Ceramiche Caesar SpA v Caesarstone Sdot-Yam Ltd

Ceramiche Caesar SpA v Caesarstone Sdot-Yam Ltd
[2017] SGCA 30

Case Number : Civil Appeal No 61 of 2016
Decision Date : 26 April 2017
Tribunal/Court : Court of Appeal
Coram : Sundaresh Menon CJ; Andrew Phang Boon Leong JA; Tay Yong Kwang JA
Counsel Name(s) : Pang Sze Ray, Melvin, Nicholas Ong and Nicholas Tong (Amica Law LLC) for the appellant; Prithipal Singh s/o Seva Singh, Chow Jian Hong and Denise Anne Mirandah (Mirandah Law LLP) for the respondent.
Parties :

Trade Marks and Trade Names – Registration criteria – Conflicts with earlier marks

[LawNet Editorial Note: The decision from which this appeal arose is reported at [2016] 2 SLR 1129.]

26 April 2017

Judgment reserved.

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

Background

1       The appellant, Ceramiche Caesar SpA ("the Appellant”), is an Italian company and a manufacturer of porcelain stoneware tiles for indoor and outdoor use, both for flooring and cladding installations. The respondent, Caesarstone Sdot-Yam Ltd ("the Respondent”), is an Israeli company and a manufacturer of engineered quartz and stone surface products for various applications.

2       The Appellant is the registered proprietor of the following trade mark ("the Appellant’s CAESAR Mark”) in Singapore:

The Appellant’s CAESAR Mark was registered for goods in Class 19 as follows:

Class 19

Non-metallic building materials, especially tiles for covering, floor coverings, tiles for gutters, cove moldings, corner beads not of metal, protruding wedges, stilted modular floors, special parts for finishing, stair treads.

3       The Respondent, on the other hand, applied to register the following mark ("the Respondent’s CAESARSTONE Mark”) in Singapore:

The application was specified for registration in Class 19, among others, as follows:

Class 19

Non-metallic tiles, panels for floors, floor coverings, wall cladding, flooring, and ceilings; non-metallic covers for use with floors and parts thereof; non-metallic profiles and floor skirting boards; slabs and tiles formed of composite stone for building panels, counter tops, vanity tops, floors, stairs, and walls.

4       The Appellant objected to the Respondent’s application in Class 19 and commenced opposition proceedings against the Respondent. No objection was raised against the Respondent in respect of the other classes in which registration was sought, these being Classes 20, 35, and 37.

5       Opposition proceedings were first heard before the Principal Assistant Registrar of Trade Marks ("the PAR”). Before the PAR, the Appellant relied on the following four grounds of opposition: (a) s 8(2)(b) of the Trade Marks Act (Cap 332, 2005 Rev Ed) ("TMA”); (b) s 8(4)(a) read with s 8(4)(b)(i) of the TMA; (c) s 8(4)(a) read with s 8(4)(b)(ii) of the TMA; and (d) s 7(6) of the TMA. The PAR allowed the opposition on the first two grounds.

6       Dissatisfied with the PAR’s decision, the Respondent appealed to the High Court in Tribunal Appeal No 12 of 2015. The appeal was in relation to the grounds on which the Appellant had succeeded before the PAR (namely, s 8(2)(b) and s 8(4)(a) read with s 8(4)(b)(i) of the TMA). The Judge allowed the appeal in Caesarstone Sdot-Yam Ltd v Ceramiche Caesar SpA [2016] 2 SLR 1129, finding that both grounds were not made out. The Judge further ordered that the Appellant pay the Respondent the costs of both the hearings before him and the PAR, which were to be agreed or taxed. The Appellant has brought the present appeal against the Judge’s decision.

7       In addition to the substantive issues which arise from the Judge’s decision on the two aforementioned grounds, this appeal also raises the issue of the appropriate threshold for appellate intervention in appeals from the Trade Marks Registry. Specifically, the question is whether there must be a "material error of fact or law” before appellate intervention is warranted. Consequently, the following three issues arise in this appeal:

(a)     whether, in appeals from the Trade Marks Registry, there is a threshold requirement for a "material error of fact or law” to be shown before appellate intervention is warranted ("Issue 1”);

(b)     whether the Judge erred in finding that the opposition under s 8(2)(b) of the TMA was not made out ("Issue 2”); and

(c)     whether the Judge erred in finding that the opposition under s 8(4)(a) read with s 8(4)(b)(i) of the TMA was not made out ("Issue 3”).

8       In relation to Issue 1, the Appellant submits that decisions of the registrar ought not to be disturbed unless there has been a "material error of fact or law”, and that in the instant case the PAR did not make any material error. The Respondent, on the other hand, submits that the threshold for appellate intervention cannot be uniformly high in all appeals from the Trade Marks Registry, and that the question of whether appellate intervention is warranted must depend on the nature and facts of each case. On Issues 2 and 3, the Appellant unsurprisingly urges this court to reverse the Judge’s findings while the Respondent, similarly unsurprisingly, urges us to affirm the same.

Issue 1: Threshold for appellate intervention

9       Issue 1 relates to the threshold for appellate intervention in an appeal from the registrar’s decision. The Judge observed that there was "some uncertainty” over this question. He framed this uncertainty as follows:

On one view, this appeal operates by way of a rehearing and that the court may consider afresh all the evidence adduced at the hearing below but the decision of the PAR ought not to be disturbed unless there has been a "material error of fact or law” … On another view, the threshold for appellate intervention may be lower in view of O 87 r 4(2) of the Rules of Court (Cap 322, R 5, 2014 Rev Ed) which provides that "an appeal shall be by way of rehearing and the evidence used on appeal shall be the same as that used before the Registrar and, except with the leave of the Court, no further evidence shall be given” …

It seems to us that the difference, to the Judge, turned on whether it had to be shown that there was a "material error of fact or law” in the registrar’s decision before the court could interfere on appeal. Ultimately, the Judge did not think that he needed to decide on which approach was correct because he took the view that nothing turned on what the threshold for appellate intervention was.

10     On appeal to us, the Appellant submits that the cases and "normative grounds” support the threshold requirement of a "material error of fact or law”. The main case relied on by the Appellant is our decision in Future Enterprises Pte Ltd v McDonald’s Corp [2007] 2 SLR(R) 845 ("Future Enterprises”), which has been followed by the High Court in Festina Lotus SA v Romanson Co Ltd [2010] 4 SLR 552 (at [21]) and Rovio Entertainment Ltd v Kimanis Food Industries Sdn Bhd [2015] 5 SLR 618 ("Rovio”) (at [78]–[80]). As for the "normative grounds”, the Appellant submits that three key factors support a high level of appellate deference: (a) the highly subjective nature of assessing mark similarity and the likelihood of confusion; (b) the registrar’s specialised experience in dealing with trade mark disputes; and (c) the function of the Intellectual Property Office of Singapore as a low-cost tribunal for hearing trade mark disputes.

11     On the other hand, the Respondent submits that the threshold for appellate intervention cannot be uniformly high in all appeals from the Trade Mark Registry, and that the question of whether appellate intervention is warranted must depend on the nature and facts of each case. The Respondent also submits that the Appellant’s reliance on Future Enterprises is misplaced, and that a lower threshold for intervention is justified because the courts are in as good a position as the PAR when it comes to evaluating the merits of the opposition.

12     In our judgment, the starting point of the analysis is our decision in Future Enterprises, although, for some reason, it was not referred to by the Judge. The relevant paragraphs of this decision state (at [5]–[7]) as follows:

5    … In Reef Trade Mark [2003] RPC 5, Robert Walker LJ considered the function of an appellate tribunal in relation to appeals from the UK Trade Mark Registry, and concluded (at [28]) that "an appellate court should … show a real reluctance, but not the very highest degree of reluctance, to interfere in the absence of a distinct and material error of principle”. In SC Prodal 94 SRL v Spirits International NV [2003] EWHC 2756 (Ch), Laddie J expressed the same sentiments (at [19]) as follows:

It is not the duty of this court to overturn a decision of the Trade Mark Registry simply because it comes to the conclusion that it might have decided the case differently had it, that is to say the High Court, been the court of first instance. It has to be demonstrated that the decision at first instance was wrong in a material way; that is to say there must be some significant departure from a proper assessment of the law or the facts.

6    ... Such an approach is consistent with established principles relating to appeals from tribunals that are not in the nature of a rehearing, such as an appeal from a decision of the PAR.

7    The smorgasbord of trade mark cases which has reached the appellate courts demonstrates the innumerable (and subjectively perceived) similarities and differences that can be conjured up and persuasively articulated by an imaginative and inventive legal mind. ... In the light of the highly subjective nature of assessing similarity and the likelihood of confusion, we agree with the approach that an appellate court should not disturb the findings of fact of a trade mark tribunal unless there is a material error of principle.

[emphasis added in italics and bold italics]

13     It is evident that the Appellant’s position is supported by the foregoing passages from Future Enterprises. However, Future Enterprises was, in our respectful view, wrong on this point. There are at least two reasons why this is so.

14     First, as observed in a number of subsequent High Court decisions (MediaCorp News Pte Ltd v Astro All Asia Networks plc [2009] 4 SLR(R) 496 at [26]; Ozone Community Corp v Advance Magazine Publishers Inc [2010] 2 SLR 459 at [31]–[32]; Polo/Lauren Co LP v United States Polo Association [2016] 2 SLR 667 at [7]), the court in Future Enterprises did not consider the effect of O 87 r 4(2) of the Rules of Court (Cap 322, R 5, 2006 Rev Ed) ("2006 ROC”) (which is worded identically in the current edition (Cap 322, R 5, 2014 Rev Ed) ("ROC”)). That provision expressly states that an appeal to the court from a decision of the registrar "shall be by way of rehearing”:

(2)     An appeal shall be by way of rehearing and the evidence used on appeal shall be the same as that used before the Registrar and, except with the leave of the Court, no further evidence shall be given. [emphasis added]

15     In our judgment, the words "shall be by way of rehearing” in O 87 r 4(2) of the ROC directs the appellate court to hear the matter afresh. We note, further, that these same words are used in O 57 r 3(1) of the ROC, which deals with appeals to this court. However, there is no threshold requirement of general application that a "material error of fact or law” (or, in the words of the court in Future Enterprises, a "material error of principle”) be shown before appellate intervention is warranted in such appeals. It seems likely to us that O 87 r 4(2) of the 2006 ROC was not drawn to the court’s attention in Future Enterprises.

16     Moreover, in Valentino Globe BV v Pacific Rim Industries Inc [2009] 4 SLR(R) 577, Chan Seng Onn J considered the meaning of the phrase "by way of rehearing” without (for some reason) considering Future Enterprises. His decision provides an illustration of the courts’ approach when Future Enterprises is removed from the analytical framework, and accords with our view that Future Enterprises was wrong on this point. In this regard, Chan J held (at [11]) that:

"By way of rehearing” in my view means that the Court is not constrained to determine only whether the tribunal’s decision was proper and/or contained manifest errors of fact and law. If it wishes to, the Court in its discretion may consider the entire ream of evidence before it and venture beyond determining the propriety of the tribunal’s decision or inquiring into whether there had been manifest errors of fact or law. However, I do not think that it places an irrevocable burden upon the Courts to hear the matter anew so that the substantive merits fall to be determined afresh. … [original emphasis omitted; emphasis added in italics]

17     Second, the requirement that there be a "material error of fact or law” before appellate intervention is warranted implies that deference should be accorded to the registrar. Given the nature of trade mark opposition proceedings, however, there is no principled reason why this ought to be the case. This is because trade mark opposition proceedings do not involve the exercise of a discretion. Rather, they involve questions of legal correctness and entitlement.

18     To illustrate, we use, as an example, s 8(2)(b) of the TMA. As will be evident when we come to Issue 2 (see [26] below), there are three requirements which must be satisfied in order for an opposition to succeed under this ground: (a) similarity of marks; (b) similarity (or identity) of goods or services; and (c) likelihood of confusion arising from the two similarities. Once it is shown that these three requirements are satisfied, the registrar has no discretion to disallow the opposition. This is because s 8(2)(b) of the TMA makes absolutely no provision for this. It does not, for instance, state that the registrar must additionally be satisfied that it is "just and equitable” for the opposition to succeed. What is really in issue is a question of legal correctness and entitlement.

19     Before us, counsel for the Appellant, Mr Pang Sze Ray, Melvin, argued that the exercise of discretion comes in at the (earlier) stage where the registrar determines whether there is similarity of marks. We disagree. While we accept that the question of similarity of marks is one of impression, all this means is that there is an element of subjectivity involved. However, it does not follow that what is involved is the exercise of a discretion. The nature of judicial discretion in the true sense is that there will be a range of legally-permissible outcomes or answers. From these, the judge of first instance may choose one and this will not be regarded as wrong or as affording a basis for appellate interference even if the appellate court might have come to a different conclusion had it been vested with the jurisdiction to hear the matter in place of the judge of first instance. This is not the case here. There is, for instance, detailed case law guidance on how the similarity of marks is to be assessed. The question, in the final analysis, remains one of legal correctness and entitlement.

20     In cases involving the exercise of a discretion, it is well established that an appellate court should not substitute the lower court or tribunal’s decision with its own simply because it would have reached a different decision. In Tay Beng Chuan v Official Receiver and Liquidator of Kie Hock Shipping (1971) Pte Ltd [1987] SLR(R) 123 ("Tay Beng Chuan”), we held (at [16]) as follows:

It is well settled that, although it would only be in extraordinary circumstances, that the Court of Appeal would interfere with the discretionary decision of the judge in the conduct of the business in his own court, the Court of Appeal would interfere if clearly satisfied that the decision was wrong so as to defeat the rights of the parties altogether and would be an injustice to one or other of the parties.

21     In TDA v TCZ and others [2016] 3 SLR 329, Judith Prakash J (as she then was) observed (at [25]) on the strength of Tay Beng Chuan that the standard for overturning a judge’s exercise of discretion is a "high one”. In contrast (and notwithstanding the "normative grounds” relied on by the Appellant (see [10] above)), we see no reason why deference should be accorded to a lower court or tribunal when what is involved is a question of legal correctness and entitlement. As we have explained at [17]–[19] above, the present case falls into the latter category.

22     Aside from this, there is also the trite principle that an appellate tribunal should not set aside a trial judge’s finding of fact, based as it is on evidence of witnesses, unless the appellant satisfies the appellate tribunal that the trial judge is plainly wrong (Seah Ting Soon (trading as Sing Meng Co Wooden Cases Factory) v Indonesian Tractors Co Pte Ltd [2001] 1 SLR(R) 53 ("Seah Ting Soon”) at [22]). This reluctance to interfere stems from the recognition of the simple fact that the trial judge is in a better position to assess the veracity and credibility of witnesses giving oral evidence (Seah Ting Soon at [22]). This is unsurprising as the trial judge is the primary receiver of the evidence and the trier of facts.

23     The present case, however, does not fall into this class of cases. As the Respondent points out, there was no trial and no oral evidence given at any level of hearing. All the evidence was tendered by way of statutory declarations only. In this regard, we have also previously cautioned that a distinction must be drawn between the perception of facts and the evaluation of facts, and observed that, especially where only written evidence is involved, an appellate court will be in as good a position as the trial court to make its own evaluation from the primary facts (Ho Soo Fong and another v Standard Chartered Bank [2007] 2 SLR(R) 181 at [20]; The "Andres Bonifacio” [1993] 3 SLR(R) 71 at [46]).

24     In the circumstances, we are satisfied that Future Enterprises was wrong on this point. In respect of Issue 1, therefore, we hold that there is no threshold requirement for a "material error of fact or law” to be shown before appellate intervention is warranted in appeals from the Trade Marks Registry.

Issue 2: Opposition under s 8(2)(b) of the TMA

25     Issue 2 concerns the ground of opposition under s 8(2)(b) of the TMA, which provides as follows:

(2) A trade mark shall not be registered if because —

(b)    it is similar to an earlier trade mark and is to be registered for goods or services identical with or similar to those for which the earlier trade mark is protected,

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

26     The PAR allowed the opposition on this ground, but this was reversed by the Judge. Under the step-by-step approach we endorsed in Staywell Hospitality Group Pty Ltd v Starwood Hotels & Resorts Worldwide, Inc and another and another appeal [2014] 1 SLR 911 ("Staywell”) (at [15]), the three requirements of (a) similarity of marks; (b) similarity (or identity) of goods or services; and (c) likelihood of confusion arising from the two similarities are to be assessed systematically. The first two elements are assessed individually before the final element, which is assessed in the round. It is not disputed that the second element is satisfied. Thus, only the first and third elements are in dispute in this appeal, and we consider them in turn.

Similarity of marks

27     With regard to the first element (namely, similarity of marks), the court must ultimately conclude whether the marks, when observed in their totality, are similar rather than dissimilar (Staywell at [17]). The well-established criteria of visual, aural, and conceptual similarities do not invite a formulaic consideration; rather, they are signposts towards answering the question of whether the marks are similar, and trade-offs can occur between the three aspects of similarity in the marks-similarity inquiry (Staywell at [18]). The 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 (Hai Tong Co (Pte) Ltd v Ventree Singapore Pte Ltd and another and another appeal [2013] 2 SLR 941 ("Hai Tong”) at [40(c)]). Furthermore, it is assumed that the average consumer has "imperfect recollection” (Hai Tong at [40(d)]). Therefore, 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 (Hai Tong at [40(d)]).

Distinctiveness

28     Before proceeding to consider the three aspects of similarity, we deal briefly with a preliminary point. This has to do with the distinctiveness of the Appellant’s CAESAR Mark. The relevance of distinctiveness to the marks-similarity inquiry was explained in Staywell, where we observed (at [25]) that a mark which has greater technical distinctiveness enjoys a high threshold before a competing sign will be considered dissimilar to it.

29     Whilst appreciating our pronouncement in Staywell (at [30]) that distinctiveness is a factor integrated into the assessment of mark similarity, both the PAR and the Judge analysed the issue of distinctiveness separately. The PAR held that the Appellant’s CAESAR Mark had a medium level of inherent distinctiveness. The PAR also accorded some weight to the evidence tendered in relation to de facto distinctiveness. On the whole, the PAR took the view that the Appellant’s CAESAR Mark was "at the very least, of a medium level of distinctiveness” [emphasis in original]. As for the Judge, he found, with respect to inherent distinctiveness, that there was no reason to disturb the PAR’s findings in relation to the distinctiveness of the name "Caesar”. It appears that the Judge did not make any express finding in relation to acquired or de facto distinctiveness. On the whole, the Judge found no reason to disagree with the PAR’s decision that the Appellant’s CAESAR Mark enjoyed a medium level of distinctiveness.

30     The Appellant does not challenge the Judge’s findings on distinctiveness, but submits that the Appellant’s CAESAR Mark "enjoys a higher threshold before the [Respondent’s] CAESARSTONE Mark will be considered dissimilar to it”. In our judgment, this is incorrect. What we held in Staywell (at [25]) was that a mark which has greater technical distinctiveness enjoys a high threshold before a competing sign will be considered dissimilar to it. The PAR concluded on this basis that "a mark which has a normal level of technical distinctiveness will correspondingly enjoy only a normal threshold before a competing sign will be considered dissimilar to it” [emphasis in original]. Thus, when the PAR (and the Judge) held that the Appellant’s CAESAR Mark had a medium level of distinctiveness, what she (and the Judge) probably meant was that the Appellant’s CAESAR Mark was of a normal level of distinctiveness and therefore enjoyed only a normal threshold before the Respondent’s CAESARTSTONE Mark would be considered dissimilar to it.

31     In any event, nothing in the present appeal turns on this issue. We turn, therefore, to consider the three aspects of similarity in the present case.

Visual similarity

32     In The Polo/Lauren Co, LP v Shop In Department Store Pte Ltd [2005] 4 SLR(R) 816, Lai Kew Chai J held (at [26]) that "[i]n cases where there is a common denominator, it is important to look at the differences between the mark and the sign in order to decide whether the challenged sign has been able to distinguish itself sufficiently and substantially”. We agree with this. In this connection, both the PAR and the Judge identified the following additions in the Respondent’s CAESARSTONE Mark: (a) the device; and (b) the word "stone”. The PAR found that the Appellant’s CAESAR Mark and the Respondent’s CAESARSTONE Mark were "somewhat visually similar”. On the other hand, the Judge held that the marks were visually dissimilar.

33     On appeal, the Appellant submits that the PAR was entitled to find that the device and the word "stone” played only a small role in the overall impression created in the mind of the average consumer and were insufficient to distinguish the Respondent’s CAESARSTONE Mark from the Appellant’s CAESAR Mark. The Respondent’s argument is that the Judge correctly found that the two marks were visually dissimilar as a whole.

(1)   The device

34     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 (Hai Tong at [62(b)]). In turn, the overall impressions conveyed to the public by a composite trade mark may, in certain circumstances, be dominated by one or more of its components (Hai Tong at [62(c)]). The device component has been found to be an equally significant, if not the dominant, component of a composite mark or sign where: (a) the device is significant and large; (b) the accompanying word(s) are devoid of any distinctive character, or are purely descriptive of the device component or of similar goods of a similar quality; and (c) the device component is of a complicated nature; but usually not where: (a) the device is simple and will not evoke any particular concept for the average consumer; (b) 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; and (c) the device component is more likely to be perceived as a decorative element rather than as an element indicating commercial origin (Hai Tong at [62(e)]).

35     The PAR thought that the device, while distinctive, only constituted a small part of the Respondent’s CAESARSTONE Mark when the mark was viewed as a whole. On the other hand, the Judge thought that the visual impact of the device should not be discounted as the device was "visually rather prominent”. Moreover, the Judge found that the device was not in any way allusive or descriptive of the textual component of the mark or the goods in question.

36     On appeal, the Appellant contends that if the Judge had applied Hai Tong correctly, he would have found that the device played only a minor role in the overall visual impression created by the Respondent’s CAESARSTONE Mark. On the other hand, the Respondent submits that the device is, visually, of equal or greater significance than the textual component of the Respondent’s CAESARSTONE Mark and therefore operates to sufficiently distinguish the competing marks.

37     In our judgment, the device is a somewhat insignificant component of the Respondent’s CAESARSTONE Mark. Having regard to the (relevant) factors set out in Hai Tong (at [62(e)]) (see [34] above), the device is, firstly, not significant and large. We accept the Judge’s observation that the device appears just at the beginning of the textual component, and also the Respondent’s submission that the former is larger (albeit only very slightly) than the font of the latter. On balance, however, we agree with the PAR that the device only constitutes a small part of the Respondent’s CAESARSTONE Mark when the mark is viewed as a whole.

38     Secondly, we think that the device is not of a complicated nature, but is simple and will not evoke any particular concept for the average consumer. The Appellant submits that the device comprises two swirls surrounding the letter "c”, which is the first letter of the word "caesarstone”. On the other hand, the Respondent submits that the device "consists of a complex amalgamation of coloured and stylised strokes”. In our view, however, the device appears to be made up of a series of three crescents. It is simple and does not evoke any particular concept in the eyes of the average consumer. For the same reason, we think that it is more likely to be perceived as a decorative element rather than as an element indicating commercial origin.

(2)   The word "stone”

39     The PAR found that the word "stone” appeared to be descriptive of the type of goods concerned and thus could not be considered to be the distinctive component of the Respondent’s CAESARSTONE Mark. In contrast, the Judge held, firstly, that the word "caesarstone” should be examined as a single word as opposed to its constituent components. Secondly, the Judge held that, whilst the word "stone” on its own might be descriptive of the goods in Class 19, the single word "caesarstone” was not a familiar expression that alluded to or designated the essential characteristics of the type of products falling within the class. Given the unique juxtaposition of the constituent words, the word "caesarstone” was distinctive as a whole. Accordingly, the significance of the word "stone” should not be disregarded for the purposes of assessing visual similarity.

40     The Appellant argues that the Judge failed to recognise that the constituent components of a word can each have an effect on the resulting overall impression, depending on, among other things, their relative levels of distinctiveness. Here, the differentiating word "stone” was merely descriptive of the goods in Class 19. Conversely, the Respondent submits that the word "stone” should not be disregarded when assessing visual similarity – it operates in combination with the rest of the Respondent’s CAESARSTONE Mark to distinguish the competing marks.

41     We agree with the Appellant that the word "stone” is merely descriptive of the goods in Class 19. In this regard, the public will not generally consider a descriptive element forming part of a complex mark as the distinctive and dominant element of the overall impression conveyed by that mark (Honda Motor Europe Ltd v Office for Harmonisation in the Internal Market (Trade Marks and Designs) (Case T-363/06) [2009] ETMR 34 at [39]) and, for this reason, we do not think that the presence of such a descriptive element can fairly be regarded as being effective to displace similarity.

42     The Respondent contends that the amalgamation of the words "caesar” and "stone” creates a unique textual juxtaposition which is distinctive and not descriptive of the relevant goods. We disagree. The question is not so much one of the distinctiveness of the Respondent’s CAESARSTONE Mark on its own. Rather, the real question here is whether the word "stone” serves to sufficiently and substantially distinguish the Respondent’s CAESARSTONE Mark from the Appellant’s CAESAR Mark (see [32] above). In our judgment, and as we have explained in the previous paragraph, because the word "stone” is merely descriptive of the goods in Class 19, it does not.

(3)   Conclusion on visual similarity

43     In the premises, we consider that the word "caesar” is the distinctive and dominant component of the Respondent’s CAESARSTONE Mark: the device is a somewhat insignificant component of the Respondent’s CAESARSTONE Mark (see [37]–[38] above), while the word "stone” is merely descriptive of the goods in Class 19 (see [41]–[42] above). We add that it is also quite clearly the case that the word "caesar” is the distinctive and dominant component of the Appellant’s CAESAR Mark. This being the case, we do not think that the differences between the two marks (namely, (a) the device; and (b) the word "stone”) serve to distinguish the Respondent’s CAESARSTONE Mark sufficiently and substantially. The overall impression conveyed by both marks is dominated by the word "caesar”.

44     We appreciate that there are several other stylistic differences between the two marks. These include differences in: (a) the font; (b) the letter case; (c) the overall length of the textual components; and (d) the arrangement of the letters, in particular, the ligating of the "a” and the "e” in the Appellant’s CAESAR Mark. However, as we have observed earlier (see [27] above), the fact that the average consumer is treated as having "imperfect recollection” means that 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 (Hai Tong at [40(d)]). Hence, while these differences may be evident (and even obvious) on a side-by-side comparison, this is not how visual similarity is to be assessed. In our judgment, these stylistic differences do not substantially detract from the fact that the overall impression conveyed by both marks is dominated by the word "caesar”. In all the circumstances and having regard also to these stylistic differences, we consider that the marks are of a moderate level of visual similarity.

Aural similarity

45     We set out two approaches towards assessing aural similarity in Staywell (at [31]–[32]). The first is to consider the dominant component of the marks and the second is to undertake a quantitative assessment as to whether the competing marks have more syllables in common than not.

46     The PAR found that there was "some aural similarity” between the marks. The Judge agreed, but added that the similarity "lies toward the lower end of the spectrum”. The Appellant contends that both approaches support a finding of a high level of aural similarity, while the Respondent argues that the Judge was correct in holding that the aural similarity between the marks fell towards the lower end of the spectrum.

47     Under the first approach, we consider that, for the purposes of the aural analysis, the dominant component of the Respondent’s CAESARSTONE Mark is again the word "caesar”. The device is obviously irrelevant in the aural analysis, while what we have said in relation to the word "stone” in the context of the visual analysis (see [41]–[42] above) applies with equal force in this context. The dominant component of the Appellant’s CAESAR Mark, which, without doubt, is the word "caesar”, is identical. As for the second approach, the two marks clearly have more syllables in common than not. In these premises, it is our view that the two marks have a high level of aural similarity.

Conceptual similarity

48     The conceptual analysis seeks to uncover the ideas that lie behind and inform the understanding of the mark as a whole (Staywell at [35]). The PAR found that the marks were only conceptually similar to a low extent, whereas the Judge found that there was little or no conceptual similarity.

49     With respect to the Appellant’s CAESAR Mark, the PAR observed that it comprised the sole word "Caesar”, which was commonly understood to refer to the Roman General Gaius Julius Caesar. Thus, the mark "may convey the idea of supremacy, power and authority”. The Judge did not disagree with this.

50     As for the Respondent’s CAESARSTONE Mark, the PAR took the view that the addition of the word "stone” would "colour the [mark] to the effect of referring to the style of masonry during the reign of Gaius Julius Caesar”. The Judge, on the other hand, made three findings, among others. First, he held that the Respondent’s CAESARSTONE Mark "does not offer a clear concept”. Second, if this was wrong and the mark evoked the image or idea of stonemasonry, this was a concept which set it apart from the idea of power, supremacy, and authority emanating from a Roman imperial statesman. Third, and "[f]rom a different perspective”, the Appellant’s CAESAR Mark was minimalist in its design as compared to the design of the Respondent’s CAESARSTONE Mark which was more complex.

51     The Appellant submits that the PAR was entitled to find that both marks were conceptually similar as both marks connote the idea of supremacy, power, and authority, even though the Respondent’s CAESARSTONE Mark’s impression might be affected somewhat by the addition of the word "stone”. The marks are therefore similar to a moderate or high extent. The Respondent, on the other hand, contends that the Judge correctly concluded that the marks had little or no conceptual similarity.

52     We are doubtful that the Appellant’s CAESAR Mark necessarily conveys the idea of supremacy, power, and authority in the context of Class 19 goods. We say this because there does not seem to be any connection that can be immediately drawn between the word "caesar”, on the one hand, and Class 19 goods, on the other. In our view, it may well be the case that the Appellant’s CAESAR Mark is of no conceptual significance. In a similar vein, the Respondent’s CAESARSTONE Mark also appears to us (as it did to the Judge) to offer no clear concept. In so far as the device is concerned, we have noted earlier that it is simple and will not evoke any particular concept for the average consumer (see [38] above). As for the word "stone”, it is merely descriptive of the goods in Class 19 (see [41]–[42] above). We are therefore once again confronted with a similar conundrum, in that there does not seem to be any connection that can be immediately drawn between the word "caesar” and Class 19 goods. To the extent that both marks do not offer a clear concept, it cannot be said that the marks are conceptually dissimilar; the most that can be said is that this is a neutral consideration.

53     Even if we accept that the Appellant’s CAESAR Mark conveys the idea of supremacy, power, and authority, we consider that the same idea would be conveyed by the Respondent’s CAESARSTONE Mark. In this regard, we think that, for the purposes of the conceptual analysis, and for the same reasons set out at [52] above, the distinctive and dominant component of the Respondent’s CAESARSTONE Mark is again the word "caesar”. Accordingly, the same idea of supremacy, power, and authority would be conveyed by both marks if it is so conveyed at all. In this connection, we do not agree with the PAR’s view that the word "stone” colours the mark by invoking the idea of the style of masonry during the reign of Gaius Julius Caesar. As we have already mentioned, the word "stone” is merely descriptive of the goods in Class 19 (see [41]–[42] above). On this analysis, therefore, there would be a high level of conceptual similarity.

54     For completeness, we add that we do not see the relevance of the Judge’s finding that the Appellant’s CAESAR Mark was minimalist in its design as compared to the design of the Respondent’s CAESARSTONE Mark which was more complex. As the Appellant points out, this is, at most, relevant to visual (rather than conceptual) similarity, and we have in fact touched on this at [44] above.

Conclusion on similarity of marks

55     To summarise our findings on the similarity of marks, there is, in our judgment, as between the Appellant’s CAESAR Mark and the Respondent’s CAESARSTONE Mark, a moderate level of visual similarity (see [44] above) and a high level of aural similarity (see [47] above). As far as conceptual similarity is concerned, this is either neutral or there would be a high level of conceptual similarity; but as we see it, there is no basis for finding any conceptual dissimilarity (see [52]–[53] above). On the whole, we find that the marks are similar rather than dissimilar.

Likelihood of confusion

56     The likelihood of confusion inquiry directs the court to look at: (a) how similar the marks are; (b) how similar the goods or services are; and (c) given this, how likely it is that the relevant segment of the public will be confused (Staywell at [55]). In opposition proceedings, the inquiry takes into account the actual and notional fair uses of both the existing and the application mark (Staywell at [60]). The factors which are admissible in the confusion inquiry include:

(a)      Factors relating to the impact of marks-similarity on consumer perception: (i) the degree of similarity of the marks themselves; (ii) the reputation of the marks; (iii) the impression given by the marks; and (iv) the possibility of imperfect recollection of the marks (Staywell at [96(a)]).

(b)      Factors relating to the impact of goods-similarity on consumer perception: (i) the normal way in or the circumstances under which consumers would purchase goods of that type; (ii) whether the products are expensive or inexpensive items; (iii) the nature of the goods and whether they would tend to command a greater or lesser degree of fastidiousness and attention on the part of prospective purchasers; and (iv) the likely characteristics of the relevant consumers and whether they would or would not tend to apply care or have specialist knowledge in making the purchase (Staywell at [96(b)]).

57     There are at least two specific aspects to the element of confusion. The first is mistaking one mark for another. 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 two marks emanate from the same source or from sources that are economically linked or associated (Hai Tong at [74]). Further, in Sarika Connoisseur Cafe Pte Ltd v Ferrero SpA [2013] 1 SLR 531 ("Sarika”), we held (at [57]) (following our earlier decision in Mobil Petroleum Co, Inc v Hyundai Mobis [2010] 1 SLR 512 ("Mobil”) at [77]–[82]) that the test to be adopted in determining likelihood of confusion is whether a substantial portion of the relevant public will be confused. Although in Sarika and Mobil we were concerned with other provisions of the TMA and the Trade Marks Act (Cap 332, 1999 Rev Ed), in our judgment, the same principle applies in the present case. In this regard, we also said in Sarika (at [57]) 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 …

58     The PAR took the view that there was a reasonable likelihood of confusion. In contrast, the Judge found that there was no likelihood of confusion.

59     The Judge also considered the preliminary question concerning the relevant date of assessment. This arose because the Respondent had adduced evidence of its sales and promotional expenditure figures after the date of application (which was 21 November 2007) in order to show the reputation of its mark in Singapore. The Judge, however, did not make a conclusive determination on this point as he took the view that there was no likelihood of confusion whether the date of application or the date of the opposition hearing was used as the basis for assessment. This matter is not seriously pursued by either party on appeal. In any event, we are of the view that nothing in the present appeal turns on this issue. In this regard, we agree with the Judge that even if it is proper to take account of developments in the reputation of the Respondent’s CAESARSTONE Mark between the application and the hearing, this is, at best, just one factor to be assessed in the round.

60     In this connection, the Judge considered four factors in his analysis: (a) the nature of the goods; (b) the mode of purchase; (c) the similarity of marks; and (d) the similarity of goods and services. All of these factors come within the ambit of the permissible factors outlined at [56] above. As the parties’ respective cases largely adopt this analytical framework, we propose to do the same.

Nature of goods

61     The Judge found that the nature of the goods went towards reducing the likelihood of confusion. In his view, the average consumer was likely to pay close attention to the purchase of Class 19 goods for decoration of homes or commercial establishments. He agreed with the Respondent that the purchase of such goods was likely to be an intensely personal and relatively expensive endeavour with long-term consequences. It was probable that the purchase would not be made lightly, and consumers would be fastidious in selecting the brand of goods which they wished to purchase.

62     The Appellant contends that the Judge arrived at a finding of fact which was not reasonable as a matter of inference from the evidence before him. It also submits that the purchase of tiles can be a relatively inexpensive endeavour (for instance, if it was for a small area), and even impersonal (say, for an area in a building that may not be prominent). On the other hand, the Respondent, in seeking to affirm the Judge’s findings, highlights the "intrinsically expensive and specialised nature of the relevant goods” and the "high level of attention and specialist knowledge which would be occasioned as a result”.

63     In our judgment, the nature of the goods does point towards a likelihood of confusion. In this regard, Class 19 covers:

Building materials (non-metallic); non-metallic rigid pipes for building; asphalt, pitch and bitumen; non-metallic transportable buildings; monuments, not of metal.

64     Before us, Mr Pang made the point that consumers of Class 19 goods focus more on the price and attributes (such as water or scratch resistance, colour, finishing and style) of the goods, rather than the brand. Having regard to the goods covered by Class 19, we agree with this submission, although, as we will explain shortly, we think that this proposition is more accurately stated in terms that consumers of Class 19 goods are likely to be indifferent towards the mark used in relation to the goods (and not, as Mr Pang framed it, the brand of the goods).

65     Mr Pang further submitted that, given this, consumers of Class 19 goods would be less cognisant of the differences between the marks, leading to a higher likelihood of confusion. This submission is, in itself, unproblematic but it is first necessary to consider an anterior question that came to the fore at the hearing before us: if consumers of Class 19 goods are likely to be indifferent towards the mark used in relation to the goods, but tend to make their purchasing decisions based on the price and attributes of these goods, does the question of a likelihood of confusion even arise in the first place? Aside from this, there is also the Judge’s holding that the purchase of Class 19 goods was likely to be an intensely personal and relatively expensive endeavour (see [61] above). We deal with these various issues in turn.

(1)   Distinction between brands and marks

66     We begin with some brief observations on the distinction between brands, on the one hand, and marks, on the other. In essence, the latter is a subset of the former and is also what is protected under trade mark law. In this regard, Prof Susanna H S Leong observes in Intellectual Property Law of Singapore (Academy Publishing, 2013) (at para 27.069) that:

There is a common misconception that trade marks are brands and vice versa. A simple way to understand the relationship between brands and trade marks is to visualise brands as an all encompassing marketing concept that comprises names, logos, designs, unique packaging and other markers …

67     Similarly, the English High Court held in O2 Holdings Ltd v Hutchison 3G Ltd [2006] RPC 29 (at [4] and [7]) that:

Brands are big business. They can be worth many millions of pounds. The value of the Coca Cola brand has been said to be worth 60 per cent of the market capitalisation of the Coca Cola Corporation. Defining a brand is not easy. A lawyer would tend to think of goodwill, trade marks and so on. But a brand includes more elements; such as image and reputation; the values that the brand owner tries to inculcate in the buying public. A brand is what customers choose to buy. Many decisions about brands are made by customers emotionally or intuitively rather than rationally. Successful brands create a relationship of trust between the customer and the brand.

… English law does not, however, protect brands as such. It will protect goodwill (via the law of passing off); trade marks (via the law of trade mark infringement); the use of particular words, sounds and images (via the law of copyright); shapes and configurations of articles (via the law of unregistered design right) and so on. But to the extent that a brand is greater than the sum of the parts that English law will protect, it is defenceless against the chill wind of competition.

68     In light of this distinction, and as we have already mentioned (see [64] above), we think that Mr Pang’s proposition is more accurately stated in terms that consumers of Class 19 goods are likely to be indifferent towards the mark used in relation to the goods.

(2)   Likelihood of confusion vis-à-vis consumer indifference

69     We turn then to the more difficult question of whether the question of a likelihood of confusion even arises when consumers are indifferent towards the mark used in relation to the relevant goods or services. Somewhat surprisingly, there does not seem to be a ready answer to this question in the authorities.

70     Answering this question in the negative seems intuitively attractive. Indeed, we note that the Judge held, in relation to this argument, that:

… This argument, in my view, misses the point. If it is indeed the case that consumers merrily purchase Class 19 goods without paying attention to what trade mark is attached to those goods, there can be no confusion as to their trade origin, which is precisely what the trade mark regime aims to protect. [emphasis added]

71     In our judgment, consumer indifference towards the mark used in relation to the relevant goods or services does not preclude the question of a likelihood of confusion from arising. On the contrary, we accept Mr Pang’s suggestion to the effect that consumer indifference would cause consumers to pay less attention to the differences between the marks, thereby pointing towards a likelihood of confusion (see [65] above). We arrive at this view for a number of reasons.

72     The first is that consumer indifference towards the mark used in relation to the relevant goods or services is a conceptually distinct matter from the likelihood of confusion in relation to the marks. It is not difficult to envisage scenarios where there exists a likelihood of confusion (and, indeed, actual confusion) even where consumers are indifferent towards the mark used in relation to the relevant goods or services. Taking the present case as an example, we do not find it unlikely that, amongst the end-consumers of Class 19 goods, there would be some among the relevant public who would play an active role in the purchasing process. Suppose a consumer who falls into this category does a Google search for "affordable and good non-slip tiles”. His focus would be on the price ("affordable”) and attributes ("good” and "non-slip”) of the goods. Although his focus is not on the precise mark used in relation to the goods, his search turns up the Appellant’s tiles. He is indifferent to the fact that the Appellant’s tiles bear the Appellant’s CAESAR Mark, but he makes a note of this because (and only because) the Appellant’s tiles match his price and attribute requirements. When he subsequently makes his purchase (with or without the help of specialists – a point which we will turn to shortly), he then sees the Respondent’s tiles and, because of the similarity of marks, confuses them for the Appellant’s. In such a scenario, there exists a likelihood of confusion (and, indeed, actual confusion) even though the consumer is indifferent towards the mark used in relation to the goods.

73     The short point is that it does not follow that there can be no likelihood of confusion just because consumers are indifferent towards the mark used in relation to the relevant goods or services. The two are conceptually distinct.

74     The second reason is related to the first and has to do with the function of trade mark law in preventing origin-based confusion. Indeed, s 2(1) of the TMA defines a "trade mark” as follows:

"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 … [emphasis added in italics and bold italics]

75     We have repeatedly affirmed this function of trade mark law. In Hai Tong, we noted (at [72]) that the "proper province” of trade mark protection "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”. A similar observation was made in Staywell, where we said (at [95]) that "the risk of origin-based confusion is the primary interest sought to be protected by trade mark law”.

76     The same proposition has been emphasised by leading local commentators. Prof Ng-Loy Wee Loon writes in Law of Intellectual Property of Singapore (Sweet & Maxwell, 2nd Ed, 2014) (at para 15.1.4) that the function of trade marks as "badges of origin” forms the "core” of trade mark protection. In a similar vein, Tan Tee Jim SC notes in Law of Trade Marks and Passing Off in Singapore vol I (Sweet & Maxwell, 3rd Ed, 2014) (at para 2.004) that:

… a trade mark in effect operates in the marketplace to protect consumers against the risk of confusion as to the origin of the goods or services deriving from the unauthorised use of an identical or similar mark in respect of identical or similar goods or services. …

77     Consequently, a trader has a real interest in being identified as the origin of the goods or services bearing his mark. Once this function of trade mark law is appreciated, it becomes evident that the reference to a "likelihood of confusion” in s 8(2) of the TMA refers really to a likelihood that the public will be unable to differentiate between the trade origins of the goods or services bearing the competing marks. Indeed, in Staywell, we held (at [78]) that the plain words of s 8(2) (and s 27(2)) of the TMA "direct the court to ultimately assess the likelihood of inaccurate consumer perception as to the source of goods” [emphasis added]. This may be distinguished from a situation where no such likelihood exists because the public is able to make such a differentiation. The inquiry into a likelihood of confusion therefore revolves around the ability or inability to differentiate between trade origins. Accordingly, consumer indifference (or otherwise) towards the mark used in relation to the relevant goods or services might be relevant to the extent that it impacts this inquiry. However, we do not think that it is an answer in opposition proceedings under s 8(2)(b) of the TMA to say that consumer indifference means that the question of a likelihood of confusion does not even arise in the first place. Such a proposition does not relate to the ability or inability of the public to differentiate between trade origins.

78     The third (and final) reason is policy. The entire trade mark regime would be seriously undermined if a finding of a likelihood of confusion were excluded every time it was thought there was consumer indifference towards the competing marks because of the nature of the relevant goods or services. In the case of s 8(2)(b) of the TMA, this would mean that, as against an incumbent trader, new traders would be free to apply confusingly similar marks on identical or similar goods or services with impunity. This is an outcome which we think cannot be right. In Staywell, we expressed the concern (at [78]) that the registration regime should remain "robust and effective to protect the value of the trade mark itself as a badge of origin and to vindicate the exclusive and proprietary rights of the owner”. We echo and affirm these sentiments.

79     For these reasons, we are satisfied that consumer indifference towards the mark used in relation to the relevant goods or services does not preclude the question of a likelihood of confusion from arising. Consumer indifference would, in the normal course of events, point towards a likelihood of confusion in so far as consumers would pay less attention to the differences between the marks. Conversely, where consumers are especially sensitive towards the mark used in relation to the relevant goods or services (for example, in the case of luxury goods), this would likely point away from a likelihood of confusion as consumers would pay more attention to the differences between the marks. Accordingly, the finding that consumers of Class 19 goods are likely to be indifferent towards the mark used in relation to the goods (see [64] above) points towards, rather than away from, a likelihood of confusion in the present case.

(3)   Expensive and personal endeavour

80     We turn to the Judge’s holding that the purchase of Class 19 goods was likely to be an intensely personal and relatively expensive endeavour (see [61] above). We are unimpressed by the Appellant’s attempt to challenge this holding. Quite apart from the atypical and unrepresentative examples suggested by the Appellant (see [62] above), we do not see how it can be seriously argued that the purchase of Class 19 goods is anything other than as the Judge found. The purchase of Class 19 goods is doubtlessly personal – one is likely to have to live with one’s choice for years. Also, while we are inclined to agree with the Judge that the evidence could have been stronger, we nonetheless think that Class 19 goods are more likely to be expensive rather than inexpensive. Before us, Mr Pang highlighted a sample invoice showing the Respondent’s tile being sold at a unit price of $2.00. However, this is unrepresentative as the other invoices that are exhibited clearly show a much wider range of prices. We therefore agree with the Judge that the purchase of Class 19 goods is likely to be a personal and expensive endeavour.

81     In the normal course of events, such a finding would point away from a likelihood of confusion due to the greater care that the average consumer would exercise in making such a purchase (see, for example, Staywell at [96(b)]). However, this is premised on the assumption that the greater care so exercised relates to the differences between the two competing marks. In the present case, however, we have found that consumers of Class 19 goods are likely to be indifferent towards the mark used in relation to the goods; their focus, rather, tends to be on the price and attributes of the goods (see [64] above). As such, the greater care exercised by the average consumer would relate to these factors instead, and not to the differences between the two competing marks. This point is made in Thomas R Lee, Glenn L Christensen & Eric D DeRosia, "Trademarks, Consumer Psychology, and the Sophisticated Consumer” (2008) 57 Emory LJ 575 (at pp 642–643) in the following terms:

Indeed, and more fundamentally, our model suggests that there is an over-simplification in the case law’s treatment of consumer care and sophistication. The courts generally assert that a careful consumer is less likely to be confused as to source or sponsorship. But that conclusion follows only where the indicators of consumer care – motivation and ability – are likely to be directed at the source-identification judgment and not at something else. Since some highly motivated consumers may care mostly about price or some other consideration other than source, the so-called sophisticated consumer may not ultimately be one who is less likely to suffer source confusion. [emphasis added]

82     Accordingly, while we agree that the purchase of Class 19 goods is likely to be a personal and expensive endeavour, we do not consider this to be a factor pointing away from the likelihood of confusion on the facts of the present case. It is, at best, a neutral factor.

Mode of purchase

83     The Judge found that the mode of purchase tended towards a lower likelihood of confusion. In this regard, he made two findings, among others. First, he found that it "would not at all be surprising” if a large proportion of consumers would consult specialists when purchasing Class 19 goods (although he also accepted that there was no evidence indicating whether specialists could find themselves confused as to the origin of the goods supplied by the Appellant and the Respondent). Second, he took the view that if the competing goods were displayed side by side, in close proximity or on the same supplier’s webpage, it would surely be easier for the end-user to see the differences in the trade marks.

84     The Appellant (once again) submits that the Judge arrived at a finding of fact which was not reasonable as a matter of inference from the evidence before him. It also submits that the visual aspect of the marks is the most important factor in determining the likelihood of confusion. On the other hand, the Respondent submits that the relevant viewpoint should be that of specialists or persons who receive professional advice. Even if the relevant viewpoint were that of an ordinary consumer not under professional advice, the mode of purchase would still point away from a likelihood of confusion.

85     We deal first with the Judge’s first finding, which is that it "would not at all be surprising” if a large proportion of consumers would consult specialists when purchasing Class 19 goods. In our judgment, the relevant public in the present context is a composite comprising both specialists and lay end-consumers. Before us, Mr Pang accepted (rightly, in our view) that specialists would not likely be confused. As for the lay end-consumers, Mr Pang accepted (again rightly, in our view) that the predominant mode of their purchase would tend to be with the assistance of specialists (notwithstanding his written submissions otherwise). However, his point is that the likelihood of confusion on the part of these consumers would not be dispelled even with the assistance of specialists. As Mr Pang explained, this is because their discussions with the specialists would revolve around the price and attributes of the goods, rather than trade origin. Put another way, the submission, as we understand it, is that the likelihood of confusion on the part of these consumers is not dispelled because any discussions they might have with the specialists would not be about trade origin in the first place. As a result, the likelihood of confusion is not addressed even with the advice of specialists and, therefore, persists. We agree with this submission, as it is simply a corollary of our earlier finding that consumers of Class 19 goods tend to focus more on the price and attributes of the goods rather than the mark used in relation to the goods (see [64] above).

86     As regards the Judge’s second finding, we agree in so far as it is said that a side-by-side display of the parties’ goods (assuming for present purposes that this is in fact how the goods would likely be displayed) would make it easier for consumers to see the differences in the trade marks. However, this relates only to one aspect of the element of confusion, which is mistaking one mark for another. As we have emphasised earlier (see [57] above), the element of confusion extends to the relevant segment of the public perceiving that the competing marks are different, but yet remaining confused as to the origin which each mark signifies and perceiving that goods bearing the two marks emanate from the same source or from sources that are economically linked or associated. A side-by-side display of the parties’ goods does not serve to point away from the likelihood of this second type of confusion.

87     Accordingly, neither of the Judge’s two findings points away from a likelihood of confusion. The mode of purchase is, at best, a neutral factor in this inquiry.

Similarity of marks

88     The Judge seemed to be of the view that even if the marks were held to be similar as a whole, the strong visual dissimilarity between the marks would mean that there was no likelihood of confusion. In this connection, the Judge observed that the visual aspect was more important as it was more likely that the average consumer would have occasion to view the marks during the purchase process.

89     It is clear to us that the Judge’s view was very much premised on his earlier finding of visual dissimilarity between the Appellant’s CAESAR Mark and the Respondent’s CAESARSTONE Mark. Given our (contrary) finding that the marks are of a moderate level of visual similarity (see [44] above) and that, on the whole, the marks are similar rather than dissimilar (see [55] above), we are satisfied that this points towards a likelihood of confusion.

90     On appeal, the Respondent points to the alleged peaceful coexistence of the Appellant’s CAESAR Mark and the Respondent’s CAESARSTONE Mark in the United States (we note that this involves a "CAESARSTONE” word mark rather than the Respondent’s CAESARSTONE Mark) and Taiwan. We do not accord much weight to this. Apart from the fact that the trade mark regimes in these jurisdictions may be different from ours, we tend to agree with the Appellant that the Respondent has not shown how this alleged peaceful coexistence stems from an absence of confusion.

Similarity of goods

91     As mentioned earlier (see [26] above), it is not disputed that the element of similarity of goods is satisfied in the present case.

Conclusion on likelihood of confusion

92     To summarise our findings on the likelihood of confusion, we consider that: (a) the nature of the goods, on the whole, points towards a likelihood of confusion (see [79] and [82] above); (b) the mode of purchase is, at best, a neutral factor (see [87] above); and (c) the similarity of marks points towards a likelihood of confusion (see [89] above), as does the similarity of goods (see [91] above). On the whole, we think that there exists a likelihood of confusion on the part of the public.

Conclusion on Issue 2

93     In light of our findings that the marks are similar rather than dissimilar (see [55] above) and that there exists a likelihood of confusion on the part of the public (see [92] above), the Appellant’s opposition under s 8(2)(b) of the TMA ought to have succeeded. In respect of Issue 2, therefore, the Judge erred in finding that the opposition under s 8(2)(b) of the TMA was not made out.

Issue 3: Opposition under s 8(4)(a) read with s 8(4)(b)(i) of the TMA

94     Issue 3 concerns the ground of opposition under s 8(4)(a) read with s 8(4)(b)(i) of the TMA, which provides as follows:

(4) Subject to subsection (5), where an application for registration of a trade mark is made on or after 1st July 2004, if the whole or an essential part of the trade mark is identical with or similar to an earlier trade mark, the later trade mark shall not be registered if —

(a)    the earlier trade mark is well known in Singapore; and

(b)    use of the later trade mark in relation to the goods or services for which the later trade mark is sought to be registered —

(i)    would indicate a connection between those goods or services and the proprietor of the earlier trade mark, and is likely to damage the interests of the proprietor of the earlier trade mark …

95     The PAR allowed the opposition on this ground, but this was reversed by the Judge. The four elements which need to be shown for an opposition under this ground are: (a) the marks are identical or similar; (b) the earlier mark is well known in Singapore; (c) the use of the applicant mark would indicate a connection between the applicant’s goods/services and the incumbent proprietor; and (d) the connection is likely to damage the interests of the proprietor (Staywell at [119]).

96     We deal briefly with the first and third elements at the outset, because we find them to be clearly satisfied. With respect to the first element (namely, identity or similarity of marks), the Judge referred to his earlier decision in Rovio (at [146]), where he had held that there was no difference between the similarity of marks comparison in s 8(2)(b) and s 8(4) of the TMA. Neither party disputes this, nor do we see any reason to hold otherwise. Given our finding that, on the whole, the marks are similar rather than dissimilar (see [55] above), we find this element to be satisfied. As for the third element (which is the use of the applicant mark indicating a connection between the applicant’s goods/services and the incumbent proprietor), this will be satisfied where there is a likelihood of confusion (Staywell at [120]). Given our finding that there exists a likelihood of confusion on the part of the public (see [92] above), we find this element to be satisfied as well. For completeness, we note that, in light of his findings in relation to Issue 2, the Judge found that these two elements were not satisfied.

97     The main dispute in the present appeal concerns the second element, the question being whether the Appellant’s CAESAR Mark is well known in Singapore. As regards the fourth element (which is the likelihood of damage to the interests of the proprietor), it is evident that this only arises for consideration when all the three prior elements are satisfied. We turn, therefore, to consider the second element.

Whether the Appellant’s CAESAR Mark is well known in Singapore

98     The starting point with respect to the second element is s 2(7) and (8) of the TMA, which states as follows:

(7) Subject to subsection (8), in deciding, for the purposes of this Act, whether a trade mark is well known in Singapore, it shall be relevant to take into account any matter from which it may be inferred that the trade mark is well known, including such of the following matters as may be relevant:

(a)    the degree to which the trade mark is known to or recognised by any relevant sector of the public in Singapore;

(b)    the duration, extent and geographical area of —

(i)    any use of the trade mark; or

(ii)   any promotion of the trade mark, including any advertising of, any publicity given to, or any presentation at any fair or exhibition of, the goods or services to which the trade mark is applied;

(c)    any registration or application for the registration of the trade mark in any country or territory in which the trade mark is used or recognised, and the duration of such registration or application;

(d)    any successful enforcement of any right in the trade mark in any country or territory, and the extent to which the trade mark was recognised as well known by the competent authorities of that country or territory;

(e)    any value associated with the trade mark.

(8) Where it is determined that a trade mark is well known to any relevant sector of the public in Singapore, the trade mark shall be deemed to be well known in Singapore.

99     In Novelty Pte Ltd v Amanresorts Ltd and another [2009] 3 SLR(R) 216 ("Amanresorts”), we held (at [137] and [139]–[140]) that the factors in s 2(7) of the TMA were not exhaustive. Further, the court is ordinarily free to disregard any or all of the factors as the case requires (except for s 2(7)(a) of the TMA), and to take additional factors into consideration.

100    The PAR found that the Appellant’s CAESAR Mark was well known in Singapore. In making this finding, it appears that the PAR had given considerable weight to the following passage from Amanresorts (at [229]):

Finally, it will be recalled that it is not too difficult for a trade mark to be regarded as "well known in Singapore” – essentially, the trade mark in question need only be recognised or known by "any relevant sector of the public in Singapore” [emphasis added] (see s 2(7)(a) of the current TMA), which sector could in certain cases be miniscule. … [emphasis in original in italics; emphasis added in bold italics]

In contrast, the Judge held that the Appellant had not discharged its burden of proving that the Appellant’s CAESAR Mark was well known in Singapore. In so holding, the Judge considered: (a) the use of the mark; and (b) the overseas registrations of the mark and the successful enforcement of rights. We consider these two factors in turn. Before that, however, we make some observations concerning the passage from Amanresorts that we have cited above.

"Not too difficult”

101    Although we said in Amanresorts that it is "not too difficult” for a trade mark to be regarded as well known in Singapore (see [100] above), the Judge thought that this comment should not be taken to mean that the hurdle that trade mark owners had to cross was minimal. Rather, the comment had to be applied with judicious caution to the actual facts and circumstances of each case.

102    We do not think that this comment in Amanresorts was made to lay down a general principle. In this regard, we agree with the Respondent’s submission that the context of this comment was the desire to clarify that, in order for a mark to be well known in Singapore, the relevant sector to which a mark must be shown to be well known can be any relevant sector of the Singaporean public, and this sector need not be large in size. Beyond this, it should not be read as suggesting (more generally) that the threshold for a trade mark to be regarded as well known in Singapore is a low one.

Use of mark

103    With respect to the use of the mark, the Judge considered a few categories of evidence and we deal with these in turn.

(1)   The Appellant’s sales figures

104    The Appellant’s sales figures, as set out by the Judge, were as follows:

Appellant’s sales figures

Year

Singapore sales ($)

Worldwide sales ($)

2002

3,390,094

119,929,925

2003

1,874,266

108,568,093

2004

935,055

119,036,242

2005

1,433,751

159,243,918

2006

203,634

173,550,987

2007

577,656

176,795,005



105    The Judge thought that the Appellant’s sales figures in Singapore, whilst not insubstantial, did not constitute a significant percentage of its global sales. On appeal, the Appellant contends that the Judge was wrong to compare its Singapore sales figures with its global sales figures. Otherwise, if the Appellant had much weaker sales globally but a similar level of sales in Singapore, it would have a better chance of establishing that its mark is well known. On the other hand, the Respondent submits that the evidence in this regard is "extremely scanty and inadequate”. Before us, counsel for the Respondent, Mr Prithipal Singh s/o Seva Singh, pointed out that the invoices used to support the Appellant’s Singapore sales figures do not show that the goods sold were the Appellant’s goods and therefore do not show the use of the Appellant’s CAESAR Mark in Class 19. Moreover, even if this issue were overlooked, the exhibited invoices, when added up, do not yield the Singapore sales figures set out in the table above, but instead yield significantly lower figures (said to average less than €200,000 per year).

106    With respect to the Judge, we agree with the Appellant that it was wrong for the Judge to compare the Appellant’s Singapore sales figures with its global sales figures. Such a comparison is, in our respectful view, irrelevant to the inquiry at hand. In this regard, we agree with the Appellant’s contentions as set out in the previous paragraph. Moreover, while there is some merit in Mr Singh’s submissions concerning the state of the evidence, we are mindful that the evidence was ultimately given in the form of a statutory declaration, which would have been given on oath or affirmed. We are therefore prepared to accept that: (a) the invoices relate to the use of the Appellant’s CAESAR Mark in Class 19; and (b) the Singapore sales figures set out in the table above are accurate. With respect to (b), we add that it appears to us that the invoices were likely intended to be a representative sample rather than a comprehensive exhibition of all the relevant invoices. The Appellant’s sales figures therefore do go towards establishing that the Appellant’s CAESAR Mark was at least known in Singapore.

(2)   The Appellant’s promotional expenditure figures

107    The Appellant’s worldwide promotional expenditure figures, as set out by the Judge, were as follows:

Appellant’s worldwide promotional expenditure figures

Year

Amount ($)

2002

273,387

2003

247,812

2004

1,339,131

2005

2,092,412

2006

2,234,788

2007

1,120,784



108    The Judge held that it was unclear how much of these expenditures related to promotional activities in Singapore. Before us, Mr Pang pointed to two listings in the Singapore Green Book and the Singapore eGuide by the Appellant’s clients. However, we do not see how these listings add anything to the Appellant’s case because they do not shed any light on the extent of promotional activities in Singapore. Moreover, we note that the Appellant’s CAESAR Mark is not even used on these listings; all these listings show is that the Appellant’s clients carry goods under the Appellant’s brand. In these premises, we do not think that the Appellant’s promotional expenditure figures go towards establishing that the Appellant’s CAESAR Mark is well known in Singapore.

(3)   The Appellant’s Italian website

109    The Judge also considered the Appellant’s information concerning the number of hits on its Italian website between 1999 and 2008. The Judge accepted that the numbers were "rather significant” and that the website had English text which would be relevant to consumers outside Italy. However, the Judge ultimately held that there was no indication as to whether the hits were from members of the relevant sector of the public in Singapore. The Appellant has not made any serious attempt to dispute this finding on appeal and we see no reason to disagree with the Judge. The Appellant’s Italian website therefore does not go towards establishing that the Appellant’s CAESAR Mark is well known in Singapore.

(4)   The Appellant’s invoices and marketing materials

110    Finally, the Judge considered the Appellant’s invoices and marketing materials. He noted that these documents bore marks which were slightly different from the Appellant’s CAESAR Mark but agreed that the Appellant’s CAESAR Mark remained the dominant component of these marks. He therefore accorded some weight to these documents.

111    With respect to the Appellant’s invoices, these appear to be the same invoices referred to at [105]–[106] above. We add only that, before us, Mr Singh did not make much of the fact that a slightly different mark was used. As for the marketing materials, it appears that the PAR had regard to two sets of marketing materials but only one of these is before us. In any event, neither Mr Pang nor Mr Singh had much to say about these marketing materials during the oral hearing. We simply observe that we do not think that the evidence sufficiently establishes that these marketing materials were used in Singapore. They therefore do not go towards establishing that the Appellant’s CAESAR Mark is well known in Singapore.

Overseas registrations of mark and successful enforcement of rights

112    In relation to the overseas registrations of the mark and the successful enforcement of rights, the Judge noted the Appellant’s submission that it had applied for and/or obtained several registrations or applications for the Appellant’s CAESAR Mark in other jurisdictions. The Judge further noted that the successful enforcement of the Appellant’s CAESAR Mark in Australia was a relevant factor in determining whether the mark was well known in Singapore. On appeal, the Appellant contends that the Appellant’s CAESAR Mark is registered or pending registration in almost 50 countries worldwide, and that it has successfully enforced its rights in the mark in Australia and Canada.

113    In Formula One Licensing BV v Idea Marketing SA [2015] 5 SLR 1349, Tay Yong Kwang J (as he then was) stated (at [66]) that "registrations do not automatically equate to trade mark use”. As for foreign decisions, he held (at [72]) that they do not assist in determining the threshold issue of whether the mark in question is well known to the relevant sector of the public in Singapore. Indeed, although the overseas registrations of the mark and the successful enforcement of rights are relevant factors under s 2(7)(c) and (d) of the TMA, the language of s 2(7) of the TMA makes it abundantly clear that the ultimate inquiry is whether a trade mark is well known in Singapore. The crucial point, therefore, is that the Appellant has to show how the overseas registrations of the Appellant’s CAESAR Mark and the successful enforcement of its rights has led to its mark being well known in Singapore. In our judgment, this has not been done. These factors therefore do not go towards establishing that the Appellant’s CAESAR Mark is well known in Singapore.

Conclusion on whether the Appellant’s CAESAR Mark is well known in Singapore

114    The burden lies on the Appellant to show that the Appellant’s CAESAR Mark is well known in Singapore. Based on what we have found at [103]–[113] above, it is evident that the evidence adduced by the Appellant is, on the whole, weak. Aside from the Appellant’s sales figures, none of the other factors goes towards establishing that the Appellant’s CAESAR Mark is well known in Singapore. As for the Appellant’s sales figures, we accept that they demonstrate that the Appellant had a tangible business in Singapore. However, this alone does not, in our judgment, suffice to establish that the Appellant’s CAESAR Mark is well known in Singapore. The fact that a trader has some business within Singapore will generally be insufficient in itself to establish that the mark is well known. Accordingly, we find that the Appellant’s CAESAR Mark is not well known in Singapore.

Conclusion on Issue 3

115    Our conclusion at [114] above makes it unnecessary for us to consider the fourth element under s 8(4)(a) read with s 8(4)(b)(i) of the TMA (which is the likelihood of damage to the interests of the proprietor). To summarise our views on Issue 3, although we reach a different conclusion from the Judge with respect to the first and third elements of s 8(4)(a) read with s 8(4)(b)(i) of the TMA (see [96] above), we agree with the Judge with respect to the second element (see [114] above). The Judge therefore did not err in finding that the opposition under s 8(4)(a) read with s 8(4)(b)(i) of the TMA was not made out.

Conclusion

116    In relation to the three issues set out at [7] above, we are therefore satisfied that:

(a)     there is no threshold requirement for a "material error of fact or law” to be shown before appellate intervention is warranted in appeals from the Trade Marks Registry (see [24] above);

(b)     the Judge erred in finding that the opposition under s 8(2)(b) of the TMA was not made out (see [93] above); and

(c)     the Judge did not err in finding that the opposition under s 8(4)(a) read with s 8(4)(b)(i) of the TMA was not made out (see [115] above).

117    Accordingly, we allow the appeal. The Respondent’s CAESARSTONE Mark shall not proceed to registration. The Respondent shall pay the Appellant its disbursements and half its costs here and below. For the avoidance of doubt, the latter includes the costs of: (a) this appeal; (b) the hearing before the Judge and; (c) the hearing before the PAR. The costs and disbursements are to be taxed if not agreed. We also make the usual consequential order for the payment out of the security.