Sundaresh Menon CJ (delivering the judgment of the court):
1. These appeals arise from opposition proceedings commenced by Sheraton International, Inc ("Sheraton") and Starwood Hotels & Resorts Worldwide, Inc ("Starwood") (collectively, "the Opponents"). The Opponents opposed the application ("the Application") made by Staywell Hospitality # PARK REGIS Group Pty Ltd ("Staywell") to register a series of two marks ^ PARK REGIS ("the Applicant Mark") in Singapore in respect of marketing and hospitality services. Civil Appeal No 148 of 2012 ("CA 148/2012") is the Opponents’ appeal against the finding of the High Court Judge ("the Judge") that there was no likelihood of confusion or sufficient indication of a connection damaging to the Opponents' interests in order to ground an opposition to registration under ss 8(2)(b) and 8(4)(b)(i) of the Trade Marks Act (Cap 332, 2005 Rev Ed) ("the Act"); and against the Judge's finding that there was no goodwill in the Opponents' ST. REGIS Singapore hotel for the purpose of an opposition under s 8(7)(a) of the Act. Civil Appeal No 147 of 2012 ("CA 147/2012") is Staywell's cross-appeal against the Judge's finding that the Applicant Mark was similar to the Opponents' ST. REGIS registered trade mark ("the Opponent Mark").
Background to the appeals
2. The Opponents are hotel and leisure companies incorporated in the United States. They own and operate 32 St Regis hotels worldwide (17 at the time of the Application)1. In Singapore, Sheraton has been the registered proprietor of the word trade mark "ST. REGIS" in Classes 36 and 37 of the International Classification of Goods and Services ("ICGS") since 2005 (relating to real estate and construction services respectively), and in Classes 41 (relating to entertainment services) and 42 (relating to hotel and hospitality services) since 19952. Class 42 is equivalent to Class 43 of the latest edition of the ICGS, and will be referred to hereafter as "Class 43".
3. Staywell is an Australian hotel operator. It currently owns and operates 27 hotels (24 at the time of the Application) under its two brands, Park Regis and Leisure Inn. These hotels are located mainly in Australia and New Zealand. The Application was Staywell's first application for trade mark registration in Singapore3.
4. Staywell filed the Application to register its series of two marks
LAF PARK REGIS
PARK REGIS on 3 March 2008 in Singapore in Classes 35 and 43 of the ICGS. Class 35 relates to advertising and marketing services. Class 43 relates to hotel services, hospitality, food and beverage and related services. The Opponents' ST. REGIS Singapore hotel opened on 20 April 2008, shortly after the filing of the Application. On 7 May 2008, the Application was accepted and published for opposition purposes. Staywell's hotel, the Park Regis Singapore, only opened for business in November 2010.
5. The Opponents filed their Notice of Opposition4 to the Application on 8 September 2008. They relied on absolute grounds of opposition under ss 7(4)(b) and 7(6) of the Act, and relative grounds of opposition under ss 8(2)(b), 8(4)(b) and 8(7)(a) of the Act. On 19 July 2011, the Principal Assistant Registrar ("the PAR") allowed the opposition under ss 8(2)(b) and 8(4)(b)(i) of the Act, but rejected the opposition under ss 8(4)(b)(ii), 8(7)(a) and 7 of the Act. In relation to the relative grounds of opposition, she found that the competing marks5 and the parties' services were similar6, and that such similarity resulted in a likelihood that the public would be confused into thinking that the Park Regis Singapore was economically linked to the ST. REGIS Singapore7. However, she refused the opposition under s 8(4)(b)(ii) of the Act because she found that the ST. REGIS mark was not "well known to the public at large in Singapore". The opposition under s 8(7)(a) also failed because the Opponents were unable to show that goodwill attached to the ST. REGIS Singapore at the date of the Application8.
6. Staywell appealed against the PAR's decision in relation to ss 8(2)(b) and 8(4)(b)(i) of the Act. The Opponents filed a cross-appeal against the PAR's rejection of the opposition under s 8(7)(a) of the Act. The appeal and cross-appeal were heard by the Judge on 18 April 2012.
The decision below
7. The Judge reversed the PAR's decision under ss 8(2)(b) and 8(4)(b)(i) of the Act and affirmed the decision under s 8(7)(a) of the Act.
8. The Judge's findings in relation to the opposition under s 8(2)(b) of the Act can be summarised as follows:
(a) The Opponent Mark was inherently distinctive in relation to hotel services9. On the whole, the marks were aurally and conceptually
(b) The Class 35 and 43 services for which Staywell sought registration were undoubtedly similar to the Opponents' services registered in Class 43. Staywell's registration in Class 35 was merely complementary to its registration in Class 4311.
(c) Considering the factors set out in The Polo/Lauren Co, LP v Shop In Department Store Pte Ltd  2 SLR(R) 690 ("Polo (CA)")12, including the steps Staywell had taken to differentiate its services from that of the Opponents', there was no likelihood of confusion resulting from the similarity of the competing marks and the similarity of their services13.
9. As regards the s 8(4)(b)(i) opposition, the Judge found as follows:
(a) The level of knowledge of the Opponent Mark in Singapore, as evidenced by the extensive advertising and the number of Singaporean patrons of ST. REGIS hotels worldwide, was sufficiently high to warrant a finding that the Opponent Mark is a well-known mark in Singapore14.
(b) Following the finding that no likelihood of confusion resulted from the similarity of the marks, the use of the Applicant Mark would not indicate a confusing connection between the Park Regis Singapore and the Opponents15.
(c) Even if a confusing connection existed, this would not cause damage to the Opponents. They had shown no intention of moving into the market segment in which the Park Regis Singapore operated16.
10. As regards the s 8(7)(a) opposition, the Judge found as follows:
(a) The Opponents' pre-opening activity in Singapore, and the worldwide reputation in the ST. REGIS brand, was not sufficient to establish that goodwill attached to the ST. REGIS Singapore at the date of the Application17.
11. In the result, the Applicant Mark was allowed to proceed to registration18. On 9 November 2012, the Opponents filed their Notice of Appeal in CA 148/201219 against the whole of the Judge's decision. Staywell also filed its Notice of Appeal20 in CA 147/2012 challenging the Judge's finding of aural and conceptual similarity between the Opponent and Applicant Marks.
Issues in the appeals
12. The following issues and sub-issues arose for our consideration in the two appeals:
(a) Whether the Judge erred in disallowing the opposition under s 8(2)(b) of the Act; in particular:
(i) whether the Judge erred in finding that the competing marks were aurally and conceptually similar ("the marks- similarity inquiry"); and
(ii) whether a likelihood of confusion arose from the similarity between the competing marks and the parties' services ("the confusion inquiry");
(b) Whether the Judge erred in disallowing the opposition under s 8(4)(b)(i) of the Act; in particular:
(i) whether use of the Applicant Mark in relation to hotel services was likely to damage the Opponents' interests; and
(c) Whether the Judge erred in disallowing the opposition under s 8(7)(a) of the Act; in particular:
(i) whether there was goodwill in the ST. REGIS Singapore as at the date of the Application.
Whether the Judge erred in disallowing the opposition under s 8(2)(b) of the Act
Were the marks aurally and conceptually similar?
13. Staywell's case in CA 147/2012 was that that the Judge came to the wrong conclusion as to the aural and conceptual similarity of the competing marks. In particular, it was argued that she fell into error when she dissected the marks and analysed their components discretely in considering their distinctiveness. The Opponents' response was that the Judge did consider the marks holistically, but was entitled to focus on the dominant and distinctive components of each mark. As a preliminary point, the Opponents also argued that as long as a minimal level of marks-similarity was established, the court could and should go on to analyse the likelihood of confusion arising from the similarity of the marks and the goods or services in question21.
Is the threshold of marks similarity a low one?
14. We begin with the preliminary point. The Opponents rely on the decision of Lindsay J in esure Insurance v Direct Line Insurance plc  RPC 6 ("esure Insurance") for the proposition that the similarity of marks is subject to a "low threshold test". They contend that the competing marks more than satisfied this low threshold test. We take this opportunity to restate a number of principles relevant to the marks-similarity inquiry.
15. To the extent that the Opponents' argument treated the similarity of competing marks as a threshold requirement that had to be satisfied before the confusion inquiry is undertaken, we agree. It is clear from the plain words of ss 8 as well as 27 of the Act that the only relevant type of confusion for the purpose of grounding an opposition or an infringement action, is that which is brought about by the similarity between the competing marks and between the goods and services in relation to which the marks are used. Since this court's decision in Polo (CA), our courts have given effect to this statutory wording by applying what is now known as the "step-by-step" approach, as opposed to the competing "global appreciation approach" applied in Europe after Sabel BV v Puma AG, Rudolf Dassler Sport  RPC 199 ("Sabel v Puma"). Under the step-by-step approach, the three requirements of similarity of marks, similarity of goods or services, and likelihood of confusion arising from the two similarities, are assessed systematically. The first two elements are assessed individually before the final element which is assessed in the round. Under the global appreciation approach the elements of similarity between marks and goods or services, whilst still necessary ingredients in the confusion inquiry, are elided with other factors going towards the ultimate question of whether there is a likelihood of confusion (see Sabel v Puma at 223-224, and Canon Kabushiki Kaisha v Metro-Goldwyn-Mayer Inc  RPC 117 ("Canon'") at 132). Whilst there have been suggestions that the two approaches might be distinct without being different, we maintain this dichotomy and endorse the step-by-step approach as being conceptually neater and more systematic and, importantly, as being more aligned with the requirements imposed under our statute (see Polo (CA)) at ).
16. However we do not agree with the Opponents' approach to the extent that it suggests that any particularly or notably low threshold of marks-similarity applies. For one thing, such an approach finds no support in the case law. The English Court of Appeal disapproved of Lindsay J's "low threshold test" when esure Insurance went on appeal in esure Insurance Ltd v Direct Line Insurance PLC  EWCA Civ 842. Arden LJ there stated that "no useful purpose" was served by holding that there was a minimal level of similarity that had to be shown; and she went on to hold that such an approach was unsupported by European jurisprudence (at ). Nor does our own case law lend any support to such an approach.
17. More fundamentally, the minimal threshold approach is inconsistent with the reality that the similarity of marks is ultimately and inevitably a matter of impression rather than one that can be resolved as a quantitative or mechanistic exercise (see Polo (CA) at  and Wagamama Ltd v City Centre Restaurants Plc  FSR 713 at 732). The learned Amicus, Prof Ng-Loy Wee Loon submitted that while the court should consider each of the three established aspects of similarity, viz, visual, aural or conceptual similarity, as long as it found that there was some degree of similarity in any one of these three aspects, no matter how weak, the marks-similarity requirement should be considered to have been met and the court would then be obliged to proceed to the next stage of the inquiry which will generally be the likelihood of confusion. We respectfully disagree. The court must ultimately conclude whether the marks, when observed in their totality, are similar rather than dissimilar. The three aspects of similarity are meant to guide the court's inquiry but it is not helpful to convert this into a checkbox exercise in which a tick, however faint it might be, in any one box must compel the court to find that the marks are similar when a sensible appraisal of the marks as a whole would show otherwise.
18. We observed this in Hai Tong Co (Pte) Ltd v Ventree Singapore Pte Ltd  SGCA 26 ("Hai Tong") at [40(a)]. Congruously, there is no prescribed requirement that all three aspects of similarity must be made out before the marks can be found to be similar: MediaCorp News Pte Ltd v Astro All Asia Networks plc  4 SLR(R) 496 ("MediaCorp") at  and Sarika Connoisseur Cafe Pte Ltd v Ferrero SpA  1 SLR 531 ("Sarika") at . In short, the 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. Trade-offs can occur between the three aspects of similarity in the marks-similarity inquiry: Ozone Community Corp v Advance Magazine Publishers Inc  2 SLR 459 ("Ozone'), see also Bently and Sherman, Intellectual Property Law (Oxford University Press, 3rd Ed, 2009) ("Bently & Sherman") at p 864.
19. A further problem with adopting an approach in which any modicum of similarity would compel the court to make a finding of marks-similarity is that this would backload much of the court's assessment to the confusion inquiry stage. We are wary of this and, in any event, do not view it as being faithful to the scheme of our statutory framework. A productive and appropriate application of the step-by-step approach necessitates that the court reach a meaningful conclusion at each stage of the inquiry.
20. Finally, on this issue, we reiterate that the assessment of marks similarity is mark-for-mark without consideration of any external matter: see Mediacorp at , Sarika at  and Hai Tong at [40(b)]. This means that at the marks similarity stage this even extends to not considering the relative weight and importance of each aspect of similarity having regard to the goods. This does not mean that the court ignores the reality that the relative importance of each aspect of similarity might vary from case to case and will in fact depend on all the circumstances including the nature of the goods and the types of marks, as we observed at [40(b)] of Hai Tong. Rather, such considerations are properly reserved for the confusion stage of the inquiry, because that is when the court is called upon to assess the effect of objective similarity between the marks, on the perception of consumers. We recognise that this reflects a slight departure from the approach taken by the High Court in Festina Lotus SA v Romanson Co Ltd  4 SLR 552 at -, and by this court in Sarika at . We think that this is conceptually clearer because it recognises that the issue of resemblance between the competing marks is distinct from the question of the effect of such resemblance. A practical application of this approach can be found in European jurisprudence: see Mystery drinks GmbH v OHIM, T 99/01  ETMR (18) 217 and Phillips-Van Heusen Corp v OHIM  ETMR 60. In these cases the court considered, respectively, the particular significance of aural similarity in relation to beverages normally sold by oral order, and visual similarity in relation to clothing normally sold based on the consumer's direct perception, both for the specific purpose of determining whether consumer confusion was likely to arise.
Whether the Judge erred in finding that the marks were aurally and conceptually similar, having regard to their distinctiveness
21. We turn to consider the Judge's finding that the competing marks in this case were similar. Staywell's case was that the Judge wrongly focussed on the "REGIS" element of both marks, instead of considering the aural and conceptual similarity of the marks as a whole. In particular, Staywell contended that the Judge erred in treating "REGIS" as distinctive when in fact the Opponents had never used the word "Regis" on its own, but always as part of the composite word "St Regis".
22. Before turning to the particular facts before us, we first distinguish between two different aspects of "distinctiveness" in trade mark law. This was articulated by Lord Walker in BUD And Budweiser Budbrau Trade Marks  RPC 25 at , as follows:
It is not necessary to go far into the authorities on trade mark law to see that the term "distinctive" is used with two different shades of meaning. Sometimes it has its normal, non-technical meaning, with a flavour of that which is unusual (or stands out in a crowd) and is therefore easy to recognise and to remember. Sometimes it is used in a more technical sense, in contrast to "descriptive" (reflecting Art.3.1(c) of the Directive and s 3(1)(c) of the Act ). Purely descriptive or laudatory words, however striking and memorable, cannot normally be distinctive in this sense. ... But they may in some circumstances acquire distinctiveness...
23. Distinctiveness in the ordinary and non-technical sense simply refers to what is outstanding and memorable about the mark. Such components tend to draw the consumer's attention, bearing in mind the imperfect recollection of the average customer. This reflects the reality that the average consumer "only rarely has the chance to make a direct comparison between the different marks and must place his trust in the imperfect picture of them that he has in his mind" (Lloyd Schuhfabrik Meyer v Klijsen Handel BV, Case C-342/97  2 CMLR 1343 ("Lloyd') at 1358 and Ozone Community Corp v Advance Magazine Publishers Inc  2 SLR 459 ("Ozone Community") at ). The distinctive (in the non-technical sense) and memorable components of the mark are those that tend to stand out in the consumer's imperfect recollection. That is why the court is entitled to have special regard to the distinctive or dominant components of a mark, even while it assesses the similarity of the two marks as composite wholes.
24. Distinctiveness in the technical sense on the other hand, usually stands in contradistinction to descriptiveness. Where the latter connotes words that describe the goods or services in question, or of some quality or aspect thereof, the former refers to the capacity of a mark to function as a badge of origin. Distinctiveness can be inherent, usually where the words comprising the mark are meaningless and can say nothing about the goods or services; or acquired, where words that do have a meaning and might well say something about the good or services, yet come to acquire the capacity to act as a badge of origin through long-standing or widespread use (see Windsurfing Chiemsee Produktion und Vertriebs GmbH v Boot und Segelzubehor Walter Huber and Franz Attenberger  ETMR 585 (ECJ), Love & Co Pte Ltd v The Carat Club Pte Ltd  1 SLR(R) 561 at -  and Hai Tong at -).
25. Technical distinctiveness is an integral factor in the marks-similarity inquiry (see Sarika at , Ozone Community at  and Polo (CA) at ); a mark which has greater technical distinctiveness enjoys a high threshold before a competing sign will be considered dissimilar to it (Sarika at ). While the components of a mark may be inherently technically distinctive, ultimately the ability of the mark to function as a strong badge of origin must be assessed by looking at the mark as a whole. Conversely, the components of a mark may not be inherently distinctive, but the sum of its parts may have sufficient technical distinctiveness. We examined this in Hai Tong in relation to the composite phrase "Lady Rose", and there stated (at ):
...the attempt to contend that this should be the case because of any lack of distinctiveness in the Composite Mark was unfounded. Of course, the words "Lady" and "Rose", when used separately, are not inventive words or phrases. However, in our judgment, it would not be correct to analyse this by reference to each of the words standing alone. It is the juxtaposition of the words that engenders distinctiveness, in that when used together, the words do not convey a sensible meaning. ...
26. When speaking of the assessment of a mark as a whole, it may be noted that the cases have consistently stated that the "visual, aural or conceptual similarity of the marks in question, must be based on the overall impression given by the marks, bearing in mind, in particular, their distinctive and dominant components" (Sabel v Puma at 224, Matratzen Concord GmBH v OHIM, Case T-6/01  ECR II-4335 ("Matratzen"), Doctor's Associates Inc v Lim Eng Wah (trading as SUBWAY NICHE)  3 SLR 193 ("Subway Niche") at  and Bently & Sherman at p 864).
27. In Matratzen the European Court of First Instance dismissed the contention that a consideration of the dominant and distinctive components of a mark would mean a failure to consider the mark as a whole. After stating that the similarity between two marks must be assessed globally, but in light of their distinctive and dominant components, the court said (at ):
It should be made clear that that approach does not amount to taking into consideration only one component of a complex trade mark and comparing it with another mark. On the contrary, such a comparison must be made by examining the marks in question, each considered as a whole. However, that does not mean that the overall impression created in the mind of the relevant public by a complex trade mark may not, in certain circumstances, be dominated by one or more of its components.
This was endorsed by the European Court of Justice ("ECJ") on appeal in Matratzen Concord GmBH v OHIM (C-3/03 P) at -. More recently in 32Red Plc v WHG (International) Ltd  RPC 2012 the English Court of Appeal dismissed the argument that the judge below had erred in deconstructing the competing composite word marks (viz, "32RED" and "32vegas") and finding "32" to be the dominant component (at  and ). Our courts have taken the same approach in cases involving competing marks with a common denominator, such as in The Polo/Lauren Co, LP v Shop In Department Store Pte Ltd  4 SLR(R) 816 ("Polo (HC)") at  and Richemont International SA v Goldlion Enterprise (Singapore) Pte Ltd  1 SLR(R) 401 at  ("Richemont"). The question in those cases was whether the common element of the competing marks was so dominant as to render the different elements ineffective to obscure the similarity between the marks.
28. It follows from this that it is not wrong for a court to find that a component of a mark is inherently technically distinctive - for example if it is a non-descriptive word or an elaborate and inventive graphic device. Where a particular element or component has a high degree of technical distinctiveness, this can have a bearing on whether as a result of this, that component or element is found to be the dominant and distinctive element of the mark in the non-technical sense.
29. The finding of distinctiveness of the separate components of the mark must ultimately be related back to the impression given by the mark as a whole. The distinctiveness of a particular component of a mark is but one factor feeding into the ultimate question of whether the mark, in the form it is registered and/or used, has strength as an indicator of origin to the exclusion of other trade sources. This latter question clearly must be considered by looking at the mark as a whole, because it is the entire mark, and not only a component of it, that must function as the badge of origin.
30. We turn to examine whether the Judge applied these principles correctly in the instant case. Perhaps for ease of analysis the Judge approached the inquiry in a two-stage process as was done in Ozone Community (see the Judge's Grounds of Decision ("GD") at ). We reiterate, as was held in Sarika (at ) and in Hai Tong (at ), that distinctiveness (in both its technical and non-technical senses) is a factor integrated into the visual, aural and conceptual analysis as to whether the competing marks are similar. It is not a separate step within the marks-similarity inquiry.
31. At ,  and  of the GD, the Judge considered the word "REGIS" to be the dominant and distinctive component of both marks when each is read out. In our view, she was perfectly entitled to come to this view. Staywell's argument that the Judge erred in considering the dominance of "REGIS" because the Opponents had never used or registered the component "REGIS" alone was misconceived simply because, as we have pointed above, a component can clearly be dominant even if it is part of the mark as a whole and is not and has never been used on its own. In fairness to the Judge, it should be noted that she specifically considered that the "ST." and "Park" portions of the competing marks, though not as dominant as the "Regis" portion, were not to be ignored. Staywell's argument that any distinctiveness of the Opponent's mark lay in "ST. REGIS" as a whole rather than in "REGIS" overlooked the permissibility of examining the distinctive components of the competing marks in both the technical and non-technical senses. At least in relation to hotels and hospitality services, "REGIS" enjoys a substantial degree of technical distinctiveness. In relation to both the competing marks "REGIS" is the element that is distinctive in the nontechnical sense because it is what will stand out in the imperfect recollection of the consumer. The Judge was therefore entitled to find this the common dominant element of both marks in assessing the question of whether the competing marks as a whole were similar. The Judge found on this basis that the marks were aurally similar and we agree.
32. An alternative approach which does not involve considering the dominant components of the marks would have been simply to undertake a quantitative assessment as to whether the competing marks have more syllables in common than not. This was the approach taken in Ozone Community at . Even on this approach it is clear that the two marks, each taken as a unified whole, have more syllables in common than not.
33. We therefore see no reason to disturb the Judge's finding that the marks had aural similarity. In fact we find that there is a high degree of aural similarity between the competing marks because of the distinctiveness of the common "REGIS" component in both the technical and non-technical senses.
34. Turning to the conceptual analysis, there are a few aspects of the Judge's decision with which we find ourselves, with respect, in disagreement.
35. First, the Judge gave weight to the Opponents' argument that the competing marks shared the underlying connotation of royalty, since the "REGIS" component of the marks was dominant. As we have already noted, the Judge was entitled to conclude that "REGIS" was the distinctive component from an aural perspective. However, we do not agree that "REGIS" was a dominant component on a conceptual analysis. Unlike the aural analysis, which involves the utterance of the syllables without exploring the composite meaning embodied by the words, the conceptual analysis seeks to uncover the ideas that lie behind and inform the understanding of the mark as a whole (Bently & Sherman at p 866). Greater care is therefore needed in considering what the conceptually dominant component of a composite mark is, because the idea connoted by each component might be very different from the sum of its parts. The case of Vedial SA v OHIM - France Distribution (HUBERT), Case C-106/103 P  ECR I-9573 is illustrative. There, the European Court of Justice upheld the Court of First Instance's finding that the mark for which registration was sought was visually, aurally and conceptually dissimilar to the earlier word trade mark "SAINT-HUBERT 41". At  the Court of First Instance in Vedial SA v OHIM - France Distribution (HUBERT), Case T-110/01  ECR II-5275 stated that:
With regard to conceptual analysis of the marks in question, it must be observed that the ideas suggested by the terms 'SAINT-HUBERT' and 'HUBERT' are different. The combination of the terms 'SAINT and "HUBERT' with a hyphen between them creates a concept and logical unit that is distinct from those of its components. Thus the words 'SAINT-HUBERT' form an inseparable whole which is likely to evoke in the mind of the targeted public a saint of the Catholic religion or a name of a place. The word 'HUBERT', on the other hand, corresponds to a common masculine French Christian name.
36. In the present case, while the term "REGIS" might, perhaps reasonably, be said to connote royalty, in our view this connotation becomes secondary once the prefixes of "Park" and "ST." are introduced. As the Judge # PARK REGIS noted at  of the GD, the composite mark VPARK REGIS connotes a geographical location, while the composite phrase "ST. REGIS" connotes a saintly character. If any connotation of royalty attaches at all, it is far more significant in the "Park Regis" mark, considering the inclusion of the fleur-de- lis device. In this connection we note that the Judge erred in treating the device as a reinforcement of the concept of royalty. The device was not part of the Opponent Mark, and the authorities are clear that marks-similarity must be assessed without regard to "external added matter or circumstances" (Hai Tong at [40(b)]). In light of these considerations, we find that the concept of a geographical location perhaps bearing some class and prestige as connoted by the inclusion of the fleur-de-lis device in the Applicant Mark, is quite different from the concept of a Catholic saint that is connoted by the Opponent Mark.
37. However, we find that there is some conceptual similarity between the marks but for a different reason. There is some strength in the Opponents' argument that both the Applicant and Opponent Marks evoke the idea of a place or location22. Although the ST. REGIS mark might conjure the image of a saintly person, in Singapore by far the most common manifestation of saintly names is in relation to places ranging from schools (St Joseph's Institution, St Margaret's School), roads (St George's Road, St Martins Lane), hospitals (St Luke's Hospital) and condominiums (St Nicholas View, St Patrick's Court), to an assortment of other public places (St Michael's flatted factory, St James Power Station (bar), Le Saint Julien (restaurant)), many of which have no particular association whatsoever with the saintly character whose name they have adopted. Similarly, the word "Park", as noted by the Judge, connotes the idea of a geographical location though not necessarily of any particular type (see  of the GD). Therefore, in the Singapore context, the marks share a tendency to connote a place or location or building which, as between the competing marks in question, are linked by the common reference to "Regis".
Conclusion on similarity of the competing marks
38. Based on the foregoing analysis, we find that the marks have a substantial degree of aural similarity and a fair degree of conceptual similarity. The PAR's finding that the marks were visually dissimilar was not appealed. On the whole, and based on what we have set out above, we have no difficulty finding that the marks were similar. On this basis, we dismiss CA 147/2013.
Was there similarity between the parties' services?
39. The parties did not appeal the Judge's finding that the services in relation to which their respective marks were used or for which registration was sought, were similar. However, we take this opportunity to clarify certain aspects of the Judge's reasoning.
40. The comparison before us was between the services falling under Staywell's intended Class 35 and 43 registrations, and the Opponents' existing Class 43 registration. In relation to the Class 43 registration, we agree with the Judge's observation that the fact that the parties' hotel services were branded for different market segments did not render the services dissimilar. Following the dictum of Lai Kew Chai J in Polo (HC) at , the Judge stated that registration in the same category establishes a prima facie case for similarity. This invites some clarification. We think that what Lai J was referring to was registration in the same specification. We would go further to say that registration in the same specification within a class establishes a prima facie case for identity. This is because it is not within the scheme of the classification system to make distinctions within a specification based on whether the particular product is targeted at one or another market segment.
41. Hotel services are hotel services, whether these concern a luxury hotel or a more modest one. We adopt the view that while "trade mark registrations should not be allowed such a liberal interpretation that their limits become fuzzy and imprecise... [w]here words or phrases in their ordinary and natural meaning are apt to cover the category of goods in question, there is equally no justification for straining the language unnaturally so as to produce a narrow meaning which does not cover the goods in question." (per Floyd J in YouView TV Ltd v Total Ltd  EWHC 3158 (Ch) ("YouView") at ; see also Omega Engineering Inc v Omega SA  FSR 25 at ). Where a good or service in relation to which registration is sought falls within the ambit of the specification in which the incumbent mark is registered, the competing goods or services would be regarded as identical (see Gerard Meric v OHIM, Case T-133/05 at ).
42. In this case there was a precise overlap between Staywell's Class 43 hotel, food and beverage services, and the Opponents' Class 43 hotel, food and beverage services. We do not think that there was any qualitative difference between the parties' services that displaced the prima facie case for identity where registration was sought in the same specification within Class 43. Therefore the requirement of similarity or identity between services under s 8(2)(b) is satisfied in this case. Given that the services in question are identical, there is no further need to consider whether they are similar or the extent of their similarity (see Hai Tong at  and Susanna Leong, Intellectual Property Law of Singapore (Academy Publishing, 2012) ("SusannaLeong') at paras 28.354 and 28.358).
43. In relation to its Class 35 registration, the Judge accepted the PAR's finding that Staywell's advertising, marketing and business management services were adjunct to and inseparable from its primary services under Class 43, and therefore similar to the Opponent's Class 43 services (GD at - ). We agree with the Judge's reasoning to the extent that Staywell's Class 35 services are closely related and indeed inseparable from its Class 43 hotel and restaurant services. This is not to say that Staywell's application for registration under the two classes was to be considered and compared collectively as against the Opponents' Class 43 services. Rather, the real question is whether Staywell's services that were sought to be registered under Class 35 are similar to the Opponents' services under Class 43, having regard to all relevant factors relating to the services themselves. Some of the factors set out in British Sugar Plc v James Robertson & Sons Ltd  RPC 281 ("British Sugar") at 296 would be relevant at the goods-similarity stage of the inquiry, in particular the consideration of the uses and the end-users of the services. The question is how the services are regarded, as a practical matter, for the purposes of trade. Applying this approach, we agree with the PAR's findings at  of the PAR's GD. There is an overlap between the uses and users of Staywell's Class 35 business management and administration services, and its Class 43 service of providing hotel venues for business conferences and receptions. As for the advertising and marketing services under Class 35, these relate to the promotion of the hotel and restaurant services under Class 43. Therefore we agree that there is similarity between Staywell's services in Class 35 and the Opponents' services in Class 43.
Was there a likelihood of confusion?
44. We come to one of the key areas of dispute in CA 148/2012. The question was whether, having found that the competing marks and services were similar or identical as the case may be, the Judge erred in her assessment of whether this gave rise to a likelihood of confusion for the purpose of s 8(2)(b) of the Act, or a confusing connection under s 8(4)(b)(i) of the Act. In concluding that such a likelihood did not exist, the Judge considered factors external to the marks and services themselves, taking into account the actual circumstances of the uses to which Staywell had put its mark prior to the Application. Her considerations extended to the parties' marketing and website design, the modes in which hotel bookings are usually made, and the star-ratings of the parties' hotels. Two issues arose in relation to the Judge's approach. First, whether there is a difference between the approach to the confusion inquiry in opposition as opposed to infringement proceedings. Second, whether factors which are extraneous to the competing marks themselves and to the products in relation to which the marks are used ("extraneous factors"), are relevant to the confusion inquiry, and if so whether there are any applicable limits to this.
45. We will address each of these issues in turn.
Difference in approach to the confusion inquiry as between opposition and infringement proceedings
46. The starting point must be the Act itself. The language used in each of ss 8(2) and s 27(2) of the Act is similar. There is no indication within the Act itself that these sections should be interpreted differently, and we have previously observed in passing that the approach to determining confusion for the purpose of opposition to registration is no different than that in an infringement action (see Hai Tong at [84(b)] and [85(a)], Valentino Globe BV v Pacific Rim Industries Inc  2 SLR 1203 (" Valentino") at  and Polo (CA) at ). However, the present appeal presents the opportunity to examine the question in greater detail.
47. In Europe and the UK, the courts have drawn a distinction between the conduct of the confusion inquiry in opposition and in infringement proceedings. In the latter, the court assesses the likelihood of confusion in relation to the defendant's actual and allegedly infringing use of his sign. On the other hand, in the former context, ie, in opposition proceedings, the assessment is conducted by considering the notional fair use of the applicant mark across the scope of the specification of goods and services in respect of which registration is sought. Such a test is obviously more stringent and offers a greater degree of protection to the incumbent than one that examines only the actual and allegedly infringing use of the later mark.
48. The leading European case is O2 Holdings Ltd v Hutchison 3G UK Ltd, Case C-533/06  ECR I-4231 ("O2 (ECJ)"). A question concerning the interpretation of Art 5(1) of the European First Council Directive 89/104 of 21 December 1988 ("the Directive") was referred to the ECJ pursuant to infringement proceedings commenced under ss 10(2) and (3) of the Trade Marks Act 1994 (c 26) (UK) ("the UK Act"). The dispute involved the defendant's use of a sign similar to the claimant's UK trade mark in its comparative advertisement campaign. Of relevance to the present case is the ECJ's statement at - where it clearly articulates the point:
It is true that the notion of likelihood of confusion is the same in Art.4(1)(b) of Directive 89/104 [which deals with the prevention of registration of similar marks likely to cause confusion].
Art.4(1)(b) of Directive 89/104, however, concerns the application for registration of a mark. Once a mark has been registered its proprietor has the right to use it as he sees fit so that, for the purposes of assessing whether the application for registration falls within the ground for refusal laid down in that provision, it is necessary to ascertain whether there is a likelihood of confusion with the opponent's earlier mark in all the circumstances in which the mark applied for might be used if it were to be registered.
By contrast, in the case provided for in Art.5(1)(b) of Directive 89/104, [which deals with the prevention of use of similar marks likely to cause confusion] the third-party user of a sign identical with, or similar to, a registered mark does not assert any trade mark rights over that sign but is using it on an ad hoc basis. In those circumstances, in order to assess whether the proprietor of the registered mark is entitled to oppose that specific use, the assessment must be limited to the circumstances characterising that use, without there being any need to investigate whether another use of the same sign in different circumstances would also be likely to give rise to a likelihood of confusion.
49. The English High Court followed this approach in Och-Ziff Management Europe Ltd v Och Capital LLP RPC 33 ("Och-Ziff"). Arnold J there stated at  that:
It is common ground that it is now clear that there is an important difference between the comparison of marks in the registration context and the comparison of mark and sign in the infringement context, namely that the former requires consideration of notional fair use of the mark applied for, while the latter requires consideration of the use that has actually been made of the sign in context. This was established by the judgment of the Court of Justice in .
50. O2 (ECJ) and Och-Ziff both concerned infringement proceedings, and the distinction drawn there has been endorsed in other infringement decisions (see Red Bull GMBH v Sun Mark Ltd  EWHC 1929 (Ch) ("Red BulF) at  and Datacard Corp v Eagle Technologies Ltd  RPC 17 at ). In addition, the English court considers the notional fair use of the incumbent mark when assessing for confusion in infringement proceedings, thereby illustrating the breadth of the proprietary rights which the court protects once a mark is on the register. In Compass Publishing BV v Compass Logistics Ltd  RPC 41 ("Compass Publishing') the High Court stated as follows (at  and ):
It must be borne in mind that the provisions in the legislation relating to infringement are not simply reflective of what is happening in the market. It is possible to register a mark which is not being used. Infringement in such a case must involve considering notional use of the registered mark. In such a case there can be no confusion in practice, yet it is possible for there to be a finding of infringement. Similarly, even when the proprietor of a registered mark uses it, he may well not use it throughout the whole width of the registration or he may use it on a scale which is very small compared with the sector of trade in which the mark is registered and the alleged infringer's use may be very limited also. In the former situation, the court must consider notional use extended to the full width of the classification of goods or services. In the latter it must consider notional use on a scale where direct competition between the proprietor and the alleged infringer could take place.
The reason why the rule of thumb referred to above [that absence of actual confusion in the market means no infringement of a registered trade mark] does not give a safe indication of whether there is infringement in this case is because of the nature of the parties' respective presences in the market. They are not in competition with each other. The business consultancy field is enormous. Indeed, on the basis of the evidence before me, the logistics section of the business consultancy field is enormous. The claimant's core activities are not in the logistics field, the defendant's are. Furthermore, even within that field, the defendant is a very small player, as will be explained below. In those circumstances it is not surprising that there has been no confusion in the marketplace. To date the claimant and the defendant trade in different parts of the market. This does not come close to imitating the notional world used for determining likelihood of confusion under Art.9.1 (b).
51. There are also other English cases in which the notional fair use approach has been applied when assessing the likelihood of confusion in the opposition context. Some of these cases were decided under s 11 of the now- repealed Trade Marks Act 1938 (c 22) (UK). While the wording of that section differed from s 8 of the Act, it did require the court to refuse registration upon finding that the use of the later mark is "likely to deceive or cause confusion". In that context it has been held that the question is whether the mark, "if used in a normal and fair manner in connection with any of the goods covered by the registration proposed, will not be reasonably likely to cause deception and confusion amongst a substantial number of persons" (see re Smith Hayden & Co's Application (1945) 63 RPC 97, Berlei (UK) Ltd v Bali Brassiere Co Inc (No.1)  1 WLR 1306 ("Bali Trade Mark"), Gymboree Corp's Trade Mark Application (No.1577166) (1999) 22(6) IPD 2205, OPEN COUNTRY Trade Mark  RPC 477 ("Open Country")). In Bali Trade Mark, where there had been actual use of the applicant mark, Lord Morris noted as follows (at 1314):
There is no reason to suppose that the learned judge did not remain mindful of such evidence as pointed to certain differences (such as those of method of manufacture and of style and of price) which existed between the articles respectively manufactured. Thus, there was evidence that the Bali brassieres are at present tailor-made and might be regarded as of different quality from those which are mass produced by Berlei. There is a difference in the colours of the packages in which the respective goods are marketed. But methods of manufacture and style of product may change from time to time and prices may vary and be adjusted. What has to be had in mind is the use that could reasonably and properly be made of a mark if fairly and normally used.
52. Lord Morris' observations demonstrate that the court had its eye on both the evidence of actual use of the marks on current products and also the potential uses to which the marks could fairly be put in the future in relation to the class of goods in question. In Open Country, the Court of Appeal stated that the approach in Bali Trade Mark was applicable whether the applicant had used his mark or not. However where there was actual use, such use would be considered as one instance of the normal and fair uses of the mark unless the opponent submitted otherwise. In this regard Aldous LJ stated (at 482) that:
It does not follow that the way that the applicant has used his trade mark is the only normal and fair manner. However in many cases actual use by an applicant can be used to make the comparison.
53. The notional fair use approach is also applied to opposition cases under the current UK Act. The cases show that the notional fair use of the mark is the basis for the analysis not only at the confusion stage (see Advanced Perimeter Systems Ltd v Keycorp Ltd  RPC 14 at  and -) but also when assessing whether the mark has distinctive character in relation to the products for which it is sought to be used (see Bongrain SA's Trade Mark Application (No 2134604)  EWHC 531 (Ch) at ). In YouView the court stated in relation to opposition proceedings that (at ):
... [the application] must be considered for the purposes of the opposition on the basis of notional and fair use across the breadth of the goods and services for which it is registered.
Actual use is not required, and so an enquiry into the goods for which the mark has been used is of not more than background interest, if that.
54. The English approach to the confusion inquiry in opposition proceedings is therefore to consider the notional fair use to which each mark can be put, though regard will also be had to the applicant's actual use of the mark if there has been any. We set out this comparative position by way of background only because it is necessary, ultimately, to interpret and apply our Act in our own context.
55. Under our law, the two threshold requirements for successfully opposing a proposed registration, or establishing liability for infringement, are similarity or identity of the marks and similarity and identity of the services. Once these threshold criteria have been met, the issue of the likelihood of confusion arises and this in our view directs the court to look at (a) how similar the marks are (b) how similar the services are and (c) given this, how likely the relevant segment of the public will be confused. In Hai Tong we said (at [85(c)]):
Having regard to the express terms of s 27(2), there are three specific elements that plainly must be considered. These are:
(i) the similarity between the registered mark and the allegedly infringing mark; (ii) the similarity or identity between the goods or services in relation to which the marks are used; and (iii) the relevant segment of the public in relation to whom the court must consider the likelihood of confusion. Each of these elements can vary. The marks may be identical or similar, and if the latter, they can vary in their degree of similarity. In the same way, the goods or services in relation to which the marks are used may be identical or similar, and again, if the latter, they may vary in the degree or extent to which they are similar. ... And as to the relevant segment of the public, there may be characteristics that are particular to the group in question. Each of these factors will have a bearing on the likelihood of confusion. As an illustrative proposition, the likelihood of confusion must be greater where, say, the contesting marks are similar to a high degree, the goods are identical in nature and the segment of the public in question is undistinguished in its attention than would be the case if the marks and the goods are somewhat similar but not exceedingly so, and the relevant segment of the public happens to be highly knowledgeable and very fastidious. .
56. Against that background we turn to the question posed in this section of our analysis. On reflection we are satisfied there is a difference between the approach to the confusion inquiry in opposition and infringement proceedings, although there are considerable overlaps as well. If it seems surprising that different outcomes may ensue from the interpretation of essentially the same text, it is because of the context in which each provision is to be applied.
57. In opposition proceedings, the contest is between the holder of an existing registered trade mark who opposes the proposed application, and the applicant who is seeking to register a new mark. The opponent enjoys certain monopoly rights associated with the use of its mark and it opposes the registration of the applicant's mark on the grounds that such registration would entail an unwarranted interference with those monopoly rights, whether or not these are already being exercised. The applicant on the other hand, wishes not just to use the mark in a particular way but to have the associated monopoly rights that would enable it to prevent others from using their own marks which would then threaten to dilute or affect the applicant's desired monopoly.
58. It is useful to note that under s 26 of the Act, registration confers the exclusive right on the proprietor to use the mark "in relation to the goods or services for which the mark is registered". The proprietor also acquires immunity from infringement liability in his use of the mark in respect of all goods and services falling under the registered specification, by virtue of s 28(3) of the Act. This means that once the applicant has registered his trade mark, he acquires the exclusive right to use the mark not only for the goods and services which he might have actually contemplated at the time registration was granted, but for the whole spectrum of goods and services within the specification for which the mark is registered. The proprietor is entitled to sue for infringement if anyone uses an identical or similar sign on goods which are identical or similar to goods or services falling within the specification for which his mark is registered, even if he does not actually supply that good or service at the material time (see Sarika at -, see also Compass Publishing at  extracted at  supra). The principle is well put in George Wei, Some Thoughts on Intellectual Property Rights in Singapore: A Monograph for Gerald Dworkin (2009) at para 3.204:
Modern registered trade mark protection is also about rights to protect indicia of origin in related fields of use: areas of business where the registered proprietor may want to diversify into. Thus a manufacturer of "soda pop" who markets the product under a particular TM might want to expand his business into "fruit juices" or "mineral water" or "tonic drinks". Diversification is the key to survival for many companies just as specialisation and focus is the key for some others. Thus, it is important that registered trade mark protection provide rights that reach use of the same or similar registered marks on or in respect of similar goods or services:
even if the registered proprietor has not yet actually ventured into that related area.
59. It follows from the foregoing that a party that applies for registration of a mark is seeking to establish a monopoly over the right to use that mark for the immediately intended purpose as well as for a penumbra of notional fair uses.
60. Accordingly in opposition proceedings the inquiry must take into account the full range of the competing monopoly rights that are already enjoyed on the one hand, namely the actual and notional fair uses to which the incumbent proprietor has or might fairly put his registered trade mark, and compare this against the full range of such rights sought by the applicant by reference to any actual use by the applicant (assuming there has been prior use) as well as notional fair uses to which the applicant may put his mark should registration be granted. This is the setting in which the question of whether there is a likelihood of confusion is assessed. In this regard we agree with the views of Prof Ng-Loy and those of Professor David Llewelyn in his article Is There Confusion in the Law of Trade Marks in Singapore? (2013) 25 SAcLJ 339 at paras 11-12:
When deciding that there was no likelihood of confusion, the judge considered it important that the applicant used its mark on four-star hotels whereas the opponent used its registered mark in relation to six-star luxury hotels: "What is relevant in this enquiry [as to source confusion] is how both the hotels are marketed." However, this seems to ignore the fact that a trade mark proprietor is entitled to use its mark on or in relation to all the goods and/or services for which it is registered, that is, those set out in the specification of goods or services, and no registered mark may be infringed by use of another registered mark on or in relation to the goods or services for which it is registered. Thus, the owner of the PARK REGIS registered mark is free to use its now-registered mark on or in relation to luxury hotels notwithstanding the fact that it has not done so to date. If it were to do so, the only remedy available to the owners of the ST REGIS trade mark would be to sue for passing off, a cause of action far more uncertain (and expensive, as it requires proof of goodwill and actual deception) than that of trade mark infringement.
It is for this reason that, in the EU regime, the relevant test to be applied under Art 4(1) of the Trade Marks Directive (implemented in the UK in s 5(2) of the Trade Marks Act 1994, which is in terms identical to s 8(2) of the Trade Marks Act for the purposes of an application to register is to compare the earlier mark as used with the mark being applied for, taking into account notional and fair use of that later mark in relation to all the goods or services in the specification, and decide whether there is a likelihood of confusion. Conversely, although the wording of the infringement provision (s 8(2) of the Trade Marks Act in Singapore) is the same as the registration provision (s 27(2) of the Trade Marks Act in Singapore), for the purposes of infringement the court should take into account the actual use (if any) by the defendant of its sign, as that is relevant to whether there is a likelihood of confusion.
61. In infringement proceedings on the other hand, there is no question of the alleged infringer seeking to establish any monopoly rights. The only question is whether the actual use of a similar or identical sign by the infringer encroaches on the registered proprietor's monopoly rights under s 26 of the Act, to use the mark in relation to goods and services for which he is already using it as well as the penumbra of fair uses for which he might want to use it. There is therefore no need to examine any notional fair use by the alleged infringer because he is not seeking to acquire or assert any rights in respect of such penumbral uses.
62. How is the foregoing analysis affected where the applicant in opposition proceedings has been using his mark prior to the filing of his registration application? To the extent that the applicant is already using the mark the allegation by the opponent must be that this actual use is to be treated as a putative infringement, so it can be assessed from that perspective. But we do not think the inquiry ends there. If we hold in such a case that there is no putative infringement based on the actual use, it does not mean that the opposition fails and the application for registration must succeed. It would still be necessary to consider the wider question of whether the notional fair uses that the applicant might put the mark to could conflict with the notional fair uses to which the proprietor of the registered mark could put his mark to. As we have noted, this latter inquiry sets a higher threshold for the applicant than an inquiry that focuses only on whether the actual use is infringing, and it follows that as a practical matter, in opposition proceedings, the applicant will have to meet that higher threshold regardless of whether there has already been actual use. In essence, in such proceedings, he will be required to establish that the notional fair use of his mark would not infringe the notional fair use rights of the registered proprietor; whereas in infringement proceedings the only question is whether the actual use by the alleged infringer infringes the notional fair use rights of the registered proprietor of the mark. As a result, in certain cases it may be that opposition and infringement proceedings yield different outcomes. As noted in Kerly's Law of Trade Marks and Trade Names (Sweet & Maxwell, 15th Ed, 2011) ("Kerly") at para 9-030:
The global appreciation assessments in relation to the relative ground in s 5 [of the UK Act] must be made on the basis of fair and notional use of the earlier and later marks, rather than by taking into account the actual use in the marketplace as required in the equivalent assessments in relation to infringement (under s 10). This may have the consequence that, for example, a tribunal finds that on the basis of notional and fair use there was a likelihood of confusion between the earlier and the later mark under s 5(2) and therefore the later mark was refused registration; but another tribunal later finds no infringement of the later mark under s 10(2) by reason of the circumstances of actual use of the registered mark and the defendant's mark that the original tribunal refused to register.
Whether extraneous factors are relevant to the confusion inquiry
63. A related issue which arose in the course of submissions is whether the court should, in assessing the likelihood of confusion, consider factors which are external to the similarity between the competing marks and the similarity between the competing goods or services. We have previously referred to such factors as "extraneous factors" (see Polo (CA) at  and Sarika at ) because they are extraneous or, perhaps, more accurately external to the marks and the services in question. Such factors have traditionally included steps taken by the parties to differentiate their goods and services ("differentiating steps") so as to reduce the likelihood of confusion. The question arose in this case because the Judge, at - of the GD, applied this court's approach to extraneous factors as laid down in Polo (CA) at  and on this basis placed weight on extraneous factors such as the distinct branding of the hotels as 4-star and 6-star respectively, and how the perceptions of customers making purchases would be affected by the trade channels used by the parties to market their hotel services.
64. We approach this question again from the starting point of the Act. Once similarity between the competing marks and goods or services has been established, the impact of these similarities on the relevant consumers' ability to understand where those goods and services originate from falls to be considered. The only relevant confusion is that which results from the similarity between marks and goods or services. Our courts have consistently recognised this since Lai Kew Chai J articulated it in Polo (HC) at . Equally however, the plain words of ss 8(2) and 27(2) do not have the effect of making a finding of confusion automatic upon the establishment of similarity of marks and goods or services. If that was what was intended, as observed in Polo (CA) at  and in Sarika at , Parliament would have provided that once such similarity is found the matter ends there without any need to examine whether there is a likelihood of confusion as a result. However, we reiterate that the statute requires that any likelihood of confusion must arise from the similarity of marks and goods or services.
65. If the marks or the goods or services in question are not similar, and confusion stems from other factors, an action might lie in the tort of passing- off but not in trade mark infringement. The similarity of marks and that of the goods or services are threshold questions but they are not determinative questions; in short these are necessary but not sufficient conditions. In our judgment, the plain words and the scheme of s 8(2) of the Act (as well as s 27(2) of the Act), as we have set it out above do not preclude the court's discretion to consider extraneous factors to the extent that these inform the assessment of the effect of the required similarity on consumer perception, but as we explain below, there are significant limits, more than we may previously have acknowledged as to the extraneous factors that may be considered.
66. We have hitherto considered extraneous factors in the confusion inquiry in the context of both infringement and opposition proceedings (see City Chain Stores (S) Pte Ltd v Louis Vuitton Malletier  1 SLR 382, Sarika and Valentino). In the course of the arguments in this case, we were referred to a line of English and European cases which ostensibly endorsed the exclusion of extraneous factors from the confusion inquiry. The first of these cases is Perfumery Ld v June Perfect Ld andF W Woolworth & Co Ld (1941) 58 RPC 147 ("June Perfect"), where Sir Wilfred Greene MR observed that "once a mark is shown to offend, the user of it cannot escape by showing that by something outside the actual mark itself he has distinguished his goods from those of the registered proprietor" (at 161). We explained in Hai Tong (at -) that this statement was made in the context of the marks-similarity stage of the inquiry and, in this regard, we have reiterated above that the inquiry is to be conducted mark-for-mark without added material. In Julius Samaan Ltd v Tetrosyl Ltd  FSR 42 ("Julius Samaan") however, the English High Court expressly excluded extraneous factors from the confusion inquiry, stating that the court must consider the likelihood of confusion arising from the defendant's use of the offending sign, discounting added matter or circumstances (at 872).
67. As against this, the European and more recent English cases appear less restrictive. The ECJ's approach is set out in Sabel, in terms which are now replicated to some degree throughout European and English jurisprudence (at 223-224, - of the court's Judgment) :
Article 4(1)(b) of the Directive does not apply where there is no likelihood of confusion on the part of the public. In that respect, it is clear from the tenth recital in the preamble to the Directive that the appreciation of the likelihood of confusion 'depends on numerous elements and, in particular, on the recognition of the trade mark on the market, of the association which can be made with the used or registered sign, of the degree of similarity between the trade mark and the sign and between the goods or services identified'. The likelihood of confusion must therefore be appreciated globally, taking into account all factors relevant to the circumstances of the case.
That global appreciation of the visual, aural or conceptual similarity of the marks in question, must be based on the overall impression given by the marks, bearing in mind, in particular, their distinctive and dominant components. The wording of Article 4(l)(b) of the Directive - '... there exists a likelihood of confusion on the part of the public ...' - shows that the perception of marks in the mind of the average consumer of the type of goods or services in question plays a decisive role in the global appreciation of the likelihood of confusion.
68. We pause to note that the ECJ's definition of "all factors relevant to the circumstances of the case" is not without limitation. The factors listed at  show that emphasis remains on the elements associated with the marks themselves, although other factors are not excluded from consideration. There is also express focus on the perception of the average consumer, in accordance with the wording of Art 4(1)(b) of the Directive (at ). A particular outworking of the focus on consumer perception is in the weighting of each element of mark-similarity in the course of the confusion inquiry. As explained in Lloyd (at 1352 and 1358):
... the national court must consider whether there is a genuine and properly substantiated likelihood of confusion for an average consumer of the particular type of goods in the Member State concerned. In making that assessment it must make a global assessment of all relevant factors, in particular the degree of aural, visual or conceptual similarity between the mark and the sign and the distinctiveness which the mark has either per se or by virtue of the degree of recognition of the mark.
... In order to assess the degree of similarity between the marks concerned [for the purpose of the global assessment], the national court must determine the degree of visual, aural or conceptual similarity between them and, where appropriate, evaluate the importance to be attached to those different elements, taking account of the category of goods or services in question and the circumstances in which they are marketed.
69. We agree that it is permissible at the confusion-stage of the inquiry to have regard to the importance to be attached to the different elements of similarity taking into account the type of the goods and services in question and how purchases of such goods and services are typically made. We reiterate here the observation made above at .
70. It may be that the more recent English cases have taken their cue from the ECJ and moved away from the strictly restrictive stance taken in Julius Samaan. In O2 (ECJ), the ECJ endorsed the English High Court's approach of analysing the likelihood of confusion within the "context in which the sign similar to the [plaintiff's] bubbles trade mark was used by [the defendant]" (O2 (ECJ) at ). Both the English High Court (in O2 Holdings Ltd (formerly 02 Ltd) v Hutchinson 3G  EWHC 534 (Ch) ("O2 (UKHC)) and the ECJ concluded that the advertisement in which the defendant applied its impugned sign "as a whole, was not misleading, and in particular, did not suggest that there was any form of commercial link between [the plaintiff and the defendant]" (O2 (ECJ) at , emphasis added). It is evident that the courts considered the use of the sign within the context of the advertisement "as a whole" in arriving in at its conclusion on confusion. O2 (ECJ) has been seen as reflecting a more "holistic" approach on the part of the ECJ towards assessing confusion by focussing on the element of use, at least in the infringement context (see Ilanah Simon Fhima, "Trade mark infringement in comparative advertising situations" 2008 EIPR 30(10) 420 at p 425).
71. Later cases have followed O2 (ECJ) and take into account the "actual context and circumstances of use of the sign itself' (Och-Ziff at -) and Red Bull at ). While "actual context" in those cases was stated as a "limitation" on the court's range of considerations (see Och-Ziff at  and Red Bull at ), the court quite clearly said this to draw out the distinction between the context and the use-specific inquiry in infringement proceedings, and the broader inquiry in opposition proceedings where notional uses and acontextual applications of the mark are considered. In Och-Ziff, the court's "contextual assessment" also included factors extraneous to the marks themselves but existing within the context of the parties' use of the marks (see ). The claimants there claimed inter alia that the defendant's use of a number of signs (OCH, OCH CAPITAL, ochcapital, Och Capital and OCH Capital) in respect of his financial advisory and fund management business, infringed their trade marks "OCH-ZIFF" and "OCH" which were registered in Class 36 (financial services). In elaborating on what the "contextual assessment" of likelihood of confusion entailed, Arnold J first cited - of O2 (ECJ) (see above at ) on the distinction between infringement and opposition proceedings. He then went on to ask (at -):
The question which arises is this: how far do the "context" referred to by the Court at  and the "circumstances characterising that use" referred to by the Court at  extend? Counsel for Och-Ziff submitted that the context and circumstances were limited to the actual context and circumstances of the use of the sign itself. Thus, in the O2 case itself, where the sign was used in a comparative advertisement, the context was the whole of the comparative advertisement, but no more. By contrast, counsel for the defendants submitted that the context and circumstances included all circumstances relevant to the effect of the use of the sign, including circumstances prior to, simultaneous with and subsequent to the use of the sign.
In my judgment the context and circumstances are limited to the actual context and circumstances of the use of the sign itself.
The Court of Justice explicitly said at  that the referring court was right to "limit its analysis" to the context in which the sign was used. Furthermore, it referred at  to the circumstances "characterising the use", not to the circumstances more generally. Thus circumstances prior to, simultaneous with and subsequent to the use of the sign may be relevant to a claim for passing off (or, under other legal systems, unfair competition), but they are not generally relevant to a claim for trademark infringement under art.
72. What emerges from this reasoning is that at least in infringement proceedings, the whole of the actual context of use will be relevant to the confusion inquiry. Neither party in Och-Ziff even attempted arguing that the inquiry could entail jettisoning factors other than the marks and goods themselves. The court then went on to apply principle to fact. One of the defendant's arguments was that its use of the impugned signs was not confusing because "OCH", "Och" and "och" were all meant to be read as an acronym, ie, "Oh-See-Aitch", while the registered mark was meant to be pronounced as "Ock-Ziff". The court rejected the argument based on the following considerations (at  and ):
. even on the basis of a contextual assessment of those uses,
there is nothing to alert the consumer who is familiar with OCH-ZIFF to the fact that the OCH element of the sign is intended to be read as O-C-H rather than as Och. By contrast, I accept that some consumers will read OCH in "OCH Capital" as O-C-H. I do not accept, however, that all consumers would read it that way. This is for three reasons. First, the human eye has a tendency to see what the brain expects it to see.
Thus I consider that some customers will tend to read "OCH Capital" as "Och Capital" because of their knowledge of Och- Ziff and Mr Och.
Secondly, there is nothing in the context of OCH Capital's use to make consumers think that OCH is an acronym. If, for example, the sign "OCH Capital" were used in conjunction with, say, the sign "Ocean Clearing House", then consumers would understand that OCH should be read as O-C-H because
it was an acronym. OCH Capital does not do this, because OCH is not an acronym. I consider that even consumers who noticed that OCH Capital's Chief Executive was called Ochocki would be unlikely to appreciate the derivation of the name OCH Capital without it being explained to them unless they were personally acquainted with Mr Ochocki.
73. In its application of the contextual assessment the court therefore considered external factors (or in that case, the lack thereof) which would have informed the consumer as to how the sign and mark were to be read. It is noted that these factors, though undoubtedly external to the marks, were closely related to and considered for the specific purpose of assessing how the marks would be perceived or read by the average customer.
74. The latest instalment in the EU/UK caselaw is Specsavers International Healthcare Ltd v Asda Stores Ltd  FSR 19 ("Specsavers (UKCA)") and Specsavers International Healthcare v Asda Stores Ltd, Case C-252/12 ("Specsavers (ECJ)"). The plaintiff there registered Community trade marks with the word "SPECSAVERS". The defendant ran an advertising campaign for its in-store opticians that made use of a logo similar to that of the plaintiff's graphic mark and of posters bearing the straplines "Be a real spec saver at Asda" and "Spec savings at Asda". The plaintiff claimed that among other things, the defendant's advertising campaign was likely to cause confusion under Art 9(1)(b) of Community Regulation 207/2009 ("the Regulation"), which corresponds to s 27(2) of our Act. The claim was dismissed in both the English High Court ( EWHC 2035 (Ch) ("Specsavers (UKHC)") and the Court of Appeal. In particular, it was found that the defendant's use of the term "spec saver" in its straplines, although similar to the plaintiff's "SPECSAVERS", was not likely to create origin- based-confusion. Mann J in the High Court stated that "context is all important" and that considering the reference to "Asda" and the play on the word "spec saver", the public would know that the advertisement was for the defendant and not the plaintiff (Specsavers (UKHC) at ). The plaintiff's contention that Mann J erred in considering the context of the advertisement and the extraneous word "Asda", was rejected by the Court of Appeal.
75. Delivering the judgment of the Court of Appeal, Kitchin LJ traced the English court's restrictive approach toward extraneous factors taken in cases such as June Perfect (though as we have observed, we consider this related to the mark-similarity stage of the inquiry) and Julius Samann, but then went on to note that "the [ECJ] was not constrained in the same way" and had "in a number of cases...indicated that the court must take account of the specific circumstances of the use of the offending sign" (at ). He then turned his focus to the ECJ's restatement of the "contextual" approach in O2 (ECJ), and what he saw to be the importation of this approach to England in Och-Ziff. His conclusion on the question of how the confusion inquiry ought to be conducted at least in an infringement setting, was as follows (at ):
In my judgment the general position is now clear. In assessing the likelihood of confusion arising from the use of a sign the court must consider the matter from the perspective of the average consumer of the goods or services in question and must take into account all the circumstances of that use that are likely to operate in that average consumer's mind in considering the sign and the impression it is likely to make on him. The sign is not to be considered stripped of its context.
76. The English Court of Appeal's decision confirmed that "in assessing infringement of a trade mark the full context of use must be considered, including the advertising campaign as a whole, not just a comparison of the mark to sign" (see casenote in EIPR 2012, 34(5), 354-357).
77. Therefore, the latest European and English cases suggest an emphasis on the "context" and this includes the milieu in which the mark is used. While we do not, with respect, think the position in England or Europe is quite as "clear" as Kitchin LJ suggested in Specsavers (UKCA), we also think it far from correct to say that the English and European cases now eschew the consideration of extraneous factors in the confusion inquiry altogether. In the final analysis, having taken note of developments elsewhere, we must of course come to a conclusion by reference to our statute and the way in which it has been understood by our courts. Our statutory regime after all has been applied over a number of years. In this regard, we are satisfied that there must be, and is, some room for the consideration of external factors. We begin with what we said in Hai Tong at [85(c)]:
... We earlier observed at [40(b)] above that in comparing the similarity between two contesting marks ... this is done without regard to "external added matter or circumstances" .
But, this is not the case when assessing the extent to which the goods bearing the contesting marks are similar and the likelihood of confusion among the relevant segment of the public. To prevent the assessment of these elements from becoming an exercise consisting largely of guesswork, it must be legitimate to have regard to extraneous factors that affect the degree and extent of the weight to be accorded to each of these elements and how they interplay with one another.
78. It cannot be gainsaid that the court's consideration of such factors in the confusion inquiry must be conducted in accordance with the structure of the Act and consistent with the legislative purpose of the trade mark registration regime. To this end the inclusion of extraneous factors must be managed so as to strike a balance between the competing goals of the trade mark registration regime. In Sarika we recognised (at ) that the competing policy concerns were those of preventing confusion on the one hand and promoting business certainty on the other. We concluded that the primary concern was to prevent confusion. Indeed, the plain words of ss 8(2) and 27(2) direct the court to ultimately assess the likelihood of inaccurate consumer perception as to the source of goods, ie, confusion. At the same time, we are concerned that the registration regime remains 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.
79. In light of these considerations, in Hai Tong at  we said:
... while it is permissible to have regard to "extraneous factors", this should not be taken as a reference to such factors "at large" and still less to those that are inconsistent with the legislative framework because they impermissibly curtail the rights granted to the registered proprietor of a trade mark. Rather, s 27(2) of the Act contemplates the interplay among:
(a) the degree of similarity between the contesting marks;
(b) the degree of similarity between the goods or services in relation to which these are to be used; and
(c) the effect of the foregoing on the relevant segment of the public, in order to assess the likelihood of confusion. Therefore, those factors that bear upon these elements in combination or upon the similarity of the goods or services in question and the characteristics, nature and likely responses of the relevant segment of the public can and should be considered. .
80. In this regard we were following the approach initially applied in Polo (HC). We reiterate these principles but with some significant refinements and previous case law on the admissibility of extraneous factors should be understood in the light of what we set out below. In our judgment, the following principles apply to the three stages of the inquiry under both ss 8(2)(b) and 27(2) of the Act but subject to the qualifications noted below.
81. On similarity of marks - the marks are to be considered without regard to any added or extraneous matter, as is well-established in case law (see  above).
82. On similarity of goods or services - extraneous factors may be relevant to establish the degree of similarity as between goods and services that are not identical in infringement cases and in opposition proceedings where the applicant's and proprietor's goods and services are registered or to be registered in different classes or specifications. Such extraneous factors are some of those identified in British Sugar, in particular the uses and the users of the goods and services in question, their inherent nature and the extent to which they are competitive. Extraneous factors are not to be considered if the goods and services are identical, because there will be no need to determine how similar they are (see  above). Goods and services will be regarded as identical where they are registered or to be registered in the same class and specification in opposition proceedings, while in the infringement context it will be permissible to have regard to the classification in which the allegedly infringing item or service would have been inserted had the alleged infringer sought registration of his mark.
83. On the effect of the foregoing on the relevant segment of the public - extraneous factors may be considered to the extent that they inform the court as to how the similarity of marks and goods will likely affect the consumer's perception as to the source of the goods. This however, is subject to some important qualifications which we will now elaborate upon.
84. First, in opposition proceedings, it will generally not be relevant to consider extraneous factors that relate to the actual and particular circumstances and ways in which the mark was used on the goods in question. While it will be necessary to consider the notional or fair uses to which each of the marks could be put, for instance in terms of what types of goods or services are within the contemplated uses for which the mark has been registered, it will not be relevant to have regard to the particular way in which the goods or services have been affixed with the mark and are then being marketed. This would thus exclude consideration of such factors as differences in the intended market segments, trading strategies employed, websites used or the trader's chosen limitations as to his use of the mark. In opposition proceedings, it is the overlap between the notional fair uses of the registered mark on the one hand, and of the applicant mark on the other, that is in issue. It would denude of significance the critical distinction that we have drawn between infringement and opposition proceedings if the confusion inquiry were diverted into a consideration rooted in the details of the actual circumstances in which the goods or services affixed with the mark are being marketed.
85. Second, even in infringement proceedings, there is a need to be vigilant to the very purpose for which the step-by-step approach was devised and has been applied. To the extent extraneous factors are advanced in the attempt to negate a finding of likely confusion on the grounds that the marks are being deployed in such a way as to target different types of consumers or to notionally differentiate the goods in question, this must be avoided. Once the goods are found to be similar or identical at the second stage of this inquiry, it becomes irrelevant to consider such factors again at the confusion- inquiry stage.
86. To put it another way, if the competing marks and goods are found to be sufficiently similar such that, having regard to the characteristics of the likely consumer, the court concludes there would be a likelihood of confusion, then it would not be permissible to have regard to yet further extraneous considerations that might have the effect of diminishing the likelihood of confusion. In the same way that confusion stemming from sources other than the similarity of marks and goods is outside the reach of trade mark protection (see  above) so too must such factors be irrelevant to displace a finding of likely confusion if the property rights represented in the trade mark are to be meaningfully upheld.
87. Extraneous factors have been adduced in general in order to show:
(a) that even if on a mark-for-mark comparison the marks are similar, then having regard to extraneous material they are not confusingly so; or
(b) in relation to the goods, even if they are similar or for that matter identical, then having regard to extraneous material they are not confusingly so.
88. In our judgment, this seems to be wrong in principle. In relation to the analysis of the marks, it denudes of any force the principle that the marks- similarity inquiry is to be conducted without recourse to external material, if such material were then permitted at the confusion stage of the inquiry.
89. For the same reason, while external factors of the sort we have identified at  above may be considered at the goods-similarity stage of the inquiry where the goods are not identical (that is to say they do not or would not fall within the same specification on the register), once the determination is made that they are similar, then it would be wrong in principle at the confusion stage of the inquiry to examine whether the trader by means of his superficial trading choices has differentiated his goods in some way from those of the owner of the incumbent mark even though in substance the goods are similar or even identical for the purposes of trade mark law.
90. This seems to us to be sound in principle if adequate regard is to be had to the proprietary rights of the owner of a trade mark. Otherwise a subsequent trader would be able to enter the market using a trade mark that was very similar to the senior mark, applied to similar if not identical goods and yet avoid liability by means either of an express disclaimer, or by using cheaper materials and selling his goods at a much lower price and saying that because of these steps, there is no likelihood of confusion notwithstanding the high degree of similarity or even of identity having been found at the antecedent stages of the inquiry.
91. Such a notion was rejected in Rolex Internet Auction  ETMR 25 where the defendant hosted a website on which imitation or replica watches were auctioned. The court observed that:
Risk of confusion ... is not excluded by the fact that the goods offered are designated "replicas" or "imitations". This is because ... the basic consideration does not concern a specific selling situation in which a risk of confusion in itself existing can be cleared up by means of explanatory notes or in some other way, for example low price, but relates to the abstract risk of confusing the two identifying marks.
92. In our judgment, this is sensible for the reason put forward by Prof Ng- Loy in her chapter "Developments in Singapore Trade Mark Law 2005-2011:
Confusion-based Protection and Beyond" in Developments in Singapore Law between 2005 and 2011: Trends and Perspectives (Academy Publishing pp 349-373) at p 357, where she wrote that the admission of extraneous factors:
. renders the protection promised to trade mark proprietors by the registration system an illusory one . permits a trader whose mark is similar to the senior mark to avoid a finding of confusion by pricing his products or services very differently, packaging them very differently, and even by using disclaimer notices ("My goods are not XYZ brand"). This [admission of extraneous factors rule] places a weapon in the hands of the junior mark owner to circumvent infringement. In this way, the rule can be unfair to the owner of the senior mark. They would also argue that the Legislature could not have intended to allow reference to matters outside of the parties' marks and their goods or services. The language used in the Trade Marks Act is this: because similar marks are used in relation to identical or similar goods or services, there exists a likelihood of confusion on the part of the public.
93. What does this leave in terms of factors that affect how the consumer would perceive the marks and how this would bear on the likelihood of confusion? In Thomas R Lee, Eric DeRosia and Glenn L Christensen "Trademarks, Consumer Psychology and the Sophisticated Customer" 57 Emory LJ 575, the authors propose a framework for understanding how consumers are likely to behave in devoting attention to the prospective purchase so as to shed light on the central question in trade mark protection, namely the likelihood of confusion. Perhaps at the risk of oversimplifying their hypothesis, the authors suggest that the key considerations that inform the analysis should be:
(a) Factors that motivate the consumer to exercise care in the purchase in question directed at, among other things, the forming of a judgment over the source of the product (see at pp 589-594); and
(b) Factors that enable the consumer to exercise such care (see at pp 595-601).
94. Examples of factors that implicate either or both of these considerations include such things as the price of the item (ie, whether it is expensive or inexpensive); whether the item is of the sort that is usually purchased on impulse with fleeting attention or something that is very personal and so tends to attract greater interest and attention on the part of the consumer; the nature of the typical purchasing process for such items and whether it is generally attended with sales personnel, discussion or negotiation; whether the transactions are routine or infrequent, and so on. In our judgment, this provides a meaningful and potentially useful framework for analysing the question by focusing on the inherent characteristics and nature of the goods, the marks and how the purchaser is likely to approach the purchasing decision.
95. Although the risk of origin-based confusion is the primary interest sought to be protected by trade mark law, there must be a limit to the range of external factors that may be taken into account to determine whether a sufficient likelihood of such confusion exists. The permissible factors are those which (1) are intrinsic to the very nature of the goods and/or (2) affect the impact that the similarity of marks and goods has on the consumer. The impermissible factors are those differences between the competing marks and goods which are created by a trader's differentiating steps. In other words, factors which are not inherent in the goods, but are susceptible to changes that can be made by a trader from time to time, should not be permissible considerations. In particular, we are satisfied that it is unnecessary, unworkable and impermissible for the court to have regard to such issues as pricing differentials, packaging and other superficial marketing choices which could possibly be made by the trader. In contrast, extraneous factors that relate to the purchasing practices and degree of care paid by the consumer when acquiring goods of the sort in question, can be considered and assessed without descending into the details of particular differentiating steps which the trader might choose to take in relation to the goods and services falling within the specification.
96. Based on these considerations, the following represents a non- exhaustive list of factors which we regard as admissible in the confusion inquiry:
(a) Factors relating to the impact of marks-similarity on consumer perception: the degree of similarity of the marks themselves (see Hai Tong at [85(c)(iii)], the reputation of the marks (see Polo (CA) at ), the impression given by the marks (see Polo (CA) at ), and the possibility of imperfect recollection of the marks would all be relevant. Clearly, the greater the similarity between the marks, the greater the likelihood of confusion. As to the reputation of the mark, Mobil Petroleum Co, Inc v Hyundai Mobis  1 SLR 512 ("MobiF) at  makes it clear that a strong reputation does not necessarily equate to a higher likelihood of confusion, and could in fact have the contrary effect as in McDonald's Corp v Future Enterprises Pte Ltd  1 SLR(R) 177 (see at ).
(b) Factors relating to the impact of goods-similarity on consumer perception: it would be legitimate to take into account factors concerning the very nature of the goods without implicating any steps that are taken by the trader to differentiate the goods. This includes the normal way in or the circumstances under which consumers would purchase goods of that type (see  above, Mystery Drinks at , Lloyd Schuhfabric Meyer v Klijsen Handel BV, Case C-342/97  2 CMLR 1343 at 1352; and Philips-Van Heusen Corp v OHIM  ETMR 60 at ). This factor is not directly dependent on the marketing choices that the trader makes. As alluded to at  above, it would also be relevant to have regard to whether the products are expensive or inexpensive items, 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 (see generally Hai Tong at [85(c)(i)]), and 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. We refer here to In the matter of an Application by the Pianotist Company for the Registration of a Trade Mark (1906) 23 RPC 774 ("Pianotist") where it was observed that, having regard to the nature of the article in question (musical instruments), the (high) price at which it was likely to be sold, and the nature of the consumers who are likely to purchase such products ("generally persons of some education"), a man of ordinary intelligence was unlikely to be confused (at 778). The price of the type of product being sold is distinct from the issue of price disparity between the parties' products. The former consideration directly impinges on the degree of care the consumer is likely to pay to his purchase and therefore his ability to detect subtle differences. As observed in Reed Executive Plc v Reed Business Information Ltd  RPC 12 at , "a 50 pence purchase in the station kiosk will involve different considerations from a once-in-a-lifetime expenditure of 50,000 pounds". On the other hand, superficial price disparity between the competing goods, which speak more about the trader's marketing choices rather than differences in the nature of the goods themselves, is not a factor we find relevant to the inquiry.
97. We turn to consider the facts in the light of these principles.
Application to the facts
98. The Judge found that there was no likelihood of confusion arising from the similarity between the marks and the parties' services. The Judge's analysis depended to a significant degree on the differentiation between Staywell's hotel as a 4-star hotel and the Opponents' as a 6-star one (at  and  of the GD). At the outset, we reiterate that because these are not infringement but opposition proceedings (see  above) a focus on the actual and particular circumstances and the way in which the mark was used on the services in question would be misplaced for the purpose of determining whether the registration should be permitted to proceed. Even if, as the Judge found, there is no likelihood of confusion based on Staywell's actual use of the Applicant Mark in relation to a 4-star hotel, we must go further to assess whether the answer would be different having regard to the range of notional fair uses of each of the marks. What, for instance, if one party should exploit its right to use its mark in a segment in which the other operates (for instance, if Staywell applied the Applicant Mark in the luxury hotel market, or if the Opponents applied the ST. REGIS mark in the business hotel market)?
99. But there is a further point. Confusion in trade mark law extends to consumers believing that the users of the competing marks are economically linked (Subway Niche at  and Sarika at ). Given this, we do not agree that it was correct in the circumstances to place such a high degree of significance on the differentiated marketing of the hotels as 4-star and 6-star respectively.
100. In our judgment, in respect of the services, the Park Regis and ST. REGIS Singapore are not qualitatively different products. Indeed, if the notional fair uses of the marks are considered, hotel services are hotel services and there is little, if any, room for differentiation between them.
101. As to the similarity of the marks, we have already observed that there is no visual similarity, some degree of conceptual similarity, and a significant degree of aural similarity. On the whole, we have found that the marks were similar. We also agree that "Regis" is the dominant component of the mark on an aural analysis.
102. Turning to the likelihood of confusion, the Judge considered a number of factors which might affect the ability of the customer to accurately discern the economic relationship (or lack thereof) between the parties' hotels. These in the main revolved around differences in their marketing methods and channels, but for the reasons we have set out above, we do not think these were legitimate considerations. Moreover, as we have also noted, the range of notional fair uses would easily extend across the spectrum of 4- to 6-star hotel segments. But even confining our analysis to only the present, actual use of the marks on these particular types of hotels, we find it not unlikely that the public would be induced by the similarity of the marks and the common industry in which they are used into believing that there is at least some economic link between the ST. REGIS Singapore and the Park Regis Singapore. This is especially so given that it is common for large hotel chains to operate differently branded hotels carrying different logos, united only by use of a common denominator in their names. This signals to the public that the various hotels, though pitched at different segments of the market, are economically linked. The High Court noted this in CDL Hotels at  in relation to the use of the common denominator "Millennium" in respect of the defendant's chain of hotels. The common denominator serves as an assurance of source and therefore quality. Examples are Hilton E 53 Hilton H0,IIS"'IS0"S Isaa] Uaraen Inn H^A^I GRAND | HYATT PARK HYATT REGENCY 1
103. In our view, therefore, the similarity arising from the use of the common denominator "Regis" in both marks takes on particular significance and there is ample ground for finding that there is a sufficient likelihood of confusion as to the existence of an economic link between the two hotels in question.
104. Before we conclude this part of our judgment, we touch on one remaining issue.
Initial interest confusion
105. The Opponents also argued that confusion which arises initially but which would have been dispelled by the time of the purchase can constitute confusion for the purpose of s 8(2) of the Act. This is an application of the doctrine known as "initial interest confusion", derived from American trade mark law. The Opponents relied on a number of cases from the UK and EU, as well as Sarika (at ) and Novelty Pte Ltd v Amanresorts Ltd and another  3 SLR(R) 216 ("Amanresorts") (at ), arguing that initial interest confusion is already part of our law. We wish to state at the outset that the doctrine of initial interest confusion is not currently part of our law. The question before us is whether it should be.
106. We have noted the origins and the evolution of initial interest confusion (see Michael J Allen, "Who Must Be Confused and When? The Scope of Confusion Actionable Under Federal Trademark Law" (Vol 81 TMR 209)). The concept can be traced to a 1962 amendment to the American Federal trade mark legislation, the Lanham (Trademark) Act (15 USC) ("the Lanham Act") (see Dinwoodie and Janis, Trademarks and Unfair Competition: Law and Policy (Wolters Kluwer 2010, 3rd Ed) at p 563). Prior to the amendment, protection was granted (in the context of similar marks) only where use was "likely to cause confusion or mistake to deceive purchasers as to the source or origin of such goods or services" (emphasis added). This was construed as requiring proof of the likelihood of confusion of the purchaser at the point of purchase. Following the amendment, the reference to "purchasers" was deleted. The legislative history of the 1962 Amendment indicates that the purpose of the deletion was to "clarify" the provision "since the provision actually relates to potential purchasers as well as actual purchasers" (emphasis added) (S Rep No 2107, HR Rep No 1108). This led some courts to expand the application of the Lanham Act beyond actual purchasers to recognise pre-sale (initial interest) and post-sale confusion.
107. The concept of initial interest confusion has since been applied both broadly and restrictively. Under the broad interpretation, the concern is not with the risk that the consumer will eventually purchase the defendant's goods; it is sufficient to show that the consumer was drawn to the defendant's goods based on its ostensible association with the plaintiff's mark. The broad interpretation was applied by the Second Circuit Court of Appeal in Grotrian, Helfferich, Schulz, Th. Steinweg Nachf v Steinway & Sons 365 F Supp 707 (1973) ("Grotrian") concerning pianos sold under the names "Grotrian- Steinweg" and "Steinway & Sons". This was one of the first cases in which the concept of initial interest confusion was applied. There the court stated (at 1342):
We decline to hold. that actual or potential confusion at the time of purchase necessarily must be demonstrated to establish trademark infringement under the circumstances of this case.
The issue here is not the possibility that a purchaser would buy a Grotrian-Steinweg thinking it was actually a Steinway or that Grotrian had some connection with Steinway and Sons. The harm to Steinway, rather, is the likelihood that a consumer, hearing the "Grotrian-Steinweg" name and thinking it had some connection with "Steinway," would consider it on that basis. The "Grotrian-Steinweg" name therefore would attract potential customers based on the reputation built up by Steinway in this country for many years.
108. On the narrower interpretation, initial interest confusion is applicable only if it gives rise to a real risk that the potential consumer will actually purchase the defendant's goods even upon knowing that they are not the plaintiff's goods. This is illustrated in Brookfields Communications Inc v West Coast Entertainment Corporation 174 F 3 rd 1036 (9th Cir, 1999) ("Brookfields"), which applied the narrower interpretation. The case concerned the defendant's use of the plaintiff's registered trade mark, "Moviebuff" as a metatag for its own website. The effect was that when an internet user searched for the term "Moviebuff", the search results would include the defendant's website. The defendant was held liable for trade mark infringement. The Ninth Circuit Court of Appeals gave its reasons, stating (at ) that:
Using another's trademark in one's metatags is much like posting a sign with another's trademark in front of one's store. Suppose West Coast's competitor (let's call it "Blockbuster") puts up a billboard on a highway reading - "West Coast Video:
2 miles ahead at Exit 7" - where West Coast is really located at Exit 8 but Blockbuster is located at Exit 7. Customers looking for West Coast's store will pull off at Exit 7 and drive around looking for it. Unable to locate West Coast, but seeing the Blockbuster store right by the highway entrance, they may simply rent there. Even consumers who prefer West Coast may find it not worth the trouble to continue searching for West Coast since there is a Blockbuster right there. Customers are not confused in the narrow sense: they are fully aware that they are purchasing from Blockbuster and they have no reason to believe that Blockbuster is related to, or in any way sponsored by, West Coast. Nevertheless, the fact that there is only initial consumer confusion does not alter the fact that Blockbuster would be misappropriating West Coast a.cquired goodwill.
109. The Opponents cited this passage in support of their argument seeking a finding of initial interest confusion in the present case. It is noted that Brookfields, and the advent of the internet, were what gave prominence to the doctrine of initial interest in the US.
110. Under both the broad and narrow interpretations, the focus of the doctrine is on protecting the reputation and the goodwill of the prior, registered mark. The mischief in both instances lies in misappropriating or freeloading off the goodwill or attractive force of others. In both Grotrian and Brookfields the concern was not the risk of confusion as to the trade source of the defendant's good, but rather the fact that the defendant in the two cases respectively "attract[ted] potential customers based on the reputation built up by [the plaintiff]", and "misappropriate[ed] [the plaintiff's] acquired goodwill...". This is language that reflects a concern with preventing the usurpation and dilution of reputation and goodwill. Later cases have echoed this theme. In Checkpoint Systems Inc v Check Point Software Technologies Inc 269 F 3d 270 (3d Cir 2001) (at 295), it was said that initial interest confusion was to prevent the defendant from "receiving a 'free ride on the goodwill' of the established mark". In Dorr-Oliver Inc v Fluid-Quip Inc 94 F 3d 376 (7th Cir 1996) (at 382) it was used to prevent the defendant from using a "bait and switch" tactic to "get its foot in the door by confusing consumers".
111. The Opponents also pointed us to the adoption of the initial interest doctrine in the UK and Europe. The leading case is Och-Ziff Arnold J there cited Grotrian and Brookfields, before going on to cite two English cases and three ECJ cases in support of the proposition that initial interest confusion was actionable under Art 9(1)(b) of the European Union Trade Marks Directive (Council Regulation 207/2009) ("EU Directive") (equivalent to s 27(2)(b) of the Act). The English cases are BP Amoco plc v John Kelly Ltd  FSR 5 ("BP") at  and Whirlpool Corp v Kenwood Ltd  EWHC 1940(Ch) ("Whirlpool') at . The dicta in these cases suggest that the court was concerned that a buyer could be misled by the similarity of the marks to initially believe that the defendant's goods were the plaintiff's, and then by the time he had come to realise his misconception, he would already have progressed so far down the route leading to purchase that he would not turn back to purchase the claimant's goods. The claimant would therefore stand to lose his custom owing to the initial confusion (see BP at  and Whirlpool at ).
112. In the ECJ cases, namely O2 (ECJ), Die BergSpechte Outdoor Reisen under Alpinschule Edi Koblmuller GmbH v Guni, Case C-278/08  ETMR 33 and Portakabin Ltd v Primakabin BV, Case C-558/08  ETMR 52 ("Portakabin"), the court's concerns were somewhat different, being that pre-purchase factors like advertising could confuse the public into thinking that the plaintiff and defendant were economically-linked undertakings, and that this would have an "adverse effect on the function [of a trade mark] of indicating origin" (Portakabin at ). Indeed, the European cases do not specifically touch on the issue of pre-sale confusion that is dispelled by the time of purchase. Arnold J's conclusion from the European cases was that it was sufficient to find a likelihood of confusion as to trade source at the point when a customer viewed an advertisement, regardless of whether or not the advertisement led to a sale or, if it did, whether or not the consumer remained confused at the time of such sale (see Och-Ziff at  and ). Ultimately, his concern, as in the American cases, was with protecting the trade mark proprietor against damage to his goodwill by dilution, even if no diversion of sales resulted. These concerns are evident in his conclusion on the issue of initial interest confusion (at ):
... confusion arising from an advertisement is capable of causing damage to the trade mark proprietor even if such confusion would be dispelled prior to any purchase. Although there will be no diversion of sales in such circumstances, there are at least two other ways in which the trade mark proprietor may be damaged. The first is that a confusing advertisement may affect the reputation of the trade marked goods or services. It is irrelevant for this purpose whether the defendant's goods or services are objectively inferior to those of the trade mark proprietor. The second is that such confusion may erode the distinctiveness of the trade mark.
113. Having considered the relevant American, English and European authorities on the matter, our view is that the doctrine of initial interest confusion is directed at a different purpose than that of s 8(2) (and s 27(2)) of our Act. The rationale underlying the doctrine is very much the protection of the reputation of a well-known mark from dilution or the prevention of misappropriation of the owner's goodwill. But this court in Mobil (at ) and Amanresorts (at ) made clear that protection against dilution is the sole province of s 8(4)(b)(ii) of the Act, which was added to our Act for that specific purpose. The courts have repeatedly stated that the confusion element in s 8(2) is concerned with the origin and source of goods, and not simply their reputation or associative properties (see Hai Tong at , City Chain at  and Richemont International SA v Goldlion Enterprise (Singapore) Pte Ltd  1 SLR(R) 401 at  ("Richemonf) at ). If a consumer is initially confused but this is unlikely to persist to the point of purchase because of a lack of sufficient similarity in the marks or the goods then the purpose of the trade mark as a "badge of origin" has not been undermined. Moreover, the doctrine will often or even generally be incompatible with the threshold requirements of marks- and goods-similarity. Any dilution of the trade mark due to initial confusion is not actionable under s 8(2) of our Act, which is directed at protecting trade mark proprietors from origin-based confusion.
114. Besides inconsistency with our statutory scheme and objectives, there are policy concerns over extending the protection of these provisions in the way suggested by the Opponents. First among these is that it can stifle competition without corresponding benefit to consumer choice or access to information. The fact that an incumbent suffers business losses because of an increase in market competition, when there is no marks-similarity and hence no deception and consumers are not ultimately confused, is not the kind of loss which the law seeks to prevent (see Jennifer Rothman, "Initial Interest Confusion: Standing At the Crossroads of Trademark Law" 27 (2005) Cardozo Law Review 105 ("Rothman's article") at pp 108 and 163). A finding of liability when consumers are, in fact, accurately informed as to what they are purchasing, is at odds with the goal of trade mark law which is to protect the source-identifying function of a trade mark and the resulting benefits to consumer welfare (see Rothman's article at pp 129-130).
115. Furthermore, it also appears that the doctrine is not easily workable in practice and can introduce uncertainty. In a report issued by the International Trademark Association ("INTA"), it was concluded that "a review of initial interest confusion cases in the US reveals that in application the doctrine is hopelessly confused, inconsistent and sometimes incoherent" (INTA Report on the Online Use Subcommittee Concerning the Proposed Resolution on Initial Interest Confusion at p 3). In particular, even the US courts, where the doctrine originated, have yet to develop a consistent approach as to what factors or consumer behaviour would be sufficient to constitute initial interest confusion. Considering these objections, we are not convinced that the benefits of adopting this concept will outweigh the costs.
116. We therefore conclude that the doctrine of initial interest confusion should not be introduced into our law because it is inconsistent with the purpose of s 8(2) of the Act which is only to protect the trade mark as an indication of origin.
Conclusion on opposition under s 8(2)
117. Notwithstanding our rejection of the doctrine of initial interest confusion, we reverse the Judge's dismissal of the opposition under s 8(2)(b) of the Act. We affirm her finding that the marks were, on the whole, similar, and that the services were in essence identical; but we disagree that this similarity did not give rise to a likelihood of confusion at the point of sale of the parties' respective services. On the contrary, we are satisfied in the circumstances that there was a sufficient likelihood of confusion by reason of the similarity in the competing marks and the services. In our judgment, because of this similarity, the average customer would have likely been confused as to whether the owners of the competing marks were economically linked and it was therefore an impermissible infraction of the Opponents' rights to permit the registration to proceed.
118. Although it is sufficient for us to dispose of the appeal on this ground alone, we go on to consider the grounds of opposition brought under s 8(4)(b)(i) and s 8(7)(a) of the Act.
Whether the Judge erred in disallowing the opposition under s 8(4)(6)(i) of the Act
Would the use of the Applicant Mark in relation to Class 35 and 42 services be likely to damage the Opponents' interests?
119. The four elements which need to be shown for an opposition under s 8(4)(b)(i) of the Act are (1) the marks are identical or similar; (2) the earlier mark is well known in Singapore; (3) the use of the applicant mark would indicate a connection between the applicant's goods/services and the incumbent proprietor; and (4) the connection is likely to damage the interests of the proprietor.
120. We have already found the marks-similarity requirement to be satisfied (see  above). On the second element, Staywell has not challenged the PAR's and the Judge's findings that the ST. REGIS mark was well-known in Singapore at the date of the Application. As for the third element, the detailed analysis in Amanresorts has put it beyond doubt that the connection requirement of s 8(4)(b)(i) and s 55(3)(a) of the Act will be satisfied where there is a likelihood of confusion (see Amanresorts at  and ). In Mobil this court elaborated that a "connection" under s 8(3) of the Act refers to a connection as to origin, a connection as to quality, and business connection. In our view, these types of connections are relevant to 8(4)(b)(i) of the Act as well. Connection as to source and quality (insofar as the public expects goods or services which it thinks emanate from the same source to be of comparable quality (see Mobil at -)) imports classic origin-based confusion, while the business connection imports the misapprehension of an economic relationship between the applicant's products and the incumbent proprietor (see Mobil at -). In the present case, we have no difficulty in finding that the use of the Applicant Mark in Classes 35 and 43, would give rise to a confusing connection between Staywell's services and the Opponents'. This flows from our earlier finding that the use of the Applicant Mark is likely to give rise to confusion under s 8(2) of the Act, in particular, confusion that the parties' hotels are part of the same chain or are otherwise economically linked.
121. The real controversy then is the last element of s 8(4)(b)(i), namely, whether the use of the Applicant Mark is likely to damage the interests of the Opponents. The Judge found that even if the public thought that the parties' hotels were economically linked, there was no likelihood of damage to the Opponents' interests because the consumer would understand that the two hotels were pitched at different market segments. Therefore the fact that the Park Regis did not offer luxury services "would not cause damage to the ST. REGIS brand" (at  of the GD). The Opponents argued that the Judge had erred in so finding. They pointed to two heads of likely damage: first, there might be damage by "dilution" and "tarnishment" due to the public believing that the Opponents did not supply only top-of-the-range hotels but hotels of varying qualities; second, there would be a restriction on the Opponents' ability to exploit their mark in a lower market segment.
122. The Opponents rely on CDL Hotels International Ltd v Pontiac Marina Pte Ltd  1 SLR(R) 975 ("CDL Hotels") for their first head of damage. CDL Hotels was a passing off case, not one concerning trade marks. The damage recognised there was the "dilution of the respondents' goodwill in their business" because "[t]he public would not perceive the plaintiff as supplying solely top-of-the-range goods or services but of varying qualities... [and the] prestigious image of the plaintiff would in these circumstances be likely to be undermined" (at ). This court in Amanresorts clarified that the damage referred to in CDL Hotels was in fact nothing more than the conventional head of damage of erosion or tarnishment to goodwill caused by the public erroneously associating the plaintiff's 5-star hotel with the defendant's 4-star hotel (Amanresorts at ). Framed in this way, it is in fact just a species of traditional origin-based confusion. But it is evident from the architecture of s 8(4)(b)(i) that the relevant damage to the interests of the proprietor of the incumbent mark (see the fourth element at  above) must be something other than the perception of a confusing connection between the parties since this would already be subsumed within the third element (see  above). There was no submission of any other damage beyond the perception of a confusing connection in this regard.
123. Moreover, on the facts before us, we do not think that such a perceived connection could in itself be damaging to the Opponents' interest in this case. As noted above (at ) the principal source of confusion in this case was the likelihood of the relevant segment of the public thinking that the owners of the two marks were economically related or linked. We have referred to examples of hotel groups that are linked through the use of one or more common features of their individual marks. The primary point of such a marketing strategy is usually to enable a group to operate in different market segments. If this is the sort of confusion that might be engendered in this case, then we do not see how such a connection would have damaged the Opponents' interests since their primary argument is that it was within the ambit of the notional fair use of their mark to use it for a lower market segment.
124. Indeed this leads us directly to the second head of damage. The Opponents contended that there was likelihood of damage to their interests as it would limit or preclude their ability to expand into the 4-star hotel market in Singapore. This was a recognised head of damage in Amanresorts in relation to the passing-off claim. Staywell's response was that the Opponents displayed no such intention to expand, and that Amanresorts was distinguishable on the basis that the respondent there had already commenced expansion of its business into the field where the appellant operated. The Judge agreed with Staywell, stating that there was no basis for finding that the Opponents' as yet non-existent expansion plans were likely to be prejudiced (at ).
125. We are unable to agree with the learned Judge on this. Where the field in which the defendant or applicant operates is in close proximity to, or is a natural extension of, the incumbent's business damage in the form of a restriction of business expansion opportunities will more readily be inferred. Christopher Wadlow in The Law of Passing-off (Sweet & Maxwell, 2011) ( "The Law of Passing-off') at para 4-043 states that:
If the defendant's chosen field of business is a natural extension of that in which the claimant trades then, as a practical matter, damage is likely to be inferred even if the claimant has no present intention of expanding into that field.
126. This court held as much in Amanresorts, where it found that the respondent's field of business (viz, high end resorts and hotels) was "closely connected" to the appellant's residential accommodation business. The fact that the respondents had in fact already expanded into the residential accommodation business overseas only "buttress[ed]" the finding of the closeness of the two fields of business (at ). This was followed in Mobil, where we stated in relation to an action under s 8(3) of the Act that this head of damage was "premised on the close proximity between the present field of business and the prospective field to be expanded into" (at ). Only if there was no such proximity (as was the case in Mobil) would the incumbent have to display a genuine intention to expand into the prospective market. This makes sense because of the need to prevent unwarranted extension of the incumbent's protection based on the wholly speculative possibility of its future expansion into a market unrelated to its present business.
127. In the case before us, it can hardly be said that the markets for hotels of different star-ratings are not in close proximity to one another. It is perfectly plausible that an operator in the 6-star market might wish to set up a 4-star hotel within the same chain, or conversely that a 4-star hotel operator might eventually move into the 6-star market by setting up a separate hotel, or simply by improving the standards of its existing operations. Because of the close proximity of the businesses, if they can even be considered separate at all, we do not think there is a need for the Opponents to evince actual expansion plans. We are satisfied that the proximity of the markets in which the Opponents and Staywell operate give rise to a likelihood that the Opponents' business expansion opportunities will be prejudiced. Therefore, we find that damage under this head is made out.
Conclusion on opposition under s 8(4)(b)(i) of the Act
128. For the above reasons, we reverse the Judge's dismissal of the opposition under s 8(4)(b)(i) of the Act. Based on our finding as to the likelihood of confusion under s 8(2)(b) of the Act, we find that the use of the Applicant Mark in Class 35 and 43 would indicate a confusing connection between Staywell's services and the Opponents'. Moreover, we find that the Opponents' interest in the possibility of expansion into other market segments is likely to be damaged as a result. The opposition under s 8(4)(b)(i) therefore succeeds.
Whether the Judge erred in disallowing the opposition under s 8(7)(b) of the Act
129. Under s 8(7)(a) of the Act, a mark may be refused registration because its use would result in liability for the tort of passing-off. The Opponents argue that the use of the Applicant Mark would have this effect.
130. It is not disputed that to succeed in an action for passing-off, the classic trinity of goodwill, misrepresentation and damage must be proven. The dispute in this case centres upon the first element of goodwill. It is well-established that such goodwill must exist in Singapore at the date of the application for the registration of the junior mark. The Judge found that no such goodwill existed in the ST. REGIS Singapore on 3 March 2008, when the Application was filed ("the relevant date"). The Opponents' primary obstacle was the fact that at the date of the Application the ST. REGIS Singapore was not yet open for business. The Opponents submit that goodwill nonetheless subsisted in the ST.
REGIS Singapore at the date of the Application owing to (1) the conduct of extensive pre-trading activities prior to the relevant date and (2) the presence of Singapore customers of overseas ST. REGIS hotels, coupled with the reputation which the ST. REGIS brand already enjoyed in Singapore at the relevant date. With respect to the second argument, the Opponents were essentially inviting this court to depart from the so-called 'hard-line' approach to goodwill currently adopted in Singapore, and to move toward a more ' soft- line approach'. We analyse these arguments in turn.
The 'hard-line' and 'soft-line' approaches to goodwill
131. Goodwill was famously described by Lord MacNaughten in IRC v Muller & Co's Margarine  AC 217 at 223-224 as:
... the benefit and advantage of the good name, reputation, and connection of a business. It is the attractive force which brings in custom.. The goodwill of a business must emanate from a particular centre or source. However widely extended or diffused its influence may be, goodwill is worth nothing unless it has power of attraction sufficient to bring customers home to the source from which it emanates. I think that if there is one attribute common to all cases of good will it is the attribute of locality. For goodwill has no independent existence. It cannot subsist by itself. It must be attached to a business. Destroy the business, and the goodwill perishes with it, though elements remain which may perhaps be gathered up and be revived again.
132. The understanding that goodwill cannot exist apart from an associated business, has given rise to two opposing schools of thought. The traditional position is that for goodwill to exist, it is essential for the trader to have carried on his trade within the jurisdiction (see Anheuser-Busch Inc v Budejovicky Budvar NP  FSR 413 ("Anheuser-Busch') at 468, and Alain Bernadin et Compagnie v Pavilion Properties Limited  FSR 341 ("Crazy Horse ") at 345). This has become known as the 'hard-line' school of thought. Crazy Horse is the locus classicus of this line of cases. It stands for the proposition that a foreign trader which conducts no business activity in the jurisdiction can have no goodwill in the jurisdiction, notwithstanding that (1) it has advertised in the jurisdiction and/or (2) it has acquired a reputation in the jurisdiction and has customers in the jurisdiction who travel abroad to purchase the trader's goods or services.
133. More recently, a line of cases have taken a 'softer' approach to the territorial aspect of goodwill. The 'soft-line' approach states that goodwill can be established by something less than trade conducted in the jurisdiction, particularly if the trader has an international reputation and is able to draw customers to the source of its attractive force. Cases from the courts in Canada, New Zealand and Australia have taken this approach (see Orkin Exterminating Co Inc v Pestco Co of Canada Ltd et al 50 OR (2d) 726 ("Orkin"), Dominion Rent A Car Ltd v Budget Rent A Car Systems  2 NZLR 395 ("Dominion") and Conagra v McCain Foods (1992) 23 IPR 193 ("Conagra")).
134. In Singapore, the Privy Council adopted the 'hard line' approach in Star Industrial Co Ltd v Yap Kwee Kor [1974-1976] SLR(R) 581 ("Star Industrial"). There, Lord Diplock expressed the view that (at ):
... A passing-off action is a remedy for the invasion of a right of property not in the mark, name or getup improperly used, but in the business or goodwill likely to be injured by the misrepresentation made by passing off one person's goods as the goods of another. Goodwill, as the subject of proprietary rights, is incapable of subsisting by itself. It has no independent existence apart from the business to which it is attached. It is local in character and divisible; if the business is carried on in several countries a separate goodwill attaches to it in each. So when the business is abandoned in one country in which it has acquired a goodwill the goodwill in that country perishes with it although the business may continue to be carried on in other countries. ...
135. Our courts have since largely followed Star Industrial, holding that a foreign trader which does not conduct any business activity in Singapore cannot maintain an action in passing-off here (see Tan Gek Neo, Jessie v Minister for Finance and Another  2 MLJ 301, Jumbo Seafood v Hong Kong Jumbo Seafood Restaurant  1 SLR 860 and CDL Hotels). The hard-line approach, as currently applied in Singapore, draws a clear distinction between goodwill and reputation: "[a] desire to become a customer of the plaintiff without the ability to actually be one, cannot ordinarily form the basis of goodwill" (Amanresorts at , see also CDL at , both citing Anheuser- Busch).
136. One aspect of the hard-line approach has however been softened. In CDL Hotels this court relaxed the restriction that there must be actual trading for goodwill to exist. Following the lead of English and Canadian authority, as well as acknowledging industry practice, this court held that (at ):
... pre-business activities can generate goodwill. This is a commonsense approach and is in line with commercial reality where promoters of businesses usually embark on massive advertising campaigns before the commencement of trading to familiarise the public with the service or product. .
137. The conduct of pre-trading activity is one of the grounds upon which the Opponents in the present case seek to establish that goodwill subsisted in the ST. REGIS Singapore prior to the relevant date.
Goodwill can be generated by pre-trading activity
138. In CDL Hotels, this court identified two categories of activity relevant to the establishing of goodwill prior to the official commencement of trading in Singapore. The first was what the court termed "business activities which had gone into full swing and generated considerable income for the respondents" (at ). This included confirmed room reservations and retail tenants which the respondent had secured for the retail wing of its Singapore establishment. The finding that such actual trading activities are capable of generating domestic goodwill is uncontroversial and does not depart from the traditional hard-line approach. The importance of CDL Hotels was its recognition of a second category of non-income generating "pre-trading activities" as capable of contributing to goodwill. These were the respondent's "large-scale advertising" efforts costing around $1m over one year, the steady stream of press-releases and media events, entering into an operating agreement with Ritz-Carton as to the running of the hotel facility, and widely- publicised promotional events such as two topping-up ceremonies (see CDL Hotels at ). This court relied on the English cases of WH Allen & Co v Brown Watson Limited  RPC 191 (" WH Allen") and British Broadcasting Corporation v Talbot Motor Co Ltd  FSR 228 ("BBC"), where the relevant activity consisted of extensive print and television publicity. The Canadian case of Windmere Corp v Charlescraft Corp Ltd (1988) 23 CPR (3d) 60 ("Windmere Corp") was also considered. The relevant pre-trading activity there was a "concerted effort to introduce the trade mark to a wide segment of its purchasers in a short period of time", involving distribution of samples and promotional materials for the upcoming product, and exhibiting the product at trade shows.
139. CDL Hotels however identified a third category of activity which would not suffice to generate goodwill, citing Amway Corporation v Eurway International Limited  RPC 82 ("Amway"). In Amway, the plaintiff had initially conducted "minor trading activity" in the UK, before seeking to expand its UK operations by sourcing for premises and interviewing personnel to run the UK office. These activities were found inadequate to show that the plaintiff had business activity in the UK which was protectable under the law of passing-off. The latter activities were described as "mere preparations for trading" in CDL Hotels (at ).
140. In the present case, the Opponents argued that the activity conducted prior to the official opening of the ST. REGIS Singapore fell on the side of relevant pre-trading activity, while Staywell contended that they were "mere preparations to trade". We do not think that it is ultimately helpful to preserve this dichotomy or to resort to labels such as these. It is not always the case that non-publicity activities such as administrative preparations have no relevance to the establishment of goodwill in Singapore. Pre-trading activity may take many forms and it is always possible that a combination of various types of activity may suffice to generate a sufficient degree of consumer interest and demand, amounting to tangible and protectable goodwill. As always, the sufficiency of the pre-trading activity depends on its particular nature and intensity (CDL Hotels at ). Instead, we provide the following guidance.
141. First, pre-trading activity need not be revenue-generating. This stems from the understanding that goodwill does not focus on the income of the trader per se but the response of the consumer. The fact that income is generated is only a proxy for the attractive force of the business. This court's holding in CDL Hotels acknowledged that valuable demand may be created in a product or business prior to its being monetized.
142. Second, the relevant pre-trading or pre-business activities should unequivocally evince the intention of the trader to enter into the Singapore market. This is consistent with the cases which have enlarged our understanding of how and for what purpose goodwill can be created. In WH Allen, BBC and Windmere Corp, the advertising activity conducted was not of the sort that created brand awareness in general, but was geared toward creating tangible demand for an upcoming product. This court in CDL Hotels also had in mind such 'pre-launch' advertising when it recognised (at ) that "promoters of businesses usually embark on massive advertising campaigns before the commencement of trading to familiarise the public with the service or product" (emphasis added). The fact that the traders in CDL Hotels, WH Allen, BBC and Windmere Corp had proceeded with administrative arrangements such as entering into operating agreements for the soon-to-be-opened hotel, or printing and manufacturing the advertised products in preparation for sale, left no doubt that the demand created by the advertising would eventually be satisfied. This distinguishes the situation from that in Anheuser-Busch, where any demand created would not be met due to the unavailability of the product within the jurisdiction.
143. More recently in Starbucks (UK) Ltd v British Sky Broadcasting Group Plc  EWHC 3974 (Ch), Arnold J held that although advance advertising could create protectable goodwill, the activities relied on by the foreign plaintiff to establish goodwill in the UK "d[id] not take the form of advertising or promotion for a forthcoming product or service". Instead, the advertisements were primarily aimed at promoting its home business in Hong Kong (at -). The Australian position on pre-trading activity is also consistent with this approach. In Turner v General Motors (Australia) Pty Ltd  42 CLR 352, Dixon J found that protectable goodwill in General Motors (Australia) was created through widespread and prolonged advertising in Australia of the plaintiff's intention to set up its Australian undertaking. The plaintiff also secured offices and commenced the erection of works at the site. It was found that due to the advertising, the Australian public came to associate the term "General Motors" specifically with the upcoming Australian undertaking. This approach is also recognised in The Law of Passing-off at 3-063, which states that:
Despite these reservations [as to whether pre-trade activity can generate goodwill], it is suggested that advertising directed at a specific market in actual preparation for trading does generate sufficient goodwill to support the action. Lord MacNaughten's description of goodwill as "the attractive force which brings in custom" is consistent with goodwill existing because advertising has created a demand of the claimant's business before it commences.
144. Besides consistency with the authorities, filtering pre-trading activity in this way reflects a key rationale underlying the hard-line approach, namely to prevent unwarranted stifling of local enterprise. This rationale was recognised by Cooke J in Dominion where he read the hard-line group of cases as being motivated by the concern that (at 405):
. unless a trader has already entered or at least is clearly about to enter the local market, and thus contributes to the local economy or is about to do so, the local law should not allow him to stifle local enterprise. .
In softening the hard-line approach so as to consider pre-trading activity as capable of generating goodwill, our courts must strike a considered balance between the interests of foreign or aspiring entrants to the Singapore market, and existing local enterprises. Where the foreign trader has no interest in the local market, the concern of stifling local enterprises prevails. Where the foreign trader has unequivocally evinced his intention to enter the local market, the concern of preventing local traders from "free-riding" on the efforts and expenditure of the foreign trader prevails. It will be a rare case, if ever, that spill-over advertising meant primarily for a foreign audience or to increase brand reputation in general, or early stages of negotiation and sourcing without any serious commitment to investment, will displace the interests of local enterprise in favour of proprietary protection for the foreign trader.
145. Finally, the unequivocal intention to enter the local market is a necessary factor qualifying pre-trade activity for consideration in the goodwill inquiry, but it is not sufficient. It must be remembered that the ultimate question is whether the activity has generated an attractive force that will bring in custom when the business eventually materialises. If a trader has taken steps evincing his intention to trade in Singapore, such as securing premises or employees here, but has done nothing to put the business in the awareness of the public so as to create demand, then there clearly is no Singapore goodwill to be protected. Such was the case in Amway, where the plaintiff, despite taking firm steps toward establishing a business presence in the UK, had not on the facts generated actual demand for its business whether by trading, advertising or otherwise. Similarly in Athletes Foot Marketing Associates Inc v Cobra Sports Ltd  RPC 343 Ch D, the plaintiff American trader unsuccessfully argued that it had goodwill in the UK by virtue, among other things, of its seeking out English franchisees. The plaintiff was unsuccessful because no evidence was adduced to prove that the English public knew about the English franchisees or that the plaintiff had done anything to seek retail custom. Against this background, we return to the final question, which is whether the activity conducted, whatever its form, suffices to generate an attractive force in the business which will then materialise into actual custom.
Application to the facts
146. In the present appeal the Opponents pointed to three forms of pre- trading activity: (a) advertisements promoting the ST. REGIS brand in general and the opening of the ST. REGIS Singapore in particular; (b) the securing of restaurant tenants; and (c) the holding of a much-publicised job fair to hire hotel staff. Turning first to the latter two activities, we do not agree with the Judge's focus on the fact that they were non-income generating. In our view these were relevant endeavours as they indicated an unequivocal intention to enter the Singapore market-an intention which did in fact materialise. However an unequivocal intention to trade in Singapore is simply the gateway through which pre-trading activity enters the court's consideration as to the subsistence of goodwill. The question remains whether the activity was sufficient to generate demand. For this, we must examine the nature and intensity of the activity. In our view, the securing of restaurant tenants and the hiring of employees were not, on these facts, activities which themselves were capable of creating demand in the ST. REGIS Singapore as a hotel. The goodwill upon which the Opponents built their case in passing-off was not goodwill in the ST. REGIS as an employer or a landlord, but as a provider of hotel and dining services. The relevant pre-trade activity should therefore be of the kind which informed and attracted the public to consume its hotel and dining services. It follows that we must therefore consider the Opponents' advertising efforts prior to the launch of the ST. REGIS.
147. The Opponents submitted that they spent over US$300 million in 2008 on advertising for its ST. REGIS mark. This was not particularly helpful to their case, because we were unable to tell how much of this advertising related to the ST. REGIS Singapore rather than other ST. REGIS hotels or the ST. REGIS brand in general. As we have stated, our focus is on publicity activity preceding the actual commencement of trade within the jurisdiction, but directed at creating demand that would be satisfied by it. The print advertisements exhibited in the Statutory Declaration of the Opponents' Vice- President, spanning close to 200 pages, were of more relevance as most were specific to the opening of the ST. REGIS Singapore. In fact, these were pre- launch advertisements of the type referred to in CDL Hotels, BBC and WH Allen. But the level of pre-launch advertising in this case fell significantly short of that conducted in those cases. The pre-launch advertisements exhibited spanned only one calendar month (between November and December 2007), and only three of these were published in a Singapore publication, namely the Straits Times. This by no means indicates that the rest of the publications were not viewed by the Singapore public, but it does hinder our ability to conclude that the Opponents' advertising was extensive and widespread and targeted at creating demand within the jurisdiction.
148. We also note that there was no mention of physical promotional events, or any use of digital and television media. It may be that the Opponents had the same limited and targeted marketing strategy as was adopted in Amanresorts (at ). However, it is difficult for us to conclude based on this limited marketing that the Opponents had actually succeeded in creating demand amongst its intended segment of the public. This is different from Amanresorts where the success of the marketing strategy was evidenced by actual demand. Absent any kind of survey or other evidence, our finding is that the 200 pages of print advertisements in a number of foreign magazines (in fact it was substantively less, given that many pages were simply covers of the magazines containing the advertisements) are insufficient to warrant a finding of the generation of an attractive force that would bring in custom in Singapore at the date of the Application.
Can a foreign business own goodwill in Singapore?
149. The Opponents' alternative argument was that there was goodwill in the ST. REGIS Singapore prior to its launch by virtue of the international reputation of the ST. REGIS brand, coupled with the presence of Singapore customers who purchased hotel services in ST. REGIS hotels abroad. The Opponents relied primarily on the decisions in Hotel Cipriani Srl v Cipriani (Grosvenor Street) Ltd  RPC 9 ("Hotel Cipriani (HC)") and Hotel Cipriani Srl v Cipriani (Grosvenor Street) Ltd  Bus L R 1465 ("Hotel Cipriani (CA)) (collectively referred to as "Hotel Cipriani") as authority for the proposition that the presence of customers within the jurisdiction meant that a trader had business within the jurisdiction relevant to the finding of local goodwill.
150. The Opponents' submission appears to simplify the position in the Hotel Cipriani cases, which in fact reiterated the position that to sustain a claim in passing-off, mere reputation was insufficient and the plaintiff must prove that it had English goodwill. Further, the cases preserved the former English position that in order for a foreign service trader to establish English goodwill, the trader must (a) have customers in the UK and (b) bookings for the trader's foreign services must have been made from within the UK ("the direct bookings requirement"). It was the latter requirement which distinguished the dif
Please Login To View The Full Judgment!
ferent conclusions reached in Sheraton Corporation of America v Sheraton Motels Ltd  RPC 202 ("Sheraton") and in Crazy Horse. In Sheraton, the plaintiff's customers in the UK made bookings for the plaintiff's hotels abroad via the plaintiff's booking office in the UK, and the plaintiff was found to have goodwill in the UK. In Crazy Horse on the other hand, the customers from the UK who visited the plaintiff's famous saloon in Paris did not make their bookings from the UK, and the goodwill element was not made out. 151. In Hotel Cipriani, Arnold J in the English High Court and Lloyd LJ in the Court of Appeal found that the plaintiff owner of the famous Hotel Cipriani in Venice owned valuable English goodwill owing to the substantial body of customers from England, significant marketing efforts directed at the relevant English public, and a significant volume of bookings placed directly from England (see Hotel Cipriani (HC) at - and Hotel Cipriani (CA) at ). In Hotel Cipriani (HC) Arnold J summarised a number of propositions governing the goodwill requirement in English law, which included the following (at -): Fourthly, in order to found a passing off claim in the United Kingdom, the claimant must own goodwill in the United Kingdom. It is not enough to have a reputation here: see Anheuser-Busch Inc v Budejovicky Budvar NP  F.S.R. 413. Fifthly, it is sufficient for goodwill to exist in the United Kingdom that the claimant has customers or ultimate consumers for his goods here, and for this purpose it is immaterial whether the claimant (a) has some branch here or (b) trades directly with customers here without having any physical presence in the jurisdiction (for example, by mail order) or (c) trades through intermediaries such as importers and distributors (provided that the circumstances are not such that the goodwill is owned by the intermediary)... Sixthly, in the case of claimants who provide services which are physically performed abroad, it is sufficient for goodwill to exist in the United Kingdom that the services are booked by customers from here: compare Sheraton Corp of America v Sheraton Motels Ltd  R.P.C. 202 with Alain Bernardin et Cie v Pavilion Properties Ltd  R.P.C. 581. [emphasis added] 152. Both Arnold J and Lloyd J expressed the desire to review the direct bookings requirement (see Hotel Cipriani (HC) at  and Hotel Cipriani (CA) at ), noting the contrary approach taken in Pete Waterman v CBS United Kingdom Ltd  EMLR 27 and a suggestion for a different test proposed by Prof Wadlow in The Law of Passing-off (at para 3-095). However, neither Judge was prepared to adopt these alternative approaches in Hotel Cipriani, largely because the facts did not call for it. It therefore remains necessary to establish that direct bookings are made by customers within the jurisdiction in order to establish the English goodwill necessary to maintain a passing-off claim under English law. 153. The decisions in Hotel Cipriani are in fact in line with this court's decision in Amanresorts. In that case, the respondent's luxury hotel was located in Bali, but bookings were made by Singapore-based customers in the respondents' two Singapore reservations offices. This was sufficient to establish the requisite local custom and business necessary to show that there was goodwill in Singapore attached to the respondents' Balinese resort. It was therefore not necessary for this court to consider the situation of a trader with an entirely foreign presence and where bookings were not made in Singapore (at ). In the present case, as the Judge rightly noted, the Opponents were unable to prove that room reservations made by Singaporean members of the Opponents' Starwood Preferred Guest ("SPG") loyalty programme for other ST. REGIS hotels located abroad, had been made in Singapore. In fact, it was not even proven that these reservations were made for ST. REGIS hotels rather than hotels under the Opponents' numerous other brands. Under the current position, the Opponents were therefore unable to prove that protectable goodwill subsisted in the overseas ST. REGIS hotels. 154. The Opponents however went further. They invited this court to depart from the hard-line position which extends protection to local goodwill but not to reputation that is unconnected to local custom. It was argued that the hardline approach was inconsistent with the positions taken in the UK, Australia and Hong Kong. 155. In fact, the position taken in each of these jurisdictions is distinct and somewhat varied. The English position has already been discussed. Australia has most clearly gone down the soft-line approach, holding that a passing-off action may be maintained based on a trader's reputation in the jurisdiction, notwithstanding the absence of business premises, business activity or products in the jurisdiction. In Conagra v McCain Foods (1992) 23 IPR 193 ("Conagra"), after reviewing Australian, Irish, English, American and Canadian authority Lockhart J concluded as follows (at 235): I am of the opinion that it is not necessary in Australia that a plaintiff, in order to maintain a passing off action, must have a place of business or a business presence in Australia; nor is it necessary that his goods are sold here. It is sufficient if his goods have a reputation in this country among persons here, whether residents or otherwise, of a sufficient degree to establish that there is a likelihood of deception among consumers and potential consumers and of damage to his reputation. 156. The court's reasoning appears to extend protection beyond reputable foreign traders with customers within Australia, covering also reputable foreign traders with only the potential of obtaining such customers should it commence trade in Australia. As stated by Gummow J (at 263): . where the plaintiff, by reason of business operations conducted outside the jurisdiction has acquired a reputation with a substantial number of persons who would be potential customers were it to commence business within the jurisdiction, the plaintiff has in a real sense a commercial position or advantage which it may turn to account. 157. As can be seen, the Australian position goes much further than the English one in liberating the action in passing-off from even the requirement of goodwill, let alone the requirement of customers in the jurisdiction who have made purchases or bookings from within the jurisdiction. 158. As for Hong Kong, the decided cases are few and uncertain. The Opponents cited JC Penney Co Inc v Punjabi Nick  FSR 26, which concerned the grant of an interlocutory injunction in favour of the American plaintiff against a local defendant which conducted its business under the name "Penneys Fashion" in wording identical to that used in the plaintiff's logo. The plaintiff in fact had a Hong Kong subsidiary through which it conducted retail purchases. The court's decision turned on the finding that the plaintiff's Hong Kong subsidiary "does a very substantial business here", and that "the business which it does carry on through its subsidiary must necessarily enjoy a local reputation which the plaintiff is entitled to protect". Therefore, this was not a case in which the plaintiff's protection rested on nothing more than an intangible reputation. In fact, it was in line with the traditional position that a plaintiff may establish goodwill through even modest business activities conducted by its local agents or subsidiaries. The court did opine (at 27) that: In these days of expanding travel and tourism I would anticipate a development in the law of passing off whereby the owners of reputation and goodwill acquired abroad may be granted protection for that reputation and goodwill here even though they do not carry on business here and even though its goods themselves have earned no reputation here. 159. However the court did not go so far as to state that that was the current position under Hong Kong law. In the oft-cited case of Ten-Ichi Company Ltd v Jancar Limited and Others  FSR 151, also concerning the grant of an interlocutory injunction, the Judge indicated partiality to a more soft-line approach and noted that the Canadian and New Zealand courts in Dominion and Orkin respectively "have recognised international reputation", but did not proceed to analyse these cases. The Judge then used the terms "reputation" and "goodwill" interchangeably, finding in the result that the plaintiffs had "international goodwill" and that some of that goodwill existed in Hong Kong. It is unclear whether the Judge intended to conflate the two concepts (as Lockhart J did in Conagra), or whether he had unintentionally equated the two distinct concepts. The reasoning deployed in the case has been questioned and it has been noted that the case of Anheuser-Busch was not considered at all (see The Law of Passing-off at para 3-117 and Lim Cheng Saw "Goodwill hunting in passing off: time to jettison the strict "hard line" approach in England?" JBL 2010, 8, 645-669 at p 655). The Hong Kong cases are therefore not instructive. 160. Prof Ng-Loy Wee Loon suggested, perhaps having regard to the reality of an increasingly transnationalised world, that the following test proposed by Prof Wadlow in The Law of Passing-Off at para 3-095 is one we could adopt: It is now suggested that a service business operating from a place or places abroad has customers and therefore goodwill in England to the extent that persons from England consciously seek out and make use of its services in preference to those available from competitors, in England or elsewhere. So the foreign business has goodwill here if English residents are prepared to go to it (literally or figuratively) to avail themselves of its services, or if the availability of those services abroad is a material factor in their travelling to wherever the services can be acquired or experienced. 161. There is much to commend such an approach given the widespread practice of international travel which is now commonplace in Singapore, as well as the prevalence of accessing services through the internet by means of on-line bookings. Having said that, we prefer to leave this determination to a subsequent occasion when it is necessary for the determination of the issues presented. This is not that occasion, not least because even if we were to adopt Prof Wadlow's formulation in this case, it would not make a difference to the Opponents' position given that there was no evidence of Singapore residents who travelled abroad for the purpose of seeking out and staying at St Regis hotels there. Conclusion on goodwill 162. For these reasons, we affirm the Judge's dismissal of the opposition under s 8(7) of the Act. Summary of our holdings 163. For convenience and without displacing the need to read what we have set out above in its full context, we summarise our principal holdings in this case: (a) We maintain the step-by-step approach to the analysis of both opposition and infringement claims (see  above); (b) We reject the notion that there is any particular or notably low threshold at the marks-similarity stage of the inquiry. The different aspects of similarity are to be applied as signposts towards answering the question of whether or not the marks as a whole are similar. This analysis is done mark-for-mark without consideration of external material. But at the confusion-stage of the inquiry it would be appropriate to have regard to the relative importance of the various aspects of similarity having regard to the nature of the goods and services in question (see ,  and  above); (c) The marks are to be assessed as composite wholes but bearing in mind and having due regard to distinctive and dominant elements (see - above); (d) Unlike aural similarity which considers the utterance of the words without regard to their meaning, conceptual similarity seeks to uncover the ideas that lie behind and inform the understanding of the marks (see  above); (e) Where a good or service in relation to which registration is sought falls within the ambit of a specification in which the incumbent mark has been registered, the goods or services in question are to be regarded prima facie as identical (see - above); (f) There is a difference at the confusion-stage of the analysis between opposition proceedings and infringement proceedings. In the former, the court must have regard to the full range of actual and notional uses of the marks; whereas in the latter, the court compares the full range of notional fair uses of the incumbent mark against the actual use of the later mark (see - above); (g) It is not relevant to have regard to extraneous factors concerning the actual and particular circumstances and way in which the mark has been used on the goods in question in opposition proceedings. In both opposition and infringement proceedings, it is impermissible at the confusion-stage of the inquiry to consider extraneous factors consisting of steps taken by a trader to differentiate his goods or marks from those of the owner of the incumbent mark. Extraneous factors that relate to the nature of the goods or services, the typical purchasing practices that would attend such goods or services and the degree of care that would be applied by the consumer would be relevant (see ,  and - above); (h) The doctrine of initial interest confusion has no application in Singapore (see -); (i) Where the field of trade in which the defendant in infringement proceedings or the applicant for registration operates is in close proximity or is a natural extension of that of the incumbent's business, damage in the form of a restriction of business expansion opportunities will generally be inferred (see -) above; and (j) The hard-line approach to goodwill is softened in Singapore to the extent that pre-trading activity need not be revenue-generating as long as it is directed at generating demand for the plaintiff's business and the trader evinces an unequivocal intention to enter the market (see - above). Conclusion 164. For the reasons we have given, we allow the appeal in CA 148/2012 with respect to the opposition under ss 8(2)(b) and 8(4)(b)(i) of the Act. We dismiss the appeal with respect to the opposition under s 8(7)(a) of the Act. We also dismiss the appeal in CA 147/2012. The Opponents are to have 85% of their costs for both appeals and similarly 85% of their costs below having regard to the fact that the Opponents failed in their opposition under s 8(7)(a). These are to be taxed if not agreed. The Applicant Mark may not proceed for registration. 165. We are grateful to all counsel for their assistance in this matter. We are especially grateful to Prof Ng-Loy Wee Loon for the tremendous assistance she rendered us in this matter. Prof Ng-Loy made available to us the benefit of her great breadth of knowledge in this field and she clarified a number of important matters for us. 1. See Appellant's Core Bundle Vol 2 ("2ACB") (Part A) in CA 148/2012 at p5, para 4 2. See 2ACB (Part A) in CA 148/2012 at p 6 para 7, and the Principal Assistant Registrar's Grounds of Decision ("PAR's GD") in 2 ACB (Part B) in CA 148/2012 at  3. See PAR's GD in 2 ACB (Part B) in CA 148/2012 at  4. See Grounds of Opposition in Record of Appeal Vol 2 at pp 21-26 5. See 2 ACB (Part B) in CA 148/2012 at - See 2 6. ACB (Part B) in CA 148/2012 at - See 2 ACB 7. (Part B) in CA 148/2012 at  and  See 2 ACB (Part B) in CA 148/2012 at - 8. See Grounds of Decision ("GD") at - 9. See GD at - 10. See GD at - 11. see Tab 13 of Annex B 12. See GD at - 13. See GD at  14. See GD at - 15. See GD at - 16. See GD at - 17. See 1 ACB in CA 148/2012 at pp 45-46 See 18. RA Vol 2 at pp 13-16 19. See RA Vol 2 at pp 5-7 20. See RC in CA 147/2012 at  -  21. See AC in CA 147/2012 at -