TMA . Alternatively, he need not adopt a mark which is similar to the registered mark – there would then be nothing to impede his business venture.

  79. [47] We note that in British Sugar the first issue which Jacob J identified for determination was “Does the Robertson’s product fall within the specification of goods for which the mark is registered?” [emphasis added]. Interestingly and pertinently, Jacob J held that the defendant’s product, “Robertson’s Toffee Treat”, which was a toffee spread, was neither identical nor similar to the specification of the registered trademark, namely “dessert sauces and syrup”. There, the defendant had also raised the defence of revocation on the ground that the word “Treat” was a common English word which was essentially descriptive and not distinctive. This could explain why Jacob J only felt the need to compare the actual product of the trademark owner with the alleged infringing product and where he held (at 297) that there was no infringement. Therefore, we are in agreement with the Judge that to determine similarity of goods in the present case, one must not only compare the goods of the offending party with the actual product of the owner but also with the specification of the mark. As such we are unable to accept the Appellant’s contention that the comparison ought to be made with the registered specification of the “Nutella” word mark as interpreted by its use in practice in the form of the Nutella spread, as this is in substance tantamount to submitting that the comparison ought to be made with the actual Nutella spread product itself.

  80. [48] We do acknowledge that the “chocolate products” specification could, on a plain reading of the term, cover a whole range of foodstuff and items. It could include things like chocolate powder, mocha, or even items like chocolate-flavoured body lotion. However, it is not true that simply because of this holding the Respondent would ipso facto have a monopoly of the “Nutella” word, or closely related variations, over what is an extensive range of all chocolate products imaginable. This is because a registered owner still needs to establish further the third element, namely, likelihood of confusion, in order to make out a successful case of trademark infringement.

  81. [49] It was held in British Sugar that the following are relevant considerations in determining the question of similarity of goods:

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

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

  84. (c) The physical nature of the goods or acts of service;

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

  86. (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; and

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

  88. [50] At this juncture, we pause to address the Appellant’s argument that the Judge erred in relying upon the views expressed by the High Court in The Polo/Lauren Co, LP v Shop In Department Store Pte Ltd [2005] 4 SLR(R) 816; [2005] SGHC 175 (“Polo (HC)”) (at [33]) that where goods fell within the same class in the ICGS classification system, the goods ought very likely be regarded as similar. In particular, we highlight the following passage:

  89. [51] The Judge held that since the “Nutello” drink was a beverage falling within Class 30 ICGS (which is the same class as that for which the “Nutella” mark was registered), he was persuaded therefore to find that there was a similarity of goods. While this appears to have a certain logical attractiveness, it should not be determinative – consideration should also be given to the factors enumerated by Jacob J in British Sugar (see above at [48]). Indeed, in Bently & Sherman (at p 784), the learned authors have stated that the purpose of the classification process is primarily administrative, viz, to facilitate searching for earlier competing marks. That was also seemingly the view of Jacob J. As regards the factors set out in British Sugar, ultimately, the weight which ought to be accorded to each factor is for the trial judge to ascribe after assessing all pertinent factors. In other words, it is not a numbers game.

  90. [52] The difficulty with applying the British Sugar factors to the present case is that we need to compare the Appellant’s product not only with the actual Nutella product (i.e. the Nutella spread) but also with the specification of the registration, i.e., “chocolate products”. Making only a narrow comparison between the two actual products, the following observations could be made:

  91. (a) Uses: As Jacob J noted in British Sugar (at 297), it is the use “in practice” of the goods that this factor is concerned with, not the target effect of marketing as the respondent has contended. Here, the respective ordinary uses of the “Nutella” and the “Nutello” products are different. The former is a cream spread typically applied on and consumed with bread whereas the latter was a gourmet beverage consumed directly from the shot glass in which it was served.

  92. (b) Users: On the facts this factor is neither here nor there since the same person could be the user of both products. For instance, professionals who frequent the Appellant’s outlets to consume the Nutello beverage may also use the Nutella spread during breakfast.

  93. (c) Physical nature: Likewise, this factor is neither here nor there because while it is undisputed that the Nutella spread is of a viscous nature, it appears from the evidence that the Nutello beverage itself, into which Nutella spread was put was similarly of a viscous nature and required the use of a spoon during consumption.

  94. (d) Trade channels: As the Appellant rightly submitted, the trade channels of the Nutella spread and the Nutello beverage are different. The former is commonly sold in supermarkets and convenience stores whereas the latter was sold exclusively at the Appellant’s TCC cafés.

  95. (e) Substitutability: The Nutella spread and the Nutello beverage can hardly be said to be in direct competition. They are not product substitutes. The former competes with other bread spreads whereas the latter competes with other beverage products.

  96. [53] Having said that, if the proper comparison to be made is between the Nutello beverage and the “chocolate product” registration specification, which we think is the case, then many of the British Sugar factors may not be applicable as they were premised on comparing two actual products. In so far as the specification of the “Nutella” mark extends to products which have not yet been produced by the Respondent, then the British Sugar factors may not be entirely helpful in the conduct of the similarity analysis. At the end of the day, the question would depend on whether the Nutello beverage can be considered a “chocolate product”. From the evidence from the market survey and internet blogs, various consumers have commented on the Nutello beverage's chocolate nature (such as its chocolate taste and chocolate content). This indicates rather persuasively that the Nutello beverage has in practice been regarded as a "chocolate product" by those who have consumed it. While not conclusive, it shows that the Nutello beverage can be considered to fall within the "chocolate product" specification. Bearing in mind that the “chocolate product” specification is not only restricted to the traditional solid form of chocolate products, but also encompasses all forms or textures such as powder and even viscous liquids/viscous products like the Nutella spread and the Nutello beverage, the similarity is obvious. It seems to us that one reason the Judge refrained from holding that the “Nutello” product was a chocolate product (i.e., identical to a chocolate product) was essentially because under the same Class 30 of ICGS there are the following separate categories of goods: “Beverages (Chocolate-based)”, “Beverages (Cocoa-based)”, “Chocolate beverages with milk” and “Chocolate-based beverages”. Although the Respondent could have done so, it did not register the “Nutella” mark against these categories of products under the same class. In the result, we agree with the Judge that the Nutello beverage is not identical but similar to “chocolate products”. It seems to us that these products (viz “Beverages (Chocolate-based)”, “Beverages (Cocoa-based)”, “Chocolate beverages with milk” and “Chocolate-based beverages” et al), which are listed under the same Class 30 ICGS as “chocolate products”, are simply further particularisation of “chocolate products” and are thus similar to “chocolate products”. Of course, had the Respondent registered the “Nutella” mark against “Beverages (Chocolate-based)” or “Chocolate-based beverages”, it would probably have been a case of “identical” rather than “similar” products. Therefore, in light of the preceding analysis, we are of the view that the “Nutella” and “Nutello” products are similar.

  97. (iii) Likelihood of confusion

  98. [54] We now turn to consider the third element to establish infringement: likelihood of confusion. The Judge found a likelihood of confusion because of the great similarity of the marks and the goods respectively, the distinctive nature of the “Nutella” word mark and the evidence showing that approximately 30% of the relevant public surveyed had indicated that they were confused.

  99. [55] The Appellant argues that the relevant test is confusion of “a substantial number of the relevant public” and that principles of passing off should apply to determine whether this element of confusion is established. It submits that the “relevant public” should comprise the Appellant’s actual and potential customers. It adds in relation to this that the number of people must be substantial, and that there would be a real impact on the Respondent’s trade or goodwill. As the Respondent had not shown weak or negative sales, but to the contrary achieved a 20% increase of sales each year, the Appellant argues that there could not have been any confusion. It also argues that the Judge was overly reliant on survey evidence in finding that there was confusion and had failed to consider other circumstances and the requisite “extraneous” factors. As to what these “extraneous” factors are, this will be elaborated below at [60]. On the other hand, the Respondent submits that passing off principles are irrelevant to the interpretation of the likelihood of confusion element in trademark infringement. In any event, the Respondent argues that “extraneous” factors should play no part in the confusion analysis.

  100. [56] Before determining whether there is a likelihood of confusion, there are two legal principles which we need to address briefly. Both are germane to the issue of confusion. The first concerns the term “substantial portion (of the relevant public)” and the second relates to the “extraneous” factors and whether they ought to be taken into consideration in determining confusion.

  101. [57] It is settled law that the test to be adopted in determining likelihood of confusion is whether a substantial portion of the relevant public will be confused (see Hyundai Mobis (CA) (at [77]-[82] and GD at [92]). As regards what constitutes a “substantial portion” of the relevant public, this court had the occasion to deal with an identical set of words found in s 8(3)(b)(iii) TMA concerning trade mark opposition proceedings in Hyundai Mobis (CA) (at [77]-[82]). The identical words were “there exists a likelihood of confusion on the part of the public”. We agree with the Judge that the essence of this requirement is that there must not be an insubstantial number of the relevant public being confused. This standard is above de minimis and must be appreciable, though it is not necessary to show that a majority of the public is confused. It is insufficient, however, if only a “single member” of the relevant public is confused or if only a “very small and unobservant section” is confused: Hyundai Mobis (CA) at [79].

  102. [58] It is of little assistance for the Appellant to argue that evidence of rising sales of the Respondent’s Nutella spread suggests indirectly that there was no confusion. While the rise in the latter’s sales can indicate a lack of confusion, it is not necessarily the case, especially in circumstances like the present. It must be borne in mind that in the context of this case it is not alleged that the Nutella spread which the Respondent put into its product (i.e., the Nutella beverage) is not genuine Nutella spread. The confusion in question relates to the relevant public being made to labour under the impression that the Appellant was somehow related to or had some business connection with the Respondent. This being the nature of confusion, there was no reason for the Respondent’s sales to trend downwards. This aside, it does not follow as a general proposition that the existence of confusion must necessarily lead to a fall in sales. Neither would it mean that an absence of confusion would necessarily translate into observable sales increment. There may be other external and/or commercial factors which could have spurred the high sales achieved by the trademark owner, as in this case.

  103. [59] In City Chain, which also involved a s 27(2) TMA trademark infringement action, this court explained in some detail who would constitute the “relevant public”. Where a product is commonly available and purchased by the general public, the average consumer would be the general public: City Chain at [56]. Even so, the court will nonetheless try to ascertain in such cases who the target consumers really are: City Chain at [59]. In the present case, the “relevant public” would be the general public as both the Nutella spread and the Nutello beverage are available generally for the public to purchase and consume. As for the target consumers of the Nutello beverage, they would likely be sophisticated executives, professionals and business people; while those for the Nutella spread would likely be shoppers in supermarkets and provision stores in general.

  104. [60] The term “extraneous factors” has been used to encompass matters outside and beyond the marks and the goods themselves: see Ng-Loy Wee Loon, “Developments in Singapore Trade mark Law 2006-2010 – Confusion-based Protection and Beyond” in Yeo Tiong Min, Hans Tjio and Tang Hang Wu gen eds, SAL Conference 2011: Developments in Singapore Law Between 2006 and 2010 – Trends and Perspectives (Academy Publishing, 2011) at p 354. It includes steps taken by the defendant to differentiate his goods from that of the claimant registered proprietor. The question therefore is whether these “extraneous factors” should be taken into account to determine whether there was in fact a likelihood of confusion on the facts of each case. The author was of the view that such “extraneous factors” should be excluded from the confusion analysis. Understandably, this view was supported by the Respondent. The reasons given by the author were as follows: firstly, statutory interpretation of s 27(2)(b) TMA does not contemplate including such factors into the confusion analysis; secondly, the nature of protection in passing off actions where these factors are relevant is different from that under TMA; and thirdly, the policy reason of certainty for the business community would be advanced if such factors were excluded. The two English authorities relied on were Saville Perfumery Ltd v June Perfect Ltd (1941) 58 RPC 147 (“June Perfect”) and Julius Samann ltd v Tetrosyl Ltd [2006] FSR 42. However, the author was also cognisant of the fact that this Court had adopted a contrary position in Polo (CA). The Appellant here took the position as in Polo (CA). The reasons given for including “extraneous factors” in the confusion analysis are as follows: firstly, extraneous factors used to be considered under the Trade Marks Act 1939; secondly, the English High Court decision of In the matter of an Application by the Pianotist Co Ld for the Registration of a Trade Mark (1906) 23 RPC 774 (“Pianotist”) called for “all the surrounding circumstances” to be considered; and thirdly, the underlying aim of trade mark law is ultimately to prevent confusion. While we recognise the force of the arguments above, we think that, on balance, the better view is that “extraneous factors” should be included in determining if there is a likelihood of confusion on the facts. This was the view of this court in Polo (CA). The observation was made at [25] that if Parliament had intended that upon establishing similarity in marks and goods there would immediately be a finding of confusion, Parliament would have used the phrase “there shall be deemed to be confusion”. There would then be no question of there being a three-step approach to establishing trade mark infringement. We should mention that the judge in Polo (HC), having examined similar arguments before him, came to the same conclusion as well. In his view, under s 27(2)(b) TMA, the likelihood of confusion must arise from the similarity of goods and marks. The judge further observed that where steps had been taken to distinguish the goods, the likelihood of confusion can become merely hypothetical or speculative. As to the argument that the nature of protection under passing off and the TMA is different, the judge in Polo (HC) (at [23]) emphasised that the difference between the regimes was that for TMA infringement actions, confusion must result from the similarity of marks and goods and only then can the court proceed to examine if the likelihood of confusion was real. We would also add that simply because the nature of protection under both these regimes is different, this does not lead to the conclusion that a consideration of “extraneous factors” in TMA infringement actions must necessarily be erroneous. The judge in Polo (HC) (at [19]) noted that the English authorities appeared to differ on the correct approach. He referred to, inter alia, June Perfect and Pianotist. In Pianotist, Parker J held (at 777) that the court should, in determining confusion, consider “the nature and kind of customer who would be likely to buy those goods”. In June Perfect, which omitted to discuss the earlier authority of Pianotist, Sir Wilfred Greene MR said that the appropriate approach depended on the goods in question. In contrast, the judge in Polo (HC) (at [20]) noted that Singapore cases had always clearly and consistently included the “extraneous factors” in the analysis. The examples given were decisions of this Court in Kellogg Co v Pacific Food Products Sdn Bhd [1998] 3 SLR(R) 904 and McDonald’s Corp v Future Enterprises Pte Ltd [2005] 1 SLR(R) 177. Recent decisions by this court in City Chain and Hyundai Mobis (CA) continue to follow this approach. Relatedly, we note that some commentators were also in favour of including “extraneous factors” in analysing whether there was a likelihood of confusion on the facts of the Polo (HC) case: Ng-Loy Wee Loon and Tan Tee Jim, Intellectual Property Law Chapter in the Annual Review of Singapore Cases (6, 2005) at para 16.62.

  105. [61] At this juncture, we would pause to briefly touch on the policy considerations in relation to trade mark law. There appear to be two competing policy concerns here, that of preventing confusion on the one hand and promoting business certainty on the other. Additionally, the need to guard against the danger of creating a monopoly in the trademark for the registered trade mark proprietor which extends protection beyond what is necessary and fairly required in the circumstances must be kept in mind. As the judge in Polo (HC) observed (at [19]), protection offered to a registered trade mark proprietor is “wide but is not indefinite”. In the final analysis, we agree that the main concern is to ensure that consumers do not get confused as to trade sources. This view was very pertinently stated in Polo (HC) at [19]:

  106. [62] Therefore, the court has to take a holistic view of all the circumstances – including the “extraneous factors” – in order to determine whether a likelihood of confusion can be said to exist on the facts of each case.

  107. [63] Confusion will not be presumed simply because the marks and the goods are similar: Polo (CA) at [25]. As this court stated in Polo (CA) at [28], citing Associated Newspapers Ltd v Express Newspapers [2003] FSR 51, the question of likelihood of confusion has to be assessed globally, taking into account all the circumstances such as the closeness of the goods, the impression given by the marks, the possibility of imperfect recollection and the risk that it may be believed that the goods came from the same source or economically-linked sources. Mere association by the relevant public between two signs based on their similar use is insufficient to establish confusion if the public is not misapprehended about the goods’ origins: City Chain at [58]. The finding of likelihood of confusion is a finding of fact and it is a matter for the court to determine at the end of the day: Electrolux Limited v Electrix Limited (1954) 71 RPC 23 at 31.

  108. [64] Survey evidence is relevant in assessing whether there is a likelihood of confusion on the part of the relevant public. That said, we agree with the Judge that survey evidence should not be conclusive (see [129] of GD). Rather, it is only one factor in the global confusion analysis. Both the Appellant and the Respondent here had conducted surveys to ascertain the proportion of the relevant public who were confused. Interestingly, each of the parties’ surveys showed that about 30% of the relevant public were likely to be confused, after adjustments were made for shortcomings in survey methodology and data collection, and the Judge found so accordingly. The Appellant was not able to furnish any compelling reason to say that this determination of the Judge is wrong. By any standard, 30% cannot be said to be insignificant, and indeed we would hold that that constitutes a substantial portion of the relevant public (see at [57] above). Due weight ought to be given to these results in the final analysis.

  109. [65] We should also add that given the degree of similarity between the mark and the sign (see [39] above) and the highly distinctive nature of the respondent’s “Nutella” word mark (see [36] above), these circumstances would certainly go no small way in confusing the relevant public. As stated in Holyoak & Torremans, Intellectual Property Law (Oxford University Press, 6 Ed, 2010) (“Holyoak”) at p 437, there would generally be an increased likelihood of confusion where a mark is highly distinctive.

  110. [66] We recognise that there are “extraneous” factors which seem to weigh against the finding of a likelihood of confusion, such as the differences in retail outlet locations of the parties goods, purchasing process and product packaging. That said, there are other considerations which lean toward the opposite finding. These include the closeness of the parties’ goods as consumption foodstuff, the impression given by the marks that the goods are related since the words “Nutello” and “Nutella” are substantially similar, and that some portion of the relevant public had actually believed that the goods came from the same source or that authorisation had been given to the Appellant to use the “Nutello” name for its drink. Further, we would add that where confusion is in the sense of the parties being related or having a business link between them, the fact that the parties’ goods are sold through different channels would not be very significant.

  111. [67] In light of all the circumstances, we are of the view that there existed a real likelihood that a substantial portion of the general public and/or the target consumers of these goods would be confused in the sense that they would think that the parties are related or have business links between them. As this Court had observed in Polo (CA) (at [35]), quoting Laddie J in Wagamama Ltd v City Centre Restaurants Plc [1995] FSR 713 (at 732), whether there has been trade mark infringement has been said to be more of “a matter of feel than science”. Ultimately, this whole question of confusion is really a matter of perception.

  112. [68] In sum then, we find that the s 27(2)(b) TMA claim is made out.

  113. The Second Issue: Whether there is well known trade mark infringement under s 55(2) TMA?

  114. [69] Section 55 TMA is concerned with well known trademarks under which the trademark proprietor would be entitled to obtain relief by injunction to restrain the use by anyone else of an identical or similar mark, or the essential part of which mark is identical or similar to the well known mark. The Appellant does not dispute that “Nutella” is a well known trademark.

  115. [70] The Judge found, on the facts, that there was infringement by the Appellant under s 55(2) TMA. The Appellant argues that the Judge omitted to consider that part of the provision which concerns the “essential part” of the Appellant’s mark being similar to the “Nutella” trademarks. It submits that the Judge had simply adopted the reasons he used in the s 27(2)(b) TMA infringement analysis to conclude that there was infringement under s 55(2) TMA despite the fact that these two sections are worded differently, viz, s 27(2)(b) TMA lacks reference to an “essential part” of the alleged infringing mark whereas s 55(2) TMA does refer to this. We find that this point has no merit. First, as we see it, the slight difference in wording between these two sections could be due to an oversight during legislative drafting. We will explain. The legislative wording of s 55(2) TMA is modelled after Article 4(1)(b) of the World Intellectual Property Organisation’s Joint Recommendations Concerning Provisions on the Protection of Well known Marks (“Joint Recommendations”). As was noted in Novelty Pte Ltd v Amanresorts Ltd and Another [2009] 3 SLR(R) 216 (“Amanresorts”) (at [163]), Article 4(1)(b) of the Joint Recommendations is an “amalgam of Article 16 of the Agreement on Trade-Related Aspects of Intellectual Property Rights, the law of dilution in American jurisprudence and, the European Union law relating to trade marks with a reputation as set out in the First Directive 89/104/EEC of the Council, of 21 December 1988, to approximate the Laws of the Member States Relating to Trade Marks”. In contrast, the legislative wording of s 27(2)(b) TMA was closely modelled after the UK TMA 1994. That the legislative wording of s 27(2)(b) TMA and s 55(2) TMA respectively was adopted from two different sources without much amendment perhaps explains the discrepancy.

  116. [71] Secondly, and more importantly, we observe that the phrases “any trade mark” and “an essential part of (that mark)” are separated by the disjunctive word “or”. As such, assessing the “essential part” of the trademark in the s 55(2) TMA analysis is only at best an alternative and not a mandatory requirement. In any event, in the context of this case an analysis under the alternative basis does not arise as the Appellant’s mark which is under attack is the entire word “Nutello”, which is similar to the Respondent’s mark “Nutella”. Next, the Appellant highlights the fact that some of the “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 “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 trademark 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 trademark, 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.

  117. [72] The Appellant also argues that the visual and conceptual dissimilarity between the “Nutello” sign and the “Nutella” device marks would obviate any likelihood of confusion. Given our observations of the visual impact of the two marks here being relatively significant (see [71] and [38]) and our finding made earlier of similarity of goods (see [53] above), there is no need to engage this argument further.

  118. [73] Accordingly, we would affirm the finding of the Judge that infringement under s 55(2) TMA has been made out.

  119. The Third Issue: Whether there is infringement by damaging connection of well known trade mark under s 55(3)(a) TMA?

  120. [74] The Appellant argues against the Judge’s finding that there was infringement by damaging connection under s 55(3)(a) TMA. To establish a case under this provision, the proprietor of a well known trade mark must show the following: first, that the defendant’s mark, or an essential part of it, is identical with or similar to the proprietor’s mark; second, that the use by the defendant of its mark would indicate a connection between the goods or services of the defendant with the proprietor; and third, that because of the connection the interests of the proprietor are likely to be damaged. Here, we ought to clarify that unlike the position under s 27(2), the well known trade mark proprietor need not show that the goods or services of the defendant are similar to those of the proprietor.

  121. [75] The Appellant’s main points of contention are twofold. First, it reiterates the argument on confusion i.e., that the “Nutello” sign, or its essential part, is not similar to the “Nutella” mark. As we have already dealt with this point at [71] above, we do not think there is a need to traverse the same arguments. Suffice it to say that there is no merit in the argument. Indeed, because of the similarity between the sign and the mark, it suggests to the mind of the average consumer that there is a connection between the Nutello beverage and the Respondent’s product, i.e., the Nutella spread. This points to a confusing connection. Second, the Appellant contends that there is nothing to show that the interests of the Respondent would be damaged on account of that connection. In this regard, the Appellant highlights the fact that the evidence from the Respondent shows that its sales in Singapore increased by a “whopping 20% from each preceding year”. Neither is there any evidence that the Respondent’s bottom line had been affected during the three years that the Appellant had used the mark “Nutello” for its drink, namely, August 2007 to July 2010.

  122. [76] As regard the “connection” and the “likely to damage the proprietor’s interests” requirements in s 55(3)(a) TMA, the Judge adopted the holding by this Court in Amanresorts (at [234]) that the tests for both these elements were “substantively the same” as the tests relating to the misrepresentation and damage elements under the law of passing off; the only difference is that the tests in passing off concern the proprietor’s “goodwill” while the tests under s 55(3)(a) TMA concern the “interests” of the well known trade mark proprietor. It is an implicit requirement that the requisite “connection” must be a confusing one, viz, a likelihood of confusion must be shown in relation to the connection between the parties’ products: Amanresorts at [218], [226] and [233]; Hyundai Mobis (CA) at [36].

  123. [77] The Judge found that there was a confusing connection based on his finding of confusing misrepresentation in the passing off action (see GD at [166]), which was in turn based on his finding of a likelihood of confusion under s 27(2)(b) TMA (see GD at [212]). Given that the test for the “connection” requirement is similar in substance to the test for the misrepresentation requirement in passing off (the findings of which were in turn based on the findings made in relation to the likelihood of confusion element in s 27(2)(b) TMA), and the distinction between the tests (see [76] above) does not result in any effective difference on the present facts, we find no reason to disagree with the Judge that there is the requisite confusing connection under s 55(3)(a) TMA here.

The judgment continues past this page. Read the rest at eLitigation.