Digi International Inc v Teraoka Seiko Co, Ltd [2021] SGHC 165

This is a cross-appeal against the Intellectual Property Adjudicator’s (“the Adjudicator”) decision dated 2 August 2019 in Digi International Inc. v Teraoka Seiko Co., Ltd [2020] SGIPOS 1 (“the GD”) to refuse the registration of Trade Mark No. (“SGTM No.”) 40201700142X (“the Application Mark”) under International Registration No. (“IR No.”) 1326111. The Application Mark is as follows:

and is sought to be registered in Classes 9, 38 and 42 (“the Application Mark’s Specifications”). The specification of goods and services under each Nice Class is set out in full in Annex 1.

In the GD, the Adjudicator held that the respondent’s opposition to the registration of the Application Mark succeeded under ss 8(2)(b), 8(4)(b)(i), read with 8(4)(a), and 8(7)(a) of the Trade Marks Act (Cap 332, 2005 Rev Ed) (“the TMA”). However, the Adjudicator dismissed the ground of opposition under s 7(6) of the Act. The appellant is appealing against the portion of the Adjudicator’s decision granting the respondent’s opposition under s 8 of the TMA, while the respondent is appealing the refusal to grant its opposition under s 7(6).

The parties

The appellant is a company headquartered in the US. It was constituted in 1985 under the name “Digiboard Inc.”, but changed its name in 1994 to “Digi International Inc.” to reflect the growing scope of its products and business. The appellant’s focus has always been on connecting things, starting with intelligent multiport serial boards for personal computers (“PC”) sold under the mark “DIGIBOARD”. As wireless data technologies evolved, the appellant expanded its product lines. Today, its products include computer hardware such as cellular routers, gateways, wireless communication adapters, serial servers, intelligent console servers and electronic sensors and cameras used to monitor and collect data and transmit it to a host or a remote computer system. Such collection and transmission of data is done to provide alerts, in the case of triggering events, and surveillance and real-time environmental sensors for measuring and transmitting temperature, humidity and light information. Its products are used “wherever connection with machines is required”, thereby forming the essential layer of machine-to-machine (“M2M”) communications, and remote monitoring and management of devices. However, the appellant claims that it does not manufacture scales, weighing devices or printers.

The appellant also sells “many of [its] products…to Original Equipment Manufacturers (‘OEMs’)” which are then incorporated into products sold under other brands. Outside of the US, the applicant has offices in countries including Singapore, Germany, Japan, Australia and China and has distributors in over 200 locations worldwide.

In Singapore, the appellant has the following trade mark registrations (collectively, “the Appellant’s DIGI-related Marks (Singapore)”) which feature the word “DIGI” or variations thereof:

The respondent is a Japan-based company headquartered in Tokyo. It was founded in November 1934. It began selling electronic scales named “DIGI” in 1971 and commenced exporting them in the same year. By the latter half of the 1980s, the respondent had established production bases in Singapore, China and the UK. Today, its products include Point-of-Sale (“POS”) equipment and systems, scales, labellers, sorting machines and systems, vending machines and cloud and server systems. The respondent presently caters to four main business fields: retail, food, logistics and hospitality.

The respondent is the registered proprietor of, inter alia, SGTM No. T8605807D (“the Respondent’s Mark (Singapore)”):

The Respondent’s Mark (Singapore) has been registered in Class 9 for “[b]alance and scale, scale equipped with printer, printer” since 18 December 1986. The respondent has also used the “DIGI” trade mark for scales and measuring equipment and related goods and services since at least April 1987.

The respondent’s wholly owned Singapore subsidiary is Digi Singapore Pte Ltd (“the Respondent’s Singapore Subsidiary”). It was formerly called “Teraoka Weigh-System Pte Ltd”, but was renamed in 2016. The Respondent’s Singapore Subsidiary is also the regional marketing centre for South East Asia.

Facts

The 2002 Agreement

The parties are no strangers to each other. In 2002, the appellant and respondent entered into an agreement (“the 2002 Agreement”). It arose out of a separate dispute – a trade mark opposition filed by the appellant in Germany against the respondent’s application to register a mark identical to the Respondent’s Mark (Singapore) (“the German Dispute”). The basis for the appellant’s opposition was its earlier registration of a mark identical to the Appellant’s Old Mark (Singapore) (German Trade Mark Registration 2093840, “the Appellant’s Old Mark (Germany)”):

In the German Dispute, the respondent responded by filing an action to cancel the registration of the Appellant’s Old Mark (Germany) on the basis of its earlier registration for “DIGI” (German Trade Mark Registration 970360, “the Respondent’s DIGI Word Mark (Germany)”).

Eventually, the parties’ reached a settlement in the form of the 2002 Agreement. In brief, the appellant and respondent withdrew their opposition and trade mark cancellation actions respectively. A mark identical to the Respondent’s Mark (Singapore) was registered in Germany (German Trade Mark Registration 39406977, “the Respondent’s Mark (Germany)”). The parties agreed not to oppose the registration, renewal and/or use of each other’s marks, covered by the 2002 Agreement, as long as each kept to a demarcated scope of goods and services specified in the 2002 Agreement. Salient terms of the 2002 Agreement are as follows:

(a) The appellant agrees not to attack the existing or new registrations, renewals and/or use of, for instance, the German trademark registrations 39406977 (ie, the Respondent’s Mark (Germany)) and 970360 (ie, the Respondent’s DIGI Word Mark (Germany)) by the respondent, as long as the use and registration of the aforementioned mark are for goods and services similar to those covered by the respondent’s two German trade marks;

(b) The respondent agrees not to attack existing or new registrations, renewals, and/or use of, for instance, the German trademark registration 2903840 (ie, the Appellant’s Old Mark (Germany)) by the appellant, as long as the use and registration of the mark are for goods and services similar to those covered by the appellant’s German trade mark and such registrations and/or use does not refer to identical goods and services covered by the respondent’s two German trade marks ;

(c) In summary, the 2002 Agreement permits the appellant to register the defined marks for “data processing devices and computer, namely, micro-computer hardware for use in connection with multiple user systems and local networks, especially server based asynchronous serial boards, ISDN-network access cards and synchronous boards, ISDN-LAN-bridges remote access servers, LAN-router, single and multiple channel transceivers, media transformer, micro hubs and modular repeaters, network printer server, cluster controller systems, modems and fax modems; software programmes brought on data carriers, namely micro computer software for the operation of the aforementioned micro computer hardware, software for terminal emulation, software for data communication in multi user systems, software for data communication in local networks and software for remote access to local networks”, so long as the foregoing do not include uses by the appellant as stated below at [9(d)] (ie, goods and services concerning weighing and/or measuring purposes or transfer of data connected to weighing and/or measuring processes) (collectively, “the Appellant’s Specified Goods and Services”);

(d) Under the 2002 Agreement, the respondent is permitted to register the defined marks for (collectively, “the Respondent’s Specified Goods and Services”):

(i) spring scales, electronical and electrical scales, parts of such goods and equipment therefore, namely digital or analogue display devices, keyboards, writer and printer (from German trademark registration 970360); and

(ii) packaging machines, including vacuum packaging machines; automatic filling and packaging machines; transport and conveying machines, especially band conveyors, conveyor belts, conveying devices, loading and charging devices for shelves and racks, weighing and labelling devices which are combinable with conveying machines; machines, electronical packaging devices, electronical manual labelling machines, devices and systems, for packaging, labelling and inscripting, respectively, machines and devices for weighing, packaging, labelling and assessing postage of goods which are ready to be dispatched; parts of the aforementioned goods; electrical and electronical weighing, measuring, signal and control apparatus and devices (as far as included in Class 9), systems for sorting, weighs, especially digital weighs, especially having electronic and digital display devices, weighing and pricing systems, electronical printer, displays combined with printers, label printers, especially with cassette plug-ins; code control print-devices, pricing devices, counters, electronical registration devices, apparatus and instruments for coding and decoding, apparatus and devices for register controlling, surveying and commanding charge of shelves and for the shipment; cash register, machine and devices for automatic book-keeping; storage and for the stock-keeping; parts of the aforementioned goods; calculators, data-processing devices and computers, except for, micro-computer hardware for use in connection with multiple user systems and local area networks, namely: server based asynchronous serial boards, ISDN-network access cards and synchronous boards, ISDN-LAN-bridges remote access servers, LAN-router, single and multiple channel transceivers, media transformer, micro hubs and modular repeaters, network print server, cluster controller systems, modems and fax modems, software programs brought on data carriers, namely micro computer software for the operation of the aforementioned micro computer hardware, software for terminal emulation, software for data communication in multi user systems, software for data communication in local networks and software for remote access to local networks, as far as the excluded goods or services concern subject matter without connection to weighing and/or measuring purposes or transfer of data connected to weighing and/or measuring processes; parts of the aforementioned goods; service of an engineer in the field of weighing packaging, filling and storing technique (from German trademark registration 39406977).

It is not disputed that the 2002 Agreement remains valid and binding on both parties (GD at [17]). Clause 6 defines the scope of the 2002 Agreement in these terms:

Genesis of the Application Mark

The Application Mark was launched by the appellant on its 30th anniversary to give its corporate logo a “new brand identity and look and feel”. It was intended to replace the mark identical to the Appellant’s Old Mark (Singapore).

In its announcement of the launch of the Application Mark on 26 October 2015, the appellant described the Application Mark as follows:

The appellant has since obtained protection of the Application Mark in several territories. The ten territories marked with an asterisk are those in which the respondent has also registered marks identical to Respondent’s Mark (Singapore) (“the Overlapping Territories”):

In addition to the territories in the table above, the respondent contends, and the appellant has not disputed, that the appellant has registered the Application Mark in Australia, Canada, the EU and South Korea. The respondent has opposed the registration of the Application Mark in these four territories. The Overlapping Territories will therefore include these four territories.

Proceedings in Singapore

The Application Mark was accepted for registration and published on 18 August 2017 for opposition purposes. The respondent filed its Notice of Opposition on 17 October 2017.

On 16 July 2019, about two and a half weeks prior to the hearing before the Adjudicator, the appellant wrote to the Registrar of Trade Marks to request an amendment to its specification of goods in Class 9. It sought to replace the words “computer printer servers” with “none of the aforementioned goods relating to balance and scale, scale equipped with printer, printer” (“the appellant’s Class 9 Amendment”). These changes are reflected in [A.1] and [A.2] in Annex 1. Despite the change being made, the respondent proceeded with the opposition. The Adjudicator rendered her decision on 20 January 2020.

Decision below

To re-cap, the Adjudicator allowed the respondent’s opposition under ss 8(2)(b), 8(4)(b)(i) and 8(7)(a) of the TMA. However, the ground of opposition under s 7(6) of the TMA was dismissed.

s 8(2)(b) TMA

Under s 8(2)(b) of the TMA, the Adjudicator applied the step-by-step approach set out by the Court of Appeal in Staywell Hospitality Group Pty Ltd v Starwood Hotels & Resorts Worldwide, Inc and another and another appeal [2014] 1 SLR 911 (“Staywell”). This approach requires the sequential assessment of three requirements: (a) similarity (or identity) of marks; (b) similarity (or identity) of goods or services; and (c) the likelihood of confusion arising from the two similarities.

First, the Adjudicator found that the Application Mark and Respondent’s Mark (Singapore) were similar. They were aurally identical because both comprise the letters “DIGI” which were read phonetically as “dee gee”. In terms of visual similarity, the Adjudicator observed that the text “DIGI” was not “particularly” inherently distinctive. Whilst it was not a word with a dictionary meaning, it was intuitively a shorthand for the word "digital" and connoted goods and services which were digital in nature. However, she noted that the stylised manner of the text and the triangle device accompanying it conferred the composite mark a “medium degree of distinctiveness” (at [45]):

In comparison, the Adjudicator noted that the Application Mark was also set against a white background, used a sans serif font type with no hooks, had all capitalised letters and a “D” with an unclosed border at the top end. It also had a triangular mark in green which lay to the top right corner of the text pointing towards the right. While not identical, the Adjudicator regarded both marks as being “visually similar in terms of the textual presentation and use of similarly shaped device”. She concluded that “adopting the position of an average consumer with an imperfect recollection, it would be understandable if a consumer considered the … Application Mark to be similar to, like a more modernised version of, the [Respondent’s Mark (Singapore)]”.

As for conceptual similarity, in both marks, the text “DIGI” evoked the word “digital” and connoted a relation to products and services with a digital element. As such, the marks had a high degree of conceptual similarity.

Second, in the next stage of the step-by-step approach, the Adjudicator held that there was similarity between the respondent’s goods in Class 9, and the appellant’s goods and services in Classes 9, 38 and 42. The Adjudicator first noted that in obiter dicta, the Court of Appeal in Staywell at [40] stated that registration in the same specification within a class establishes a prima facie case for identity. In this case, while the appellant’s Class 9 specification is defined extensively as compared to the respondent’s much shorter specification, the 2002 Agreement evinced the parties’ recognition that their respective descriptions could encroach on the other’s goods and services. By the appellant’s own evidence, its products were “used wherever connection with machines is required”. Such products would include even machines relating to weighing or measuring that have elements of connectivity, including the products sold by the respondent.

The Adjudicator then compared Classes 38 and 42, individually, to the respondent’s specification under Class 9. Classes 38 and 42 are named “Telecommunications” and “Scientific and technological services and research and design relating thereto; industrial analysis and industrial research services; design and development of computer hardware and software” respectively. It was held that the specifications in these two classes “closely correspond[ed] with [the appellant’s] specification of goods set out in Class 9” and that “[b]y this connection, it would also overlap with the goods offered by the [respondent] under Class 9”.

The Adjudicator added that the appellant’s Class 9 Amendment did not affect her conclusion. Neither it nor the 2002 Agreement, and the demarcation of each party’s sphere of goods and services thereunder, made the appellant’s goods and services dissimilar to the respondent’s.

Third, a likelihood of confusion as to the existence of an economic link between the appellant’s and respondent’s goods and services was found. The Adjudicator was “inclined to agree” with the respondent that the relevant segment of the public included IT and electronics professionals and enthusiasts as well as the general public. However, she noted that the goods and services marketed by both parties were not typically marketed and sold to the general public at large. Nevertheless, the parties could have “target[ed] similar groups of consumers using technological products and machinery”, ranging from “business consumers to individuals such as IT and electronics professionals or enthusiasts”. The parties’ goods and services covered a broad price range and could have been sold online or in physical stores. Simpler and less expensive products were more likely to be purchased by individuals who would generally have less specialised needs and knowledge and who would pay less attention when selecting the products for purchase. The aural, visual and conceptual similarity of the marks also “significantly increase[d]” the likelihood of confusion. All these factors together led to the satisfaction of the third stage in the step-by-step inquiry. The opposition under s 8(2)(b) therefore succeeded.

s 8(7)(a) TMA

s 8(7)(a) of the TMA sets out the ground of opposition based on the tort of passing off. To succeed in proving passing off, the respondent must establish the classic trinity of goodwill, misrepresentation and likelihood of damage.

First, the Adjudicator accepted that the respondent’s evidence of sales, profits and expenses incurred in promoting the goods and services in association with the Respondent’s Mark (Singapore) in Singapore established its goodwill here.

Second, under element of misrepresentation the Adjudicator was satisfied that the threshold requirement of proving that the Respondent’s Mark (Singapore) was distinctive of its goods and services was met. Further, given the similarity of the parties’ specifications and likelihood of confusion under s 8(2)(b) of the TMA, misrepresentation was established.

Finally, the damage in this case was a real tangible risk of the respondent’s goodwill being adversely affected through a diversion of custom.

s 8(4)(b)(i) read with s 8(4)(a) TMA

To succeed on the ground of opposition under s 8(4)(b)(i) of the TMA, the respondent must show that:

The earlier trade mark in this case is the Respondent’s Mark (Singapore) which was filed on 18 December 1986. As the marks-similarity analysis under ss 8(2)(b) and 8(4) of the TMA is fundamentally the same (Rovio Entertainment Ltd v Kimanis Food Industries Sdn Bhd [2015] 5 SLR 618 (“Rovio”) at [146]), the Adjudicator held that the first element at [31(a)] was satisfied.

The Adjudicator also found that the Respondent’s Mark (Singapore) was well known in Singapore. The principles and factors for determining if a trade mark is well known in Singapore are found in ss 2(1) and 2(7)–2(9) of the TMA:

The relevant sector of the public was consumers who had used the respondent’s products and this did not need to be the public at large. The Adjudicator considered the respondent’s evidence showing, inter alia, the degree and extent to which the Respondent’s Mark (Singapore) had been used on products sold in Singapore since 1987 and the sales and advertising figures of the products”. She accepted that the well known status of the Respondent’s Mark (Singapore) was borne out by the evidence. She also noted that s 8(4) of the TMA grants proprietors of well known trade marks in Singapore protection from registration and/or use of identical or similar marks on dissimilar goods or services (Novelty Pte Ltd v Amanresorts Ltd and another [2009] 3 SLR(R) 216 (“Novelty”) at [229]). Accordingly, the absence of evidence showing the respondent using the Respondent’s Mark (Singapore) on services in Classes 38 and 42 was irrelevant.

Finally, the Adjudicator cited Novelty at [234] for the proposition that the tests for misrepresentation and damage in passing off are substantively the same as the tests for the third and fourth elements at [31(c)] and [31(d)] respectively. Given her views on passing off under s 8(7)(a) of the TMA, the third and fourth elements were satisfied. The ground of opposition under s 8(4)(b)(i) of the TMA therefore succeeded.

s 7(6) of the TMA

s 7(6) of the TMA provides that “[a] trade mark shall not be registered if or to the extent that the application is made in bad faith.” The Adjudicator applied the test set out in Valentino Globe BV v Pacific Rim Industries Inc [2010] 2 SLR 1203 (“Valentino”) at [29]. Namely, whether there is bad faith is to be judged according to the combined test which contains both a subjective element (viz, what the particular applicant knows) and an objective element (viz, what ordinary persons adopting proper standards would think).

First, the Adjudicator rejected the respondent’s submission that the appellant’s conduct ran counter to the 2002 Agreement. In her view, there was no requirement under the 2002 Agreement for parties to inform or consult with each other if a new “DIGI” related mark was developed. She also rejected the suggestion that the 2002 Agreement prohibited parties from adopting a new mark which was more similar to the other’s mark for goods and services beyond the demarcated limits of the 2002 Agreement.

Second, the Adjudicator distinguished two cases relied on by the respondent: Case O-006-17 Trade Mark Application No. 3134673 by Cold Black Label Ltd (“Cold Black Label”) and Ceravolo Premium Wines Pty Ltd v Ma Kirkby (Trpl) Pty Ltd [2018] ATMO 43 (“Ceravolo”). The Adjudicator observed that in these cases: (a) the marks concerned were in fact confusing or deceptively similar (although this is not necessary to establish bad faith); and (b) in both cases, the registration was made not that long after the applicant entered into an agreement with the respondent not to use a confusingly or deceptively similar mark – about a year later in Cold Black Label and in Ceravolo the registration took place even while the agreement was being negotiated. In each case, there was a clear sense that the applicant, while entering into the agreement, was still trying to contravene the spirit of the agreement by surreptitiously seeking to register a similar mark without the other party’s knowledge.

In contrast, the appellant filed the Application Mark 14 years after the 2002 Agreement. Even if the Application Mark is more similar to the Respondent’s Mark (Singapore) than the Appellant’s Old Mark (Singapore), which the Adjudicator made no finding on, this did not translate into an obligation to inform the respondent of the Application Mark or of its registration. There was no common understanding which the appellant appears to have acted against.

As for the application to make the appellant’s Class 9 Amendment, the Adjudicator held that bad faith could not be inferred from the appellant’s Class 9 Amendment. Inter alia, despite the 2002 Agreement, the parties continued to oppose the scope of goods or services specified in various registrations around the world for their respective “DIGI” marks, and the present registration involved a different mark. The ground of opposition under s 7(6) of the TMA therefore failed.

Parties’ submissions

Appellant’s submissions

s 8(9) of the TMA

s 8(9) of the TMA provides that “[t]he Registrar may, in his discretion, register a trade mark where the proprietor of the earlier trade mark or other earlier right consents to the registration.” The appellant submits that the 2002 Agreement evidences the respondent’s consent to its use of the word “DIGI” in its trade marks, subject to the appellant confining the usage of the Application Mark to the Appellant’s Specified Goods and Services. Since the Application Mark’s Specifications fall within the Appellant’s Specified Goods and Services under the 2002 Agreement, the 2002 Agreement constitutes consent by the respondent to the Application Mark’s registration.

While Clause 6 of the 2002 Agreement states that “this agreement is valid for existing trademarks” [emphasis added], the appellant argues that it governs the registration of the Application Mark, which was launched in 2015 (see [13] above). At the hearing, it was argued that a 2003 Settlement Implementation Agreement (“the 2003 Canadian Settlement Agreement”), arising out of the respondent’s opposition to the appellant’s registration of certain “DIGI” related trademarks in Canada, elucidates the ambit of the 2002 Agreement. The recital of the 2003 Canadian Settlement Agreement states that the parties wish to “abide by the spirit of the [2002 Agreement]”. Clauses 8.0 and 9.0 of the 2003 Canadian Settlement Agreement then provide that the parties will not challenge each other’s existing or new registrations, new applications, renewals and/or use of certain “DIGI” related marks, “and/or close variations of such trademarks involving the word DIGI” [emphasis added]. Since the 2003 Canadian Settlement Agreement governs close variations of the “DIGI” related marks specified in that agreement, the appellant argues that the same should be said of the 2002 Agreement. However, the appellant accepts that the 2003 Canadian Settlement Agreement does not apply on a worldwide basis.

s 8(2)(b) TMA

The appellant asks for the ground of opposition under s 8(2)(b) of the TMA to be dismissed. It submits that the step-by-step approach under s 8(2)(b) of the TMA should be analysed with the 2002 Agreement in mind.

At the first stage concerning marks similarity, the appellant advanced two alternative cases on appeal.

Its first case is premised on the view that the respondent tacitly accepts co-existence with the Application Mark if: (a) it is not more similar to the Respondent’s Mark (Singapore) than the Appellant’s Old Mark (Singapore); and (b) the Application Mark’s Specifications keep within the limits of the 2002 Agreement. Accordingly, the appellant argues that the Application Mark is not more similar to the Respondent’s Mark (Singapore) than the Appellant’s Old Mark (Singapore). In respect of the triangle device in the Application Mark, the appellant highlights that the respondent does not claim exclusivity over such a device and that it is unlikely that the Application Mark will inflict any damage on the respondent’s interests in addition to that inflicted by the Appellant’s Old Mark (Singapore).

Its second case, in the alternative, is that the Application Mark and Respondent’s Mark (Singapore) are dissimilar. In this regard, the appellant relies on the 2 February 2021 decision of the Australian Registrar of Trade Marks, Teraoka Seiko Co., Ltd v Digi International Inc. (International Registration Designating Australia 1326111 (9, 38, 42); Australian Trade Mark Application No. 1818298) (“the Australian Registrar’s Decision”). The Australian Registrar dismissed the respondent’s opposition to the appellant’s application for registration of the Application Mark in Australia. Briefly, it was held that marks identical to the Application Mark and Respondent’s Mark (Singapore) (“Respondent’s Mark (Australia)”) were not deceptively similar. The word “DIGI” lacked distinctiveness as it was an abbreviation for “digital” and the triangle devices of the marks are “significantly different” (at [14], [21]). Any confusion that might occur would stem from the use of the descriptive term “DIGI”, and not from the deceptive similarity of the parties’ trade marks. The appellant argues that the reasoning set out by the Australian Registrar applies equally in this case.

At the second stage of the step-by-step approach, the appellant submits that the Application Mark’s Specifications: (a) fall within the scope of the Appellant’s Specified Goods and Services; and (b) are no more similar to the specifications of the Respondent’s Mark (Singapore) than the specifications of the Appellant’s Old Mark (Singapore) were. The appellant is thus compliant with the 2002 Agreement and entitled to register the Application Mark.

Alternatively, the appellant submits that the specifications of the Application Mark in Classes 9, 38 and 42 and the Respondent’s Mark (Singapore) in Class 9, are dissimilar. The parties operate in different industries and provide goods and services for different uses and users.

Finally, the appellant refutes the existence of any likelihood of confusion among the relevant public. It argues that the relevant public, OEMs, would apply care and specialist knowledge when purchasing the appellant’s products and would therefore not confuse the parties’ products. In any event, the appellant does not manufacture scales, weighing devices or printers. The appellant also highlights that: (a) the parties’ marks co-existed for 14 years under the 2002 Agreement. From 2002 to 2015, the appellant filed about 425 applications worldwide for “DIGI” related marks; and (b) the Application Mark currently co-exists with identical iterations of the Respondent’s Mark (Singapore) in the Overlapping Territories without opposition from the respondent. The trade mark offices of those territories believed that there would be no likelihood of confusion among the relevant public. The appellant also highlights the absence of any evidence of actual confusion and cites, in support of its overall position, the Australian Registrar’s Decision where it was held that the parties’ trade in different industries (at [52]).

s 8(7)(a) TMA

The appellant challenges the existence of a misrepresentation and damage to the respondent’s goodwill.

Under the element of misrepresentation, the appellant argues that the respondent’s consent to its registration of the Application Mark, pursuant to the 2002 Agreement (see [41] above), means that the appellant is not misrepresenting its goods as being those of the respondent’s. Further, it submits that the specifications of the Application Mark are, in substance, similar to specifications of the Appellant’s DIGI-related Marks (Singapore). The goods and services covered under these earlier marks must have been deemed to be distinct from the specification of the Respondent’s Mark (Singapore). Therefore, allowing the Application Mark to be registered will not result in any likelihood of confusion in the market.

Under the element of damage, the appellant submits that this is not satisfied given the absence of a likelihood of confusion under s 8(2)(b) of the TMA.

s 8(4)(b)(i) read with s 8(4)(a) TMA

Given there is no likelihood of confusion under s 8(2)(b) of the TMA, the elements of: (a) a connection between appellant’s goods or services bearing the Application Mark and the respondent; and (b) likely damage to the interests of the respondent, under s 8(4)(b)(i) TMA are not made out.

s 7(6) TMA

The appellant disputes the allegation that the Application Mark was filed in bad faith. Principally, it takes the view that Clauses 4 and 5 of the 2002 Agreement envision that parties may apply for new registrations of trade marks and that there is no obligation on either party to consult with the other before making changes to their respective marks.

Respondent’s submissions

s 8(9) TMA

The respondent disputes that it consented to the registration of the Application mark under the 2002 Agreement. The 2002 Agreement is inapplicable due to Clause 6 of the 2002 Agreement, which states that the agreement is “… valid for existing trade marks on a world wide basis …”. Seen in this context, Clauses 4 and 5 only permit “new registrations” of each party’s trade marks that were existing at the time of the 2002 Agreement.

s 8(2)(b) TMA

At the first stage of the step-by-step analysis involving marks similarity, the respondent is of the view that the Adjudicator erred in finding that the Respondent’s Mark (Singapore) only possessed a “medium degree of distinctiveness” due to its font design and triangle device. It submits that the Respondent’s Mark (Singapore) possesses a high degree of technical distinctiveness arising from its inherent and, if not, acquired distinctiveness.

In terms of inherent distinctiveness, the respondent relies on the principle that an invented word possesses a high degree of distinctiveness. It argues that “DIGI” has no dictionary meaning and that neither party led evidence proving that it is a short form of the word “digital”. It also highlights that the appellant took the position in the EU Office for Harmonisation in the Internal Market (“EU OHIM”) (Trade Marks and Designs) Opposition Division case of Opposition No. B 819 229 Digi International Inc. v DigiPoS Store Solutions Group Limited (“Digi v DigiPoS (EU)”) that “DIGI” does not mean “digital” and that it does not appear as such in the dictionary. In terms of acquired distinctiveness, the appellant contends that the Adjudicator failed to give due consideration to this latter form of distinctiveness.

The respondent otherwise adopts the Adjudicator’s findings that the Application Mark and Respondent’s Mark (Singapore) are visually similar and aurally identical. However, it submits that the Adjudicator erred in only regarding the marks as being conceptually similar. Instead, the marks are conceptually identical as each will be “remembered as a mark made up of a meaningless word, and a single triangular device.”

The respondent also disputes the relevance of the Australian Registrar’s Decision at this stage of the analysis. It argues that, unlike in Australia, the Respondent’s Mark (Singapore) does not bear an endorsement that the registration of the mark gives no exclusive right to the word “DIGI”. Additionally, the laws governing marks-similarity in Australia and Singapore are not comparable.

At the second stage of the step-by-step approach, the respondent submits that the Adjudicator was right to find that there is similarity between the respondent’s goods in Class 9 and the Application Mark’s Specifications in Classes 9, 38 and 42. At this stage of the analysis, the respondent cites the Australian Registrar’s view that the identical iteration of the Respondent’s Mark (Singapore) in Australia is registered for at least a subset of the Application Mark’s specifications in the Australian registration application. The Australian Registrar reached this view as the respondent’s goods relate predominantly to POS devices. These inevitably consist of computer hardware, the latter forming part of the Application Mark’s specifications in Australia.

s 8(7)(a) TMA

The respondent re-iterates the Adjudicator’s finding that all three elements of passing off are made out.

In respect of misrepresentation, the respondent also highlights its use of the Respondent’s Mark (Singapore) on cloud-based solutions and systems used in the Internet of Things (“IoT”) (“the Respondent’s IoT Goods and Services”) and goods with wireless capabilities (“the Respondent’s Wireless Goods”) which are similar to the Application Mark’s Specifications. This causes the parties’ fields of activity to overlap. The respondent also raises several examples of the appellant’s IoT-related goods and services being marketed in the retail, food, logistics and hospitality industries.

As for the Appellant’s DIGI-related Marks (Singapore), the respondent submits that their registration is irrelevant as the state of the register says nothing of how they came to be on the register and that the register does not necessarily reflect the situation in the marketplace.

The respondent also dismisses the relevancy of prior use of the Appellant’s DIGI-related Marks (Singapore). It argues that the differences between these earlier marks and the Application Mark are not “inconsequential”, such that their prior use is not probative of whether there will be a likelihood of confusion if the appellant trades under the Application Mark. In the alternative, it submits that there is insufficient evidence to prove co-existence of the Appellant’s DIGI-related Marks (Singapore) with the Respondent’s Mark (Singapore) (eg, sales and advertising figures in Singapore of the appellant’s products and services bearing these earlier marks).

The respondent submits that the insufficiency of evidence on prior use of the Appellant’s DIGI-related Marks (Singapore) in Singapore is crucial in another respect. It precludes the appellant from asserting independent goodwill in Singapore in the broad range of IoT goods and services it now seeks registration for. If at all, the appellant’s independent goodwill exists in respect of a narrow range of goods only. The appellant therefore cannot encroach into the range of the Respondent’s IoT Goods and Services.

s 8(4)(b)(i) read with s 8(4)(a) TMA

The respondent agrees with the Adjudicator’s findings under this ground of opposition, save for its disagreements over the Adjudicator’s marks similarity analysis highlighted at [56]–[58] above.

s 7(6) TMA

The respondent argues that the Application Mark was filed contrary to the spirit of the 2002 Agreement. This is because: (a) the Application Mark is more similar to the Respondent’s Mark (Singapore) than the Appellant’s Old Mark (Singapore); and (b) the appellant is applying for registration in respect of specifications beyond the limits permitted by the 2002 Agreement. As such, the appellant “should have known” that the respondent would likely have objected to the Application Mark and “should have broached the change with the [r]espondent with the view to updating the 2002 Agreement to avoid any ambiguity”.

Bad faith is also evidenced by the appellant’s Class 9 Amendment, which was done “late”, and the appellant began trading in contravention of this amendment. In this regard, the respondent points to the appellant’s computer hardware and software relating to weighing and/or measuring purposes or transfer of data connected to weighing and/or measuring purposes.

Issues to be determined

In light of the foregoing, the issues arising for my determination are as follows:

Whether the Application Mark should be registered under s 8(9) of the TMA?

s 8(9) of the TMA provides that “[t]he Registrar may, in his discretion, register a trade mark where the proprietor of the earlier trade mark or other earlier right consents to the registration.” The point of contention under this ground of opposition is the scope of applicability of the 2002 Agreement. Namely, does the 2002 Agreement govern the registration and/or use of trade marks which came into existence after the 2002 Agreement was formed? This is a matter of contractual interpretation.

To recapitulate, the respondent argues in favour of reading Clauses 4 and 5 of the 2002 Agreement, which permit “new registrations” of each party’s marks, together with Clause 6 of the 2002 Agreement. Clause 6 provides that the agreement is “… valid for existing trade marks on a world wide basis …”. In contrast, the appellant argues that the 2002 Agreement must be read in the context of the parties subsequent conduct, viz, entering into 2003 Canadian Settlement Agreement. The latter agreement, which is intended to implement the 2002 Agreement, permits each party to register close variations of the defined “DIGI” marks in Canada. As such, the 2002 Agreement must similarly cover registrations and use of close variations of the parties’ “DIGI” marks defined therein.

Preliminarily, I note that the appellant’s position is that the governing law of the 2002 Agreement is German law. The respondent has not disputed this. In my view, Clause 9 of the 2002 Agreement confirms the appellant’s position. Clause 9 states that “[t]he German District Court Landgericht München I, ruling according German Law [sic], is solely competent for all disputes arising from this conflict.” As contractual interpretation is governed by the proper law of the contract, the scope of the 2002 Agreement is, technically, an issue of German law.

However, the appellant has not led evidence of German law relating to contractual interpretation to support its argument that the 2003 Canadian Settlement Agreement is relevant to interpreting the scope of the 2002 Agreement. In such a case, the court may rely on the presumption of similarity – if there is no or insufficient proof of the foreign law, the court assumes that the foreign law is the same as the local law (OMG Holdings Pte Ltd v Pos Ad Sdn Bhd [2012] 4 SLR 201 at [38]). In accordance with this presumption, I will consider Singapore’s principles on contractual interpretation to determine the relevance, if any, of the 2003 Canadian Settlement Agreement to interpreting the scope of the 2002 Agreement.

The Court of Appeal in Zurich Insurance (Singapore) Pte Ltd v B-Gold Interior Design & Construction Pte Ltd [2008] 3 SLR(R) 1029 (“Zurich Insurance”) at [132(d)] stated that there “should be no absolute or rigid prohibition against evidence of … subsequent conduct, although, in the normal case, such evidence is likely to be inadmissible … ”. The Court of Appeal subsequently confirmed in MCH International Pte Ltd and others v YG Group Pte Ltd and others and other appeals [2019] 2 SLR 837 (“MCH International”) at [20] that it has not expressed a definitive view as to the admissibility of parties’ subsequent conduct for contractual interpretation. Nevertheless, in MCH International at [18]–[19], the Court of Appeal identified some “provisional parameters” [emphasis in original in bold italics] for when evidence of subsequent conduct is admissible. Namely, the court must bear the following criteria in mind:

(a) the subsequent conduct must be relevant, reasonably available to all the contracting parties, and relate to a clear and obvious context;

(b) the principle of objectively ascertaining contractual intention(s) remains paramount; and accordingly,

(c) the subsequent conduct must always go toward proof of what the parties, from an objective viewpoint, ultimately agreed upon.

In view of these provisional parameters, I find it difficult to accept that the 2003 Canadian Settlement Agreement is at all relevant to what the parties intended the scope of the 2002 Agreement to be at the time it was concluded. Accordingly, I did not consider it for the purposes of interpreting the scope of the 2002 Agreement.

First, Clause 7 of the 2002 Agreement states that the contract “constitutes the entire agreement between the parties with respect to the subject matter hereof”. Since the parties intended to embody their entire agreement in the written contract, the terms of the 2002 Agreement constitute prima facie proof of the parties’ intentions and extrinsic evidence is inadmissible to contradict, vary, add to or subtract from these terms (Zurich Insurance at [132(b)]; s 93 of the Evidence Act (Cap 97, 1997 Rev Ed)).

However, the appellant seeks to vary the terms of the 2002 Agreement by relying on the 2003 Canadian Settlement Agreement. Clause 6 of the 2002 Agreement makes clear that: (a) the 2002 Agreement only applies to trade marks existing at the time the agreement was concluded; and (b) the 2002 Agreement applies on a worldwide basis. The first portion of Clause 6, stating that the 2002 Agreement applies to existing trade marks, qualifies the scope of Clause 5, which reads:

Clause 5 of the 2002 Agreement therefore applies to a narrower class of trade marks than Clause 9.0 of the 2003 Canadian Settlement Agreement. The latter clause extends to “close variations” of the “DIGI” related marks identified therein and is not limited to marks in existence at the time the 2003 Canadian Settlement Agreement was concluded. Clause 9.0 reads as follows:

In essence, the respondent is not entitled to rely on extrinsic evidence, viz, the 2003 Canadian Settlement Agreement, to extend the ambit of the 2002 Agreement.

Second, the appellant does not dispute that the 2003 Canadian Settlement Agreement only applies in Canada. This must be correct, since Clause 12.0 states that:

In contrast, the 2002 Agreement applies on a worldwide basis by virtue of Clause 6. Therefore, while the 2002 Agreement only applies to trade marks in existence at the time the agreement was concluded, it protects the parties’ rights to register these trade marks in territories where registration had not been obtained as at the date the agreement was concluded. Whereas, on an objective view, 2003 Canadian Settlement Agreement was intended to expand the scope of non-interference in the parties’ new registrations, new applications, renewals and/or use of “DIGI” related marks in Canada, without disturbing the scope of the 2002 Agreement in other parts of the world. The very fact the 2003 Canadian Settlement Agreement was needed to extend the parties’ rights to “close variations” of marks existing in Canada shows that the scope of the 2002 Agreement is more limited.

For these reasons, this is a case where the provisional parameters laid down in MCH International have not even been met. I did not consider the 2003 Canadian Settlement Agreement when interpreting the scope of the 2002 Agreement. Instead, reading Clauses 5 and 6 of the 2002 Agreement together, I accept the respondent’s position that the 2002 Agreement does not govern the registration of marks that have come into existence after the 2002 Agreement was concluded. In other words, the 2002 Agreement is irrelevant to the present application to register the Application Mark since this mark was launched only in 2015 (see [13] above). The 2002 Agreement therefore does not represent the respondent’s consent to the Application Mark’s registration under s 8(9) of the TMA. I decline to register the Application Mark under this provision.

Whether the ground of opposition under s 8(2)(b) of the TMA is made out?

It is trite law that under s 8(2)(b) of the TMA, courts will apply the step-by-step approach endorsed by the Court of Appeal in Staywell at [15]:

Are the marks similar?

The first stage of the step-by-step approach concerns the question of whether the marks are similar. The appellant’s argument that the respondent tacitly accepts the co-existence of the two marks, under the 2002 Agreement, if the Application Mark is not more similar to the Respondent’s Mark (Singapore) than the Appellant’s Old Mark (Singapore) is a non-starter. This follows from my finding on the inapplicability of the 2002 Agreement (see [83] above). The appellant’s alternative case may be summarised as follows: the Respondent’s Mark (Singapore) lacks technical distinctiveness and the differences between the Application Mark and Respondent’s Mark (Singapore) are sufficient to render them dissimilar (see [46] above).

Before I begin my analysis, I set out some key principles relevant to determining the similarity of competing marks. While it is not a mechanistic approach, the court compares three aspects of the marks: their visual, aural and conceptual similarity. When doing so, the court should consider the marks as a whole, without taking into account any external added matter or circumstances because the comparison is mark for mark. The court assumes the viewpoint of an average consumer, with imperfect recollection, who would exercise some care and a measure of good sense in making his or her purchases, not that of an unthinking person in a hurry (Hai Tong Co (Pte) Ltd v Ventree Singapore Pte Ltd and another and another appeal [2013] 2 SLR 941 (“Hai Tong”) at [40]).

The question of whether the Respondent’s Mark (Singapore) is distinctive, in both the technical and non-technical sense, is integrated into the analysis of the visual, aural and conceptual similarity of the marks (Staywell at [30]). In the non-technical sense, distinctiveness refers to “what is outstanding and memorable about the mark”. Such components tend to draw the average consumer’s attention and stand out in the average consumer’s imperfect recollection (Staywell at [23]). Technical distinctiveness refers to the capacity of a mark to function as a badge of origin and it may arise in two forms. Inherent distinctiveness is “usually where the words comprising the mark are meaningless and can say nothing about the goods or services” (Staywell at [24]). Acquired technical distinctiveness is “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” (Staywell at [24]). Technical distinctiveness must be assessed by looking at the mark as a whole. For instance, while the components of a mark may not be inherently distinctive, the sum of its parts may have sufficient technical distinctiveness (Staywell at [25]). Importantly, a mark which has greater technical distinctiveness enjoys a high threshold before a competing sign will be considered dissimilar to it.

Technical distinctiveness of the Respondent’s Mark (Singapore)

I am mindful that determining the technical distinctiveness of the opponent’s mark is not a separate step within the marks-similarity analysis. However, given that distinctiveness features in all aspects of the mark-similarity analysis, as a matter of practicality, I shall begin with this question before proceeding to compare the competing marks visually, aurally and conceptually.

For ease of reference, I set out the Application Mark and Respondent’s Mark (Singapore):

The respondent contends that the Respondent’s Mark (Singapore) possesses both forms of technical distinctiveness. I will consider each in turn.

(1) Inherent Distinctiveness

In terms of inherent distinctiveness, the Adjudicator found that the word “DIGI” is not particularly inherently distinctive. However, the composite whole of the Respondent’s Mark (Singapore), viz, the stylised manner of the text and the triangle device accompanying it, possessed a medium degree of distinctiveness. While not explicit, the Adjudicator appears to have had inherent distinctiveness in mind when commenting on the distinctiveness of the whole of the mark (see [20] above).

The respondent disagrees with the Adjudicator and argues that the word “DIGI” is invented and has no dictionary meaning. It also highlights the irony of the appellant taking this very position in Digi v DigiPoS (EU), where the appellant relied on an identical iteration of the Appellant’s Old Mark (Singapore) to oppose the registration of the trade mark “DigiPos” by an unrelated third party (see [57] above).

Initially, the appellant did not argue that the word “DIGI” is descriptive. In its written submissions of 6 January 2021, it merely argued that the competing marks’ word component should be discounted, and only their triangle devices be compared, as the effect of the 2002 Agreement is that parties agreed that each of them could use “DIGI” in their respective trade marks. The appellant’s Counter-Statement, statutory declaration of Mr Steve Mauer (“Mr Mauer”; the Senior Counsel – Intellectual Property & Commercial Transactions of the appellant) and written submissions before the Adjudicator also do not allege that the word “DIGI” is descriptive. For the first time in its further submission of 15 March 2021, the appellant submits that “DIGI” is descriptive as it is an abbreviation for “digital”, citing the Australian Registrar’s Decision of 2 February 2021 in support (see [46] above).

Having considered the parties’ submissions, I accept that the textual component of the Respondent’s Mark (Singapore), “DIGI”, is descriptive. Technically, “DIGI” has no dictionary meaning. However, I agree with the Adjudicator that “DIGI” would intuitively be understood to denote the digital nature of the goods in the Class 9 specification of the Respondent’s Mark (Singapore). This is especially so given the average consumer in Singapore has been described as someone who would exercise some care and a measure of good sense in making his or her purchase, not that of an unthinking person in a hurry (Hai Tong at [40(c)]-). The EU OHIM in Digi v DigiPoS (EU) similarly opined that “…in the Community in general, ‘DIGI’ would normally be considered a weaker element due to its relationship with ‘DIGITAL’ …” The EU OHIM’s view in this regard was upheld by the Boards of Appeal in DigiPoS Store Solutions Group Limited v Digi International Inc. (Case R 127/2010-1) (“DigiPoS v Digi (EU) appeal”) at [48] and [64].

No definitive position has emerged in the UK as to whether the word “DIGI” is inherently distinctive. In Digipos Store Solutions Group Ltd v Digi International Inc [2008] Bus LR 1621 (“Digipos v DIGI (UK)”), the Hearing Officer for the Registrar of Trade Marks in the UK (“the Hearing Officer”) and the English High Court reached differing views. However, the English High Court did not overrule the Hearing Officer as the latter’s decision was not “plainly wrong”. The Hearing Officer found that “DIGI” was inherently distinctive. While the Hearing Officer was referred to an acronym website where “DIGI” was defined as meaning “digital”, he concluded that the average consumer would still “wonder whether this is the inference that is being alluded to [as] they would not be sure and would see it as a made up term” (at [45]). On appeal, Daniel Alexander QC (sitting as a deputy judge of the English High Court) took the opposing view for the following reasons (at [92]):

As explained above, the English High Court refrained from disturbing the Hearing Officer’s decision as the latter’s decision was not “plainly wrong” and might be said to have been supported by countervailing matters. These matters included the absence of evidence of any established definition of “DIGI” in mainstream dictionaries and the lack of evidence as to how the mark “DigiPos” was actually understood by members of the trade and public (at [93]).

Notwithstanding the English High Court’s refusal to overrule the Hearing Officer, I stand by my view that “DIGI” is descriptive in the context of the respondent’s Class 9 specification.

However, this is not the end of the inquiry pertaining to inherent distinctiveness. Technical distinctiveness must be assessed by looking at the Respondent’s Mark (Singapore) as a whole. As the Court of Appeal in Staywell observed at [25], while the components of a mark may not be inherently distinctive, the sum of its parts may have sufficient technical distinctiveness (see [87] above).

This leads me to my second point – that the Adjudicator was correct to go on to consider whether the composite whole of the textual and triangle device components of the Respondent’s Mark (Singapore) possess inherent distinctiveness.

The triangle device in the Respondent’s Mark (Singapore) is an inverted triangle, coloured black (for the purposes of registration), with white stripes that become increasingly spaced out towards the top. It is not descriptive or allusive of the goods covered by the specification of the Respondent’s Mark (Singapore). The word “DIGI” is also stylised in the manner described at [124] below. However, I do not think the stylisation of the triangle device and word component is so elaborate or inventive as to confer the Respondent’s Mark (Singapore) substantial inherent distinctiveness. Accordingly, viewing the triangle device and stylisation of the word “DIGI” in the Respondent’s Mark (Singapore) in totality, I find that the composite whole of the mark has a low degree of inherent indistinctiveness.

For completeness, while arguing that the Application Mark is not more similar to the Respondent’s Mark (Singapore) than the Appellant’s Old Mark (Singapore), the appellant states the respondent “does not claim exclusivity over such a [triangle] device. There are other marks on Singapore’s register with devices of a similar shape.” The examples provided in Mr Mauer’s first affidavit of 1 June 2020 (“Mr Mauer’s First Affidavit”) are as follows:

However, the fact that other marks on the register contain triangle devices does not diminish the inherent distinctiveness of the composite whole of the Respondent’s Mark (Singapore). The inherent distinctiveness of the Respondent’s Mark (Singapore) lies in the combination and stylisation of the word “DIGI” and a triangle device. None of the examples cited by Mr Mauer comprise of these precise elements. Further, even if the appellant is right in saying that the respondent does not claim exclusivity over the triangle device, I fail to see how the average consumer would know this. Such failure to claim exclusivity in the triangle device does not diminish the capacity of the totality of the Respondent’s Mark (Singapore) to function as a badge of origin. In fact, that the appellant now focuses on distinguishing the triangle devices in the competing marks (see [123] below) betrays a recognition that the triangle device in the Respondent’s Mark (Singapore) adds to the mark’s technical distinctiveness. As such, I re-iterate my conclusion that the sum total of the word and device component of the Respondent’s Mark (Singapore) bears a low degree of inherent distinctiveness.

(2) Acquired Distinctiveness

I next consider whether the Respondent’s Mark (Singapore) has acquired distinctiveness.

The Court of Appeal in Hai Tong endorsed the following factors as being helpful, but inexhaustive, indicia in determining whether a mark has acquired distinctiveness through use (at [33]):

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

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

(c) the amount invested in promoting the mark;

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

(e) any statements from trade associations and chambers.

The respondent submits that the Respondent’s Mark (Singapore) has acquired distinctiveness through the extensive use and promotion of it in Singapore since 1987. It provides the following evidence in support. In 1987, Singapore became the location of the respondent’s first overseas production site. “DIGI” products were produced in Singapore where the factory occupied 12,000 square metres and employed over 100 employees. Since 1987, the Respondent’s Singapore Subsidiary has continuously and extensively used “DIGI” and the Respondent’s Mark (Singapore) in respect of, inter alia, balances and scales, scales equipped with printer, printers, weighing and measuring systems in general and other such products in Class 9 in Singapore. In Singapore, the respondent’s goods bearing the Respondent’s Mark (Singapore) have, for numerous years, been supplied to local supermarkets including NTUC, FairPrice, Cold Storage and Sheng Siong, and local companies such as Singapore Post Pte Ltd, Toshiba Chemical (S) Pte Ltd, Citizen Farm and Sora. As a rough estimate, the respondent’s goods bearing “DIGI” or the Respondent’s Mark (Singapore) account for 60% of the market share for scales and commercial weighing and measuring systems in Singapore. The Respondent’s Singapore Subsidiary’s sales revenue figures increased from $420,925 in 2007 to $2,964,185.35 in 2017. The respondent also claims that its evidence demonstrates extensive promotions and advertisements in Singapore over the years. The Respondent’s Mark (Singapore) has been placed in, for instance, directories such as Green Book and Yellow Pages. Its annual advertising expenditure also increased from $2,230 in 2013 to $11,609.68 in 2017.

While the Adjudicator mentioned the above evidence at [43] of her GD, she did not make a finding of acquired distinctiveness when performing the marks similarity analysis under s 8(2)(b) of the TMA. However, when analysing the tort of passing off under s 8(7)(a) of the TMA, the Adjudicator held at [73] that “it is clear that the threshold query has been met and the goodwill is distinctive and sufficiently associated with the [respondent’s] goods and services using the [Respondent’s Mark (Singapore)]”. The threshold query the Adjudicator must have had in mind is proving that the mark (of the opponent on the passing off ground) has acquired distinctiveness through use. This is because the Court of Appeal in Singsung Pte Ltd v LG 26 Electronics Pte Ltd (trading as L S Electrical Trading) [2016] 4 SLR 86 (“Singsung”) at [38] and [70] held that the issue of acquired distinctiveness is best understood as a threshold inquiry in the context of determining whether an actionable misrepresentation has been committed. Based on the evidence above at [105], the Adjudicator further held that the Respondent’s Mark (Singapore) was well known in Singapore under s 8(4)(a) of the TMA. She also noted at [82] of the GD that the appellant “did not refute all of [the evidence at [105] above] but submitted that the evidence was confined only to scales and measuring equipment and there was no evidence of use of services in relation to Classes 38 and 42.”

Even on appeal, the appellant does not challenge the finding that the Respondent’s Mark (Singapore) has acquired distinctiveness (in connection with the element of misrepresentation in passing off) or that the Respondent’s Mark (Singapore) is well known in Singapore. I note that the Australian Registrar’s Decision, which the appellant relies on to argue that “DIGI” is descriptive, went on to say at [19] that there was “… no basis to find that the word ‘digi’ has become more distinctive since [the time of registration of the Respondent’s Mark (Australia)]”. However, I do not regard the Australian Registrar’s Decision to be persuasive, in general, in the matter before me. Namely, the material statutory provisions discussed in the Australian Registrar’s Decision, such as ss 44 and 60 of Australia’s Trade Marks Act 1995 (Cth) (“the Australian TMA”), differ from comparable provisions in Singapore. George Wei J stated a similar view in Rovio at [122]–[124], in connection with ss 44 and 60 in the Australian TMA. In particular, he observed that: (a) the statutory language in s 44 of the Australian TMA regarding “deceptive similarity” reduced the persuasive weight of authorities discussing this provision in Singapore (at [122]); and (b) that the provisions in Singapore on well known earlier trade marks are not the same as the Australian provisions (at [124]). Accordingly, the Australian Registrar’s Decision does not advance the appellant’s case.

Having regard to the factors identified in Hai Tong (see [104] above), I uphold the Adjudicator’s finding that the Respondent’s Mark (Singapore) has acquired distinctiveness.

First, with regards the market share held by the Respondent’s Mark (Singapore), Mr Heng Yong Chiang’s (“Mr Heng”) Statutory Declaration filed in the Registry states that “[a]s a rough estimate, the [respondent’s] DIGI products account for approximately 60% of the market share for scales and commercial weighing and measuring systems”. Mr Heng is a director of the Respondent’s Singapore Subsidiary. The appellant has not disputed this claim in Mr Heng’s statutory declaration before the Adjudicator or this court.

I accept that Mr Heng provided evidence of the sales and advertising figures of the Respondent’s Singapore Subsidiary increasing and identified local customers such as major supermarkets in Singapore. Nevertheless, this does not prove the extent of the market share of the Respondent’s Singapore Subsidiary described in the preceding paragraph. In the light of these considerations, whilst the respondent’s estimated market share is still a factor going in its favour, I am unable to place significant weight on it.

Second, with regards the nature of the use of the Respondent’s Mark (Singapore), I am satisfied that the it has been used extensively in connection with the respondent’s goods in Singapore.

In terms of marketing and advertisement, the respondent’s product brochures feature an array of weighing apparatus, printers and a combination of the two bearing the Respondent’s Mark (Singapore). These brochures also name the Respondent’s Singapore Subsidiary and provide its contact details in Singapore. The Respondent’s Mark (Singapore) was also displayed in connection with the respondent’s goods in, inter alia, the Green Book 2012–2013 and on the Green Book website, albeit the latter being for an unknown period of time.

In terms of actual use of the Respondent’s Mark (Singapore) on goods, the Respondent’s Singapore Subsidiary has used this mark on balances and scales, scales equipped with a printer, printers, weighing and measuring systems in general and other such products in Class 9 in Singapore since 1987 (ie, for over three decades). These goods bearing the Respondent’s Mark (Singapore) are used in various supermarkets in Singapore. The Respondent’s Singapore Subsidiary’s revenue figures have grown from 2007–2017 and, since 2012, have been consistently upwards of $1 million. The Respondent’s Singapore Subsidiary hit $2,964,185.35 in revenue in 2017. In addition, the Respondent’s Singapore Subsidiary supplies goods to buyers in the Asian and world markets, including in Germany and Japan. The invoices issued to these foreign buyers by the Respondent’s Singapore Subsidiary bear the Respondent’s Mark (Singapore).

Moreover, a portion of the further evidence of the respondent, admitted in HC/SUM 2178/2020 (“the SUM 2178/2020 evidence”), is relevant to the determination of the Respondent’s Mark’s (Singapore) acquired distinctiveness. While I do not rely on portions of the SUM 2178/2020 evidence which pertain to goods and services falling outside of the Respondent’s Mark’s (Singapore) Class 9 specification, there are certain goods disclosed in the SUM 2178/2020 evidence that fall within the Respondent’s Mark’s (Singapore) Class 9 specification, and which are highlighted as examples of the Respondent’s Wireless Goods (ie, goods which incorporate wireless capabilities). Two examples are:

(a) The SM-5300, which is a PC based weigh-labeller with inbuilt wireless modules and wireless connectivity to other products like InfoTag and e.Label Electronic Signage. The respondent’s website states that through the SM-5300, “[revised] … promotional text, product information and pricing updates can be instantaneously displayed on InfoTag via the wireless base station” [emphasis in original in bold]; and

(b) The SM-500V2, which is a scale printer that prints receipts or labels. Notably, this product has wireless network connectivity and its data can be centralised under the scale application software, e.Leap, and sales reports may be generated for management assessment.

Sales of SM-5300 and SM-500V2 in Singapore are evidenced in records from the Respondent’s Singapore Subsidiary’s accounting system exhibited in Mr Heng’s statutory declaration. In my view, sales of these goods, which fall within the Respondent’s Mark’s (Singapore) Class 9 specification, in Singapore increases the extent of the mark’s actual use here and strengthens the case for acquired distinctiveness. By way of an aside, a fuller discussion of the relevance of the SUM 2178/2020 evidence, more generally, to the ground of opposition in s 8(2)(b) TMA may be found at [142]–[144] below.

Third, with regards the amount invested in promoting the Respondent’s Mark (Singapore), the Respondent’s Singapore Subsidiary’s annual advertising has risen from $2,230 in 2013 to $11,609.68 in 2017.

Evidence relating to the Hai Tong factors at [104(d)] and [104(e)] has not been adduced. I recognise that a well-designed market survey is likely to be useful when seeking to establish the acquired distinctiveness of a highly descriptive and/or inherently non-distinctive mark, a slogan, a colour or shape mark. For such marks, the market survey may show that the relevant public has been educated that the sign is and functions as a trade mark. However, considering the mark in issue here and the evidence showing use of the mark as a trade mark, I do not consider the absence of a market survey to be fatal.

In light of the long-standing and extensive use of the Respondent’s Mark (Singapore) in Singapore, both in promotional material and on actual goods sold, the mark possesses acquired distinctiveness. Therefore, while the word “DIGI” is descriptive and the whole of the Respondent’s Mark (Singapore) only has a low level of inherent distinctiveness, the distinctiveness of the whole mark has been enhanced to a moderate degree through use.

While there may be some marks which can never possess distinctiveness (inherent or acquired) (eg, “Soap” for “soap” (Jacob J in British Sugar Plc v James Robertson & Sons Ltd [1996] RPC 281 (“British Sugar”) at 305)), the Respondent’s Mark (Singapore) is not one such example. To the contrary, viewing the composite whole of the Respondent’s Mark (Singapore), which includes the stylised word “DIGI” and a triangle device (see [100] above), and on account of its use in Singapore, the whole mark functions as a badge of origin from the perspective of consumers.

Visual similarity

I now turn to comparing the marks visually.

The Adjudicator held that the competing marks are visually similar “in terms of the textual presentation and use of similarly shaped device” (GD at [48]). She noted that the appellant owned many other “DIGI” related marks but none used a similar sans serif font to that in the Respondent’s Mark (Singapore) (GD at [47]). She also observed that there were “many other shapes that could have been used, instead of a triangular shaped device” (GD at [48]).

The respondent agrees with the Adjudicator’s reasoning, and stresses that the competing marks are both “composite marks, each made up of a single [identical distinctive] word and device”. While it acknowledges differences in the font of the marks’ word component and the design and placement of the triangle devices, it submits these differences are “simply insufficient to distinguish the marks”.

In this regard, the appellant initially accepted that the parties’ previous marks are similar, and that the Application Mark is “not more similar” to the Respondent’s Mark (Singapore) than the Appellant’s Old Mark (Singapore). In fact, Mr Mauer’s precise words are that the parties’ respective “DIGI” marks in the 2002 Agreement are “essentially identical” [emphasis added] – inter alia, they are visually similar as the most distinctive component of their marks was the word “DIGI”. However, in its further submissions, the appellant now adopts the Australian Registrar’s finding that the competing marks are not deceptively similar as the triangle devices are “significantly different” in their nature and placement. The device in the Application Mark is a small green triangle on the right of the word “DIGI”. In contrast, the device in the Respondent’s Mark (Singapore) is a large inverted triangle with graduated lines, next to a smaller word mark “DIGI”.

I find it significant that both the competing marks comprise the word “DIGI” and a triangle device. With regards the Respondent’s Mark (Singapore), given the triangle’s size relative to the word component and its graduated lines, the triangle is visually co-dominant with the word component. As for the Application Mark, the colour contrast between the light green triangle and the black letters would call attention to both components. Visual, aural and conceptual similarity of marks is based on the overall impression given by the marks (Staywell at [26]). As the competing marks contain the same elements, both of which are eye-catching in the context of each mark, I have no doubt that the impression left by each mark is highly similar. Even further still, the stylisation of the marks’ word components are also highly similar:

(a) All letters are capitalised;

(b) The letter “D” in both marks has a gap, although the gap in the Application Mark is at the top of the “D”, while the gap is at the bottom of the “D” in the Respondent’s Mark (Singapore); and

(c) Both marks employ a similar font which is plain and minimalist. In fact, in its written submissions, the appellant accepts that the Application Mark and Respondent’s Mark (Singapore) employ the “sans serif font”. Although there are minor stylistic differences between the fonts (including the location of the gap in the “D”), these will not register in the imperfect recollection of the average consumer.

Having considered the appellant’s submissions, I am not satisfied that the differences in the nature and positioning of the two triangle devices distinguish the marks visually. Keeping in mind that competing marks must not be compared side by side (Hai Tong at [40(d)]), and that the average consumer has imperfect recollection, these differences are insufficient to distinguish the competing marks in the mind of the average consumer. It also bears emphasising that the Respondent’s Mark (Singapore) has inherent and acquired distinctiveness. All the more, the differences in respect of the triangle devices are not so substantial as to distinguish the Application Mark visually. However, these differences are also not completely irrelevant. On the whole, I characterise the degree of visual similarity between the marks as fair.

Aural Similarity

The Adjudicator held that the marks are aurally identical as both comprise the letters “DIGI” and are pronounced in the same manner by both parties (GD at [40]).

I agree with the Adjudicator’s finding. “DIGI” is the only textual component in both marks and the appellant accepted, in the hearing before me, that both marks would be referred to as “DIGI”. Absent any suggestion that “DIGI” is pronounced differently in the Application Mark, the marks are aurally identical.

Conceptual Similarity

In considering whether there is conceptual similarity, the inquiry is directed at the ideas that lie behind or inform the marks or signs in question (Hai Tong at [70]). In this analysis, the court should consider 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 (Staywell at [35]).

The Adjudicator found that the marks have a high degree of conceptual similarity. In both marks, the text “DIGI” evokes the word “digital” and connotes a relation to products and services with a digital element. The respondent submits that the marks are conceptually identical as both will be remembered as marks made up of a meaningless word and a single triangle device. Alternatively, if “DIGI” means “digital”, the respondent submits that since the word “DIGI” is common to both marks, they are conceptually identical.

It follows from my finding that “DIGI” is a short form for “digital” (see [94] above) that the respondent’s first argument fails.

In my view, the Adjudicator was right to find that both marks evoke the notion of digital goods and services (GD at [49]), or technology more generally. Clearly, the marks are, at least, conceptually similar to a high degree.

The respondent goes further to say that both marks are conceptually identical. This presupposes that the conceptually dominant component of both marks is the textual component only. However, for the Respondent’s Mark (Singapore), I take the view that the device and textual components are visually and conceptually co-dominant. The triangle device is larger and “looms” over the textual component. The triangle device leaves as strong an impression in the mind of an average consumer as the textual component. As such, the idea connoted by the Respondent’s Mark (Singapore) would include both technology and the notion of a triangle. That being said, in respect of the Application Mark, the notion of a triangle is likewise evoked by its triangle device component. However, this notion will not be identical to that connected to the Respondent’s Mark (Singapore) as the two triangle devices are not the same.

The competing marks are thus not conceptually identical, but highly similar.

Conclusion on marks similarity

Based on the foregoing analysis, I find that the marks are aurally identical, conceptually similar to a high degree and visually similar to a fair degree. On the whole, I have no trouble finding that the marks are similar.

Are the marks’ specifications similar?

At the second stage of the step-by-step approach, the court considers whether the goods and services in the specifications of the Application Mark and Respondent’s Mark (Singapore) are similar.

As I have rejected the applicability of the 2002 Agreement, I need not consider the appellant’s argument that the Application Mark’s Specifications fall within the defined scope of goods and services in the 2002 Agreement. I therefore focus on whether the goods and services in the specifications of the competing marks are similar.

The appropriate comparison is between the specification for which the Respondent’s Mark (Singapore) is registered and the specifications for which the Application Mark is sought to be registered (Staywell at [40]). The Court of Appeal in Hai Tong also confirmed (at [92]) that the following factors set out by Jacob J in British Sugar at 296 (“the British Sugar factors”) are relevant to determining the question of the similarity of goods and services:

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

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

(c) The physical nature of the goods or services;

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

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

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

Further, each class of specification under which the Application Mark is sought to be registered must be compared individually to the Class 9 specification of the Respondent’s Mark (Singapore) (Staywell at [43]).

The appellant submits that the specifications of the Application Mark in Classes 9, 38 and 42 and the specification of the Respondent’s Mark (Singapore) in Class 9 are dissimilar. This is because the appellant’s goods and services and the respondent’s goods are for different uses and users and do not compete. The appellant’s products pertain to the “connection of machines” and provide the infrastructure which makes possible and facilitates connection to the Internet. Many of its goods are sold to government bodies or OEMs to be incorporated into other products. In contrast, the respondent’s goods are confined to scales and measuring equipment, labellers, label printers and sorting machines and systems. The respondent’s goods are sold to the “general public as complete end users systems”, and cater to four main business fields: retail, food, logistics and hospitality. The appellant also argues that the respondent’s SUM 2178/2020 evidence, showing the Respondent’s IoT Goods and Services and Wireless Goods, is irrelevant. These goods and services are outside the scope of the Respondent’s Mark’s (Singapore) Class 9 specification. It also submitted that the Application Mark’s Specifications are, in substance, similar to the specifications allowed under the Appellant’s DIGI-related Marks (Singapore). As the latter group of marks co-exist with the Respondent’s Mark (Singapore) without any issue, this demonstrates that the goods and services covered under the Appellant’s DIGI-related Marks (Singapore) are distinct from the specification of the Respondent’s Mark (Singapore).

In response, the respondent agrees with the Adjudicator’s finding that there is similarity of goods and services (see [23]–[24] above for Adjudicator’s reasoning). In gist, the respondent submits that the appellant’s goods and services are complementary to its own goods, are sold by the same producer/providers to the same users and are commonly found in the same distribution channels. In particular, with regards the Application Mark’s Specifications in Classes 38 and 42, the respondent emphasises that these services may be used in connection with goods in the Class 9 specification of the Respondent’s Mark (Singapore) (ie, balance and scale, scale equipped with printer, printer). Classes 38 and 42 of the Application Mark’s Specifications do not exclude services which concern weighing and/or measuring purposes or transfer of data connected to weighing and/or measuring processes (see [10(c)] above). There are few devices that exist without some form of wireless connectivity. The services that allow goods to perform wirelessly are also sold as integrated systems and provided by the same provider together with the goods in question. It follows that the services covered in Classes 38 and 42 of the Application Mark’s specifications and the respondent’s goods in Class 9 would be obtained through a single purchase of a particular good that integrates such wireless technology. The respondent also submits that the SUM 2178/2020 evidence is relevant under s 8(2)(b) of the TMA as the assessment of similarity of goods and services must take into account actual and notional fair uses of the specification of the Respondent’s Mark (Singapore). Finally, the respondent submits that the specifications of the Appellant’s DIGI-related Marks (Singapore) are irrelevant to the inquiry as the state of the register says nothing of how the marks were put on to the register.

Preliminarily, I deal with two points raised by the appellant.

The first point is the question of whether the SUM 2178/2020 evidence is relevant to the comparison of the competing marks’ specifications. At this stage, and in the context of the step-by-step approach more generally, I am confined to comparing the specification for which the Respondent’s Mark (Singapore) is registered and the specifications for which the Application Mark is sought to be registered (Staywell at [40]). The respondent has conceded in Mr Youichi Kishi’s (“Mr Kishi”) first affidavit of 3 June 2020 (“Mr Kishi’s First Affidavit”) that this further evidence pertains to goods and services beyond the scope of the Respondent’s Mark’s (Singapore) specification. Mr Kishi is the Business Unit Manager of the Intellectual Property and Technical Standard Department of the respondent. For these reasons, I am unable to take the SUM 2178/2020 evidence into account under s 8(2)(b) of the TMA insofar as it relates to goods and services beyond the Respondent’s Mark’s (Singapore) Class 9 specification. However, as explained at [114] above, there are certain goods disclosed in the SUM 2178/2020 evidence that fall within the Respondent’s Mark’s (Singapore) Class 9 specification, and which are highlighted as examples of the Respondent’s Wireless Goods (ie, goods which incorporate wireless capabilities). Two examples are the SM-5300 (a PC based weigh-labeller with inbuilt wireless modules and wireless connectivity to other products) and the SM-500V2 (a scale printer which has wireless network connectivity).

In my view, these instances of the respondent’s goods which fall within the Respondent’s Mark’s (Singapore) Class 9 specification are relevant to my comparison of the competing marks’ specifications (see [163] and [167] below).

In contrast, all of the SUM 2178/2020 evidence remains relevant under ss 8(7)(a) (ie, passing off) and 8(4)(b) (ie, well known trade mark) of the TMA. In the former ground of opposition, the court can consider actual use of the Respondent’s Mark (Singapore) when determining, inter alia, whether the respondent has goodwill in Singapore and whether the parties’ fields of business activity overlap under the element of misrepresentation. In the latter ground of opposition, such actual use of the Respondent’s Mark (Singapore) is relevant to determining whether the mark is well known to a relevant sector of the public in Singapore.

The second point is whether the fact that the parties operate in different industries militates against a finding of similarity of goods and services. To recapitulate, the appellant emphasises that many of its products are sold to “government bodies or Original Equipment Manufacturers” to be incorporated into other products. In contrast, the respondent’s goods are sold to the “general public as complete end users systems”, and cater to four main business fields: retail, food, logistics and hospitality.

The appellant’s argument is misplaced. Under s 8(2)(b) TMA, I am not comparing the parties’ actual goods and services only. In opposition proceedings, when comparing the specifications of competing marks at the second stage of the step-by-step approach, the court must consider all actual and notional fair uses of the goods and/or services in the specification(s) in which the earlier mark is registered, and the application mark is sought to be registered (Staywell at [40], [41], [60]).

Accordingly, even if the actual industries the parties presently operate in are different, this is not the end of the inquiry. As I explore subsequently at, inter alia, [150], [151], [163] and [167], the notional fair uses of the goods and/or services reflected in the competing marks’ specifications are relevant to a finding of similarity.

Class 9 v Class 9

The Respondent’s Mark (Singapore) is registered in Class 9 for “[b]alance and scale, scale equipped with printer, printer”. The Class 9 specification of the Application Mark is more extensive and is set out in full at [A.1].

I find that the competing marks’ specifications in Class 9 are similar.

Under factors (a) and (f) of the British Sugar factors (see [137] above), I agree with the respondent that the marks’ Class 9 goods are complementary in nature. The complementary nature of the marks’ goods and/or services may not, by itself, be sufficient for a finding of similarity. It is, however, a relevant consideration under British Sugar factor (f) (Johnson & Johnson v Uni-Charm Kabushiki Kaisha (Uni-Charm Corp) [2007] 1 SLR(R) 1082 at [14], cited in Monster Energy Company v Glamco Co, Ltd [2021] 3 SLR 319 at [68]). For example, a “scale equipped with [a] printer” (ie, the respondent’s specification) may involve the use of a wireless printer. This would require some “computer network connectivity hardware [and software]” (which are found in the Application Mark’s Class 9 specification) to transmit data between the scale and wireless printer. Even the appellant accepts that its products are “used wherever connection with machines is required”.

Such complementarity would also mean that the end users of both parties’ Class 9 goods are potentially the same (British Sugar factor (b)). The goods are also likely to be marketed through highly similar trade channels which, I expect, would be a combination of physical and online stores (British Sugar factor (d)). For instance, in respect of physical stores, Mr Kishi’s First Affidavit contains examples of retailers at Sim Lim Square, Singapore, offering both IT hardware (eg, printers, monitors, web cameras) and connectivity components (eg, WiFi mesh router, WiFi/LAN Adaptors”). These examples illustrate that the competing marks’ Class 9 goods will be distributed through highly similar trade channels.

In my view, the appellant’s Class 9 Amendment, that “none of the aforementioned goods [will relate] to balance and scale, scale equipped with printer, printer” is insufficient to dispel similarity. The fact remains that the appellant’s goods in Class 9 may be used in connection with the respondent’s goods in Class 9. To compound matters, the appellant has produced brochures marketing such goods falling within its Class 9 specification (eg, computer network connectivity hardware and software), which are complementary to those in the specification of the Respondent’s Mark (Singapore), bearing the Application Mark. These brochures name the appellant’s subsidiary in Singapore, Digi International Singapore (“Appellant’s Singapore Subsidiary”). Although I subsequently observe at [221(b)] that there is no evidence that these brochures have been viewed or distributed in Singapore, or of sales in Singapore of the goods featured in the brochures, the point on the complementary nature of the goods remains valid. The goods featured in these brochures include:

(a) Digi One SP IA, a serial server providing entry level serial-to-Ethernet connectivity for virtually any industrial device, including bar-code readers, scales, or temperature sensors;

(b) Hubport, a switched USB expansion hub that adds USB ports to a PC, server or thin client for instant peripheral device connectivity. It may be used to connect a PC to a printer. Hubport can also connect to another of the appellant’s products, Edgeport, a USB-to-serial convertor. The latter may connect to scales and cameras. Both Hubport and Edgeport bear the Application Mark.

(c) DIGI Connect SP, a serial server that can be connected to a deli scale and controlled through the appellant’s cloud-based device management platform. The former bears the Application Mark.

So long as the appellant’s Class 9 goods may be used generally “wherever connection with machines is required” (see [150] above), this complementarity is one of several factors establishing the similarity of both marks’ Class 9 specifications.

The following authorities raised by the respondent further support a finding of similarity in respect of the marks’ Class 9 specifications.

In Opposition No. B 2604 125 Indra Sistemas, S.A. v Kare en cours de formation (“Indra Sistemas”), a decision of the EU Intellectual Property Office (“EUIPO”) Opposition Division, the opponent’s relevant Class 9 specification was “Computer software, recorded; computer programs, recorded; computer programs, (downloadable software)”. The applicant’s trade mark was sought to be registered for, inter alia, “[p]rinters for computers” under Class 9. The EUIPO Opposition Division held that the applicant’s goods were (at p 72):

In the present case, the Application Mark’s Class 9 specification does not contain the broad category of “computer software”. It does, however, include “[c]omputer hardware”. Commonsensically, electronic scales and/or balances and printers would contain some computer hardware. My earlier analysis of the British Sugar factors – which recognises that the competing marks’ Class 9 goods are complementary, share the same end-users and distribution channels – gives voice to this reality. Indra Sistemas illustrates that when this combination of factors presents itself in a comparison between two Class 9 specifications, a finding of similarity is warranted.

The respondent also relies on DigiPoS v Digi (EU) appeal, in which the EU OHIM Boards of Appeal dismissed the appeal from the first instance OHIM decision of Digi v DigiPoS (EU) referred to at [57] above. DigiPoS sought registration in respect of, inter alia, “Class 9 – … printers …”. However, a mark identical to the Appellant’s Old Mark (Singapore) was registered for, inter alia, computer network connectivity hardware in Class 9. The EU OHIM Boards of Appeal found that these goods were “very similar” at [32]:

In our case, the Application Mark’s Class 9 specification contains “computer network connectivity hardware” while the Respondent Mark’s (Singapore) specification contains “printer”. The foregoing excerpt, which compares these very two goods, supplements my reasons at [150] and [151] on why the competing marks’ Class 9 goods are complementary.

Finally, the respondent also relies on the decision of the Korean Intellectual Property Office (“the Korean Decision”). In the Korean Decision, the respondent successfully opposed the registration of a mark identical to the Application Mark (seeking registration in Classes 9, 38 and 42) (“Application Mark (Korea)”) on the basis of a mark identical to the Respondent’s Mark (Singapore) (registered in Classes 7 and 9) (“Respondent’s Mark (Korea)”). The Korean Intellectual Property Office held that the specification of the Class 9 goods the Application Mark (Korea) was sought to be registered for, including “computer hardware, OEM computer modules, USB sensors, single board computers, computer hardware with embedded computer software, namely, modules, microprocessors, single-board computers, electronic sensors, etc.” and the Respondent’s Mark’s (Korea) specification in Class 9, including “… electronic load calculating machines and tools having automatic recording of package, …, electronic application label printer, …, label printer with electronic scale, electronic printer …” are similar. This is because they all correspond to electronic application tools automatically processing data and are identical or similar in the range of the supplier, seller and consumer of goods.

While the reasoning of the Korean Intellectual Property Office is brief, its message is clear. The Korean Intellectual Property Office was satisfied that the Class 9 goods in the Application Mark’s (Korea) specification and Respondent’s Mark’s (Korea) specification were identical or similar in terms of their suppliers, sellers and consumers. Likewise, in this case, for the reasons explained above at [150]–[152], there is similarity between the competing marks’ Class 9 specifications.

Class 38 v Class 9

The Class 38 specification the appellant seeks registration for may be found at [A.3]. The appellant has not made an amendment to the Class 38 specification equivalent to that for Class 9.

The Adjudicator reasoned that since the appellant’s Class 38 specification “closely correspond[s]” with its own Class 9 specification, the former would also overlap with the respondent’s Class 9 specification (GD at [59]). The Adjudicator appears to have applied the following proposition laid down by the Court of Appeal in Staywell at [43]. In that case, the competing marks’ specifications in Class 43 were found to be similar. Since the application mark’s specification in Class 35 was “closely related and indeed inseparable” from its own Class 43 specification, the application mark’s Class 35 specification was regarded as similar to the opponent’s mark’s class 43 specification (Staywell at [43]).

The respondent’s submissions on this point adopt the more traditional approach of comparing the appellant’s Class 38 specification and the respondent’s Class 9 specification. This is also the approach I will take.

In connection with British Sugar factors (a) and (f), I accept the respondent’s submission that the Application Mark’s services in Class 38 are complementary to the respondent’s Class 9 goods. For instance, printers that require connection to computers, and scales and balances that require connection to computers and/or printers, would utilise services stipulated in the Application Mark’s Class 38 specification, eg, “transmission of M2M (machine-to-machine) data; transmission of IoT (internet of things) data” and “wireless data network services for others, namely, electronic transmission of data related to connectivity, device management, and memory storage via wireless networks”. As a case in point, the respondent has sold scales or weighing devices with wireless capabilities and which connect to other products such as electronic labels or data storage and processing platforms (ie, SM-5300 and SM-500V2 at [114] above).

Further, I agree that these telecommunications and connectivity services in Class 38 may be marketed together with the respondent’s Class 9 goods. For example, a notional fair use of the appellant’s Class 38 specification, namely, “telematic services for sending of information, namely, providing machine-to-machine (M2M) and internet of things (IoT) mobile-device management (MDM) connectivity over long distances and/or remote locations” is to host a cloud server on which customers may store data that they want to print subsequently at a different location. This is the very service provided by Fuji Xerox’s “Cloud On-demand Print” server. That Fuji Xerox sells this service online confirms, directly, the overlap between the trade channels of the respondent’s Class 9 goods, which include online stores, and the appellant’s Class 38 services. More generally, however, the complementary nature of the appellant’s Class 38 services and the respondent’s Class 9 goods makes it likely for both to be sold through highly similar trade channels (both physical and online) (British Sugar factor (d)). To underscore this point, Mr Kishi’s First Affidavit contains multiple examples of businesses advertising goods and complementary IoT-related services together. Two such examples are: (a) Hanshow offering “integrated Omni-channel solutions for retail stores with a collection of smart hardware devices and management software”; and (b) NCR offering “the software, hardware and services [customers] need” in the areas of POS hardware and software, mobile and online ordering and consumer engagement. It follows from the foregoing that the respective “end users” of the appellant’s services and the respondent’s goods will also be highly similar (British Sugar factor (b)).

Accordingly, I reach the same conclusion as the Adjudicator, that there is similarity between the appellant’s Class 38 specification and the respondent’s Class 9 specification, albeit by a different route.

Class 42 v Class 9

The appellant’s Class 42 specification may be found at [A.4]. The appellant has not made an amendment to the Class 42 specification equivalent to that for Class 9. The Adjudicator applied the same reasoning for Classes 38 and 42 of the appellant’s specifications. In short, as the appellant’s Class 42 specification “closely correspond[s]” to its own Class 9 specification, the former would overlap with the respondent’s Class 9 specification. The respondent’s submissions on Class 38 apply equally to Class 42.

With regards the British Sugar factors (a) and (f), my view on the complementarity of the marks’ relevant specifications in the context of the appellant’s Class 38 specification would apply equally to this comparison between the appellant’s Class 42 services and the respondent’s Class 9 goods (see [163] above). To take an example, printers that require connection to computers, and scales and balances that require connection to computers and/or printers, would utilise services stipulated in the Application Mark’s Class 42 specification, eg, the “integration of disparate computer systems, networks, hardware and software through the application of wireless communication technology to facilitate M2M and IoT communication via web based browsers, personal digital assistants, mobile phones, embedded microprocessors, sensors and other electronic devices”. It bears emphasising that the respondent has sold scales or weighing devices with wireless capabilities and which connect to other products such as electronic labels or data storage and processing platforms (ie, SM-5300 and SM-500V2 at [114], [163] above). As a result, the appellant’s Class 42 services will likely be sold through highly similar trade channels (British Sugar factor (d); see [164] above) to highly similar end users as the respondent’s Class 9 goods (British Sugar factor (b)).

The respondent cites the EUIPO Opposition Division decision Opposition No. B 3 097 447 S.C. Dante International S.A. v Love Letters Limited (“Dante International”) to buttress the similarity of the appellant’s Class 42 services and the respondent’s Class 9 goods. The EUIPO Opposition Division held that “notebook computers” in Class 9 were similar to services in Class 42 such as “research and development of new products for others” and “consultancy in the design and development of computer hardware” as producers of computers and/or software would commonly render such services. Relevantly, the EUIPO Opposition Division said that (at p 4):

In the present case, the appellant’s Class 42 specification contains highly similar services to those discussed in the excerpted passage in Dante International, namely, “design and development of computer software and middleware for others; computer network design for others, consulting with regard to computer systems, computer network connectivity hardware and computer network connectivity software and middleware; technical support services, namely, troubleshooting and diagnosing computer system software and computer network connectivity software and middleware; …; technical consulting and assistance with computer-based information systems and components, namely, technological consulting services in the field of datacenter architecture, public and private cloud computing solutions, and evaluation and implementation of Internet technology and services” (see [A.4] below). The reasoning of the EUIPO Opposition Division confirms, inter alia, that these Class 42 services are “complementary [to] and target the same public [as]” related Class 9 goods. The corollary of this is that these goods and services will be marketed via highly similar trade channels. Dante International underscores my earlier views on the British Sugar factors at [167] above.

As such, I find that the appellant’s Class 42 services and the respondent’s Class 9 goods are similar, albeit for different reasons to the Adjudicator.

Conclusion on goods and services similarity

Based on the foregoing, I am satisfied that there is similarity between the Class 9 specification of the Respondent’s Mark (Singapore) and the specifications in Classes 9, 38 and 42 in which the Application Mark is sought to be registered.

Is there a likelihood of confusion?

The likelihood of confusion inquiry directs the court to look at (a) how similar the marks are; (b) how similar the goods or services are; and (c) given this, how likely it is that the relevant segment of the public will be confused as to the origin of the appellant and respondent’s goods and services (Staywell at [55]; Hai Tong at [73]). At (c), the test to be adopted in determining whether a likelihood of confusion exists is whether a substantial portion of the relevant public will be confused. This standard is above de minimis and must be appreciable, though it is not necessary to show that a majority of the public is confused (Sarika Connoisseur Cafe Pte Ltd v Ferrero SpA [2013] 1 SLR 531 (“Sarika”) at [57]). It is also useful to set out the Court of Appeal’s exposition of what confusion as to the origin of the parties’ goods and services entails in Hai Tong at [74]:

It bears re-iterating that under s 8(2)(b) of the TMA, the inquiry takes into account the actual and notional fair uses of both the opponent’s mark and the application mark in respect of their specifications (Staywell at [60]). As summarised by the Court of Appeal in Ceramiche Caesar SpA v Caesarstone Sdot-Yam Ltd [2017] 2 SLR 308 at [56], the factors relevant to the confusion inquiry are:

I begin with the question of who the relevant public is. The relevant public is the segment of the public comprising the actual or potential purchasers of the goods or services of the appellant and respondent and those who deal with these goods (Ng-Loy Wee Loon, Law of Intellectual Property of Singapore (Sweet & Maxwell, 2nd Ed, 2014) (“Ng-Loy”) at para 21.5.27). I accept that the relevant public is the general public in Singapore. I can envisage the notional fair uses of both parties’ goods and services being purchased by sophisticated customers and the general public. I therefore reject the appellant’s contention that the relevant public is confined to “OEMs” (see [49] above).

With this in mind, I turn to the confusion inquiry proper.

I first consider the overall degree of similarity of the competing marks. In this regard, the Adjudicator held that the “aural, visual and conceptual similarity in the marks, bearing the same sound and letters and similar triangular device, significantly increase the likelihood of confusion” (GD at [66]). To this, the respondent adds that the competing marks are, in totality, similar to a high degree and that their common denominators (ie, the word “DIGI” and triangle device) would give the relevant public the impression that there is an economic link between the goods and services provided under each mark. The respondent also emphasises the well-known status of the Respondent’s Mark (Singapore) and its technical distinctiveness (both inherent and acquired). Given the imperfect recollection of the relevant public, the differences between the competing marks will be barely perceptible and the relevant public will believe that the Application Mark is a modern evolution of the Respondent’s Mark (Singapore).

I agree with the respondent that the competing marks are, overall, similar to a high degree. This conclusion flows from my earlier findings that the marks are aurally identical, conceptually similar to a high degree and visually similar to a fair degree.

I next consider the degree of similarity of the goods and services under the competing marks’ specifications. I accept that the goods in the Respondent’s Mark’s (Singapore) Class 9 specification and the goods and services in the Application Mark’s Specifications in Classes 9, 38 and 42 have different functional purposes. While the Application Mark’s goods and services enable M2M communication and network connectivity, the Respondent’s Mark’s (Singapore) goods are for weighing, measuring and/or printing. However, given the complementarity of the Application Mark’s Specifications and the Respondent’s Mark’s (Singapore) specification, the fact that these goods and/or services are likely sold through highly similar trade channels and to highly similar end-users, I find that the specifications of the competing marks have a moderate degree of similarity.

Finally, I arrive at the crucial question of whether there is a likelihood of confusion given the high and moderate degree of similarity of the competing marks’ and their specifications respectively. I structure my analysis according to two broad factors which the Adjudicator discussed and the respondent submitted on.

Mode of purchase

First, the Adjudicator considered the relevance of the mode of purchase of the parties’ goods and services. She said at [66] that “[t]he goods and services may also be sold online or may be purchased physically in person at a shop. The aural, visual and conceptual similarity in the marks, bearing the same sound and letters and similar triangular device, significantly increase the likelihood of confusion.” The respondent builds on this by arguing that these modes of purchase would cause the visual similarities of the marks to assume greater importance. The differences in the marks are insufficient to dispel the risk of confusion. In particular, for online purchases, the respondent cites Hai Tong at [53] for the proposition that the textual component of the marks would be the “most important”. Since both the Application Mark and Respondent’s Mark (Singapore) have the same textual component, this significantly increases the likelihood of confusion.

I agree with the Adjudicator that the Application Mark’s goods and services and the Respondent’s Mark’s (Singapore) goods would be sold through both physical and online stores (see [151] above).

With regards sales in physical stores, I accept that the relative significance of the visual similarity of the competing marks is high. However, the aural similarity of the competing marks is also important as consumers would use “DIGI” as an identifier when making enquiries in a shop. The competing marks’ fair visual similarity and aural identity must also be appreciated in the light of the fact that the distinctiveness of the Respondent’s Mark (Singapore) has been enhanced to a moderate degree through use (see [118] above). Accordingly, the differences between the triangle devices in the competing marks are insufficient to dispel the impression of similarity in the average consumer’s mind. In fact, I believe the Adjudicator hit the nail on the head when she said at [48] that:

Consequently, there is a likelihood of consumers being confused as to whether goods bearing the Respondent’s Mark (Singapore) and goods and services offered under the Application Mark emanate from the same source.

With regards online stores, I have difficulty agreeing with the respondent’s point that the focus on the competing marks’ common textual component would significantly increase the likelihood of confusion (see [180] above). At [94], I concluded that the word “DIGI” is descriptive when seen in the context of the Respondent’s Mark’s (Singapore) Class 9 specification. As such, if it is true that the textual component becomes the “most important” element in online purchases, the likelihood of origin-based confusion may be low.

However, although Hai Tong at [53] states that the textual component of a composite mark is generally dominant where goods or services bearing the mark are marketed and sold predominantly through online trade channels, I do not think I can discount the composite whole of the competing marks entirely. This is because in the context of the goods and services in the Application Mark’s specifications and the goods in the Respondent’s Mark’s (Singapore) specification, it is not uncommon for online stores to display the marks used on the goods or in connection with the services being sold, or to provide photographs of the actual goods which allow online consumers to observe the mark on the good itself. As such, the high degree of similarity of the composite whole of the competing marks remains relevant and reinforces the likelihood of confusion in the online sphere.

Nature of goods

On this front, the Adjudicator held at [66] of the GD that the goods sold by the parties cover a broad price range, with the simpler and less expensive products being more likely to be purchased by individuals who would generally have less specialised needs and knowledge and who would pay less attention in the distinction when selecting the products for purchase. The respondent agrees with this observation.

I likewise concur with the Adjudicator’s view. Notionally, the appellant’s goods and services and the respondent’s goods may involve high-end and complex technology on one end of the spectrum, and low-end and simple technology on the other. Put another way, the parties may provide both expensive and inexpensive goods and services.

However, even for the general public purchasing the parties’ relatively inexpensive goods and services for household or personal use, it cannot be gainsaid that these are technical products. It would not be unusual for the average consumer to perform some research or comparisons in the market before completing their purchase. For “highly technical products”, a higher degree of similarity between contesting marks might have to be shown to warrant a finding of likely confusion (Hai Tong at [85(c)(i)]). Considering the fact that these goods and services for household use are unlikely to be highly technical, and that the competing marks are highly similar overall, I am satisfied that there is a likelihood of confusion among the general public. Further, the general public would be less fastidious when making their purchase as compared to the segment of the relevant public comprising sophisticated consumers.

This leads me to consider the likelihood of confusion among the segment of the relevant public comprising sophisticated consumers such as business consumers, IT professionals and IT enthusiasts. This segment of the relevant public would apply a greater degree of fastidiousness and specialised knowledge when conducting their purchases. This is especially so as I think that both parties’ goods and services would involve a not insignificant capital outlay by businesses looking to setup their IoT or network connectivity infrastructure and/or purchase electronic scales and balances and printers in bulk. However, even for this segment of the relevant public, the fact remains that the competing marks are aurally identical, visually and conceptually similar, and overall similar to a high degree. The specifications of the competing marks are also similar to a moderate degree as the goods and/or services therein are complementary in their uses and are likely to be sold through highly similar trade channels to highly similar end-users. In my judgment, there is still a likelihood of confusion in this segment of the relevant public, even if to a lower degree than among the general public.

To summarise my findings thus far, I am satisfied that (a) the marks are overall similar to a high degree; (b) the marks’ specifications are similar to a moderate degree; (c) the modes of purchase (both physical and online) point towards a likelihood of confusion; and (d) there is a likelihood of confusion despite the technical nature of the goods. On the whole, I am satisfied that there exists a likelihood of confusion among a substantial portion of the relevant public. I will now consider some remaining arguments advanced by the appellant to dispel the likelihood of confusion.

Relevance of Australian Registrar’s Decision

The appellant further argues that the Australian Registrar’s Decision supports its position that there is no likelihood of confusion in Singapore.

However, for the reason set out at [107] above, I find the Australian Registrar’s Decision to be unpersuasive. My conclusion on the issue of the likelihood of confusion remains unchanged.

Relevance of co-existence of the Appellant’s DIGI-related Marks (Singapore) in Singapore, the appellant’s other marks around the world and the Application Mark in the Overlapping Territories

The appellant also submits that the long period of co-existence of the parties’ marks after the 2002 Agreement (in Singapore and outside of Singapore) and the ongoing co-existence of marks identical to the Application Mark and Respondent’s Mark (Singapore) in the Overlapping Territories supports the conclusion that there will be no likelihood of confusion among the relevant public.

First, as a general point, I do not regard the state of the register prior to the application to register the Application Mark, in Singapore or outside Singapore, to be relevant. British Sugar supports this position. That case concerned a trade mark infringement claim. The plaintiff’s mark, “Treat”, was registered for “[d]essert sauces and syrups”. The defendant had used “Robertson’s Toffee Treat” for a toffee flavoured spread. At issue was the distinctiveness of the “Treat” mark. Jacob J reasoned as follows (at 305):

Second, in particular connection to the argument that the identical iterations of the Application Mark and Respondent’s Mark (Singapore) co-exist in the Overlapping Territories, I agree with the respondent that it is entitled to elect which jurisdiction(s) to contest the registration of the Application Mark in. This is a multi-faceted commercial decision which I am not in a position to question. As such, the co-existence of identical iterations of the Application Mark and Respondent’s Mark (Singapore) in other jurisdictions merely shows that the respondent has chosen not to oppose the Application Mark’s registration in those jurisdictions.

Conclusion on likelihood of confusion

For the foregoing reasons, I find that there is a likelihood of confusion among a substantial portion of the relevant public. I affirm the Adjudicator’s finding that the ground of opposition under s 8(2)(b) of the TMA succeeds.

Whether the ground of opposition under s 8(7)(a) of the TMA is made out?

To succeed on the ground of opposition under s 8(7)(a) of the TMA, the respondent must establish the three elements of passing off: (a) goodwill; (b) misrepresentation; and (c) likelihood of damage to goodwill (Novelty at [37] and affirmed in Singsung at [28]).

Goodwill

Goodwill has been described as “the attractive force which brings in custom” (The Commissioners of Inland Revenue v Muller & Co’s Margarine, Limited [1901] AC 217 at 224, cited in The Singapore Professional Golfers’ Association v Chen Eng Waye and others [2013] 2 SLR 495 (“SPGA”) at [21]; The Audience Motivation Company Asia Pte Ltd v AMC Live Group China (S) Pte Ltd [2016] 3 SLR 517 (“AMC”) at [81]). Goodwill is often proved by evidence, including that of the trader’s sales and the expenses incurred in promoting his goods and services in association with the mark or brand that they bear (SPGA at [22]; Tuitiongenius Pte Ltd v Toh Yew Keat and another [2021] 1 SLR 231 (“Tuitiongenius”) at [85]). In this element, the court is only concerned with whether there is goodwill in the respondent’s business. The question of whether the Respondent’s Mark (Singapore) is sufficiently associated with the respondent’s business should be dealt with, as a matter of analytical clarity, as a threshold issue under the element of misrepresentation (Tuitiongenius at [83]; AMC at [88]).

The Adjudicator held that the evidence of the Respondent’s Singapore Subsidiary’s sales and expenses incurred in promoting the goods and services in association with the Respondent’s Mark (Singapore) in Singapore was sufficient to establish its goodwill in Singapore (GD at [71]). She noted that the appellant did not advance any contrary evidence to refute the respondent’s.

In its submissions before me, the appellant did not directly challenge the existence of the respondent’s goodwill, save for arguing that no element of passing off is made out when one considers the co-existence of the Appellant’s DIGI-related Marks (Singapore) and the Respondent’s Mark (Singapore) in Singapore. I deal with the relevance of such co-existence at [214]–[222] below.

In light of the evidence at [105] above, particularly the respondent’s revenue figures in Singapore, I have no doubt that the element of goodwill is established. I am fortified in my decision by the SUM 2178/2020 evidence (see [114], [142] above), including the portions which pertain to goods and services beyond the scope of the Respondent’s Mark’s (Singapore) Class 9 specification. In Mr Kishi’s First Affidavit, he states that the respondent has “continuously and extensively provided … cloud services” and “[g]oods and systems incorporated with wireless capabilities” [emphasis in original in underline]. Mr Kishi also claims that the respondent’s Singapore sales figures for the set-up and maintenance of its cloud services (excluding figures for goods that are built with these cloud capabilities) rose from approximately $26,000 in 2011 to $165,000 in 2016, and that its Singapore sales figures for its wireless goods and services rose from approximately $16,000 in 2011 to $88,000 in 2016. I place reliance on supporting documentary evidence, viz, the respondent’s brochures, invoices and/or records from the Respondent’s Singapore Subsidiary’s accounting system, in both Mr Kishi’s First Affidavit and Mr Heng’s statutory declaration to find, at least, that the following examples of the Respondent’s Wireless Goods and IoT services were sold in Singapore:

(a) The Delious Cloud, which provides centralised management of sales, purchasing and attendance at food and beverage establishments. The Delious Cloud allows users to track shop performance in real time, access data anywhere and transmit and manage data and to perform label design. This is one example of the respondent’s IoT services; and

(b) The Delious Pico POS terminal (“the Delious Pico”), which caters to food and beverage businesses and comes with, inter alia, a built-in receipt printer. The Delious Pico is connected to the Delious Cloud to consolidate reports for multiple outlets for management and monitoring remotely online. This is one example of the Respondent’s Wireless Goods.

These sales of the Respondent’s Wireless Goods and IoT services in Singapore buttress its goodwill here.

Misrepresentation

The element of misrepresentation requires proof that the appellant’s goods and services will be held out to be, or to be connected with, the respondent’s business, thereby giving rise to confusion (Novelty at [77]; AMC at [86]; Tuitiongenius at [88]). In connection with this, the threshold issue is whether the respondent’s goodwill is sufficiently associated with the Respondent’s Mark (Singapore). Sufficient association will be established if the respondent can show that the Respondent’s Mark (Singapore) is distinctive of its business (see AMC at [87]–[88]; Hai Tong at [115]), such that the relevant public views that mark as an indicator of origin (see Rovio at [172]; Tuitiongenius at [88]).

The threshold requirement of distinctiveness is met given my findings at [118]. My conclusion that the Respondent’s Mark (Singapore) acquired a moderate degree of distinctiveness is bolstered by the fact that the Respondent’s Wireless Goods and IoT services (which fall outside of the Respondent’s Mark’s (Singapore) Class 9 specification) set out at [200] above were sold in Singapore by reference to the Respondent’s Mark (Singapore).

Preliminarily, the appellant first submits that the respondent consented to its registration of the Application Mark by virtue of the 2002 Agreement. As such, there is no element of deceit by way of the appellant trying to steal the respondent’s goodwill surreptitiously (Tuitiongenius at [99]). This argument fails as I have already found at [83] above that the respondent provided no such consent under the 2002 Agreement.

Likelihood of confusion amongst the relevant public

I am of the view that the appellant’s use of the Application Mark in connection with its goods and services would falsely represent that these goods and services and the respondent’s goods emanate from the same source (see [182] above).

The remaining, but essential, question is whether the misrepresentation would lead to confusion amongst ordinary sensible members of the relevant public (Asia Pacific Publishing Pte Ltd v Pioneers & Leaders (Publishers) Pte Ltd [2011] 4 SLR 381 at [123]; SPGA at [41]–[42]). Actual confusion is not required for a passing off action to succeed. The likelihood of confusion amongst a substantial proportion of the relevant public is sufficient (Novelty at [77]).

In passing off, the relevant public was described by the Court of Appeal in Novelty as persons who are actual or potential customers of the plaintiff in the passing off action who have goodwill vis-à-vis the plaintiff’s mark/name/get-up (Novelty at [73] and [75]). Subsequently, in SPGA at [43], the Court of Appeal stated that the relevant public “typically is that segment of the public in which the claimant’s goodwill actually or potentially, directly or indirectly subsists” [emphasis added]. Sundaresh Menon CJ elaborated as follows (at [49]):

The respondent submits that the relevant public comprises “the businesses and individuals in the [r]espondent’s trade in the retail, food, logistics, and hospitality industry.” Conspicuously, this omits members of the general public who form part of the relevant public under s 8(2)(b) TMA (see [174] above). As the respondent has not shown nor argued that its line of goods includes inexpensive consumer-grade products, I proceed on the basis that the general public does not comprise its potential customers and therefore is not part of the relevant public in passing off. However, I do not think the relevant public is confined to the trade customers in the retail, food, logistics and hospitality industries. Applying Menon CJ’s pronouncements in SPGA at [43] and [49], the relevant public includes trade customers in other industries in which the claimant’s goodwill potentially, directly or indirectly subsists.

The test for a misrepresentation which creates a likelihood of confusion is substantially the same as that for “likelihood of confusion” under s 8(2)(b) of the TMA (Ferrero SPA v Sarika Connoisseur Café Ptd Ltd [2011] SGHC 176 at [212]; Sarika at [77]). I have earlier found at [195] that there is a likelihood of confusion under s 8(2)(b) of the TMA. I re-iterate my finding at [188] that even among sophisticated consumers like business consumers, IT professionals and IT enthusiasts, there is a likelihood of confusion.

The SUM 2178/2020 evidence, pertaining to goods and services falling outside of the Respondent’s Mark’s (Singapore) Class 9 specification, is also relevant to the confusion inquiry in s 8(7)(a) of the TMA (see [144] above). As explained at [200] and [203] above, this further evidence confirms sales of the Respondent’s Wireless Goods and IoT services in Singapore by reference to the Respondent’s Mark (Singapore).

Taking the SUM 2178/2020 evidence into account, there is a high degree of proximity in the parties’ fields of activity. Not only are the parties’ goods complementary (see [152], [163] and [167] above), the very nature of the goods and services they offer are now in competition. In particular, the respondent has begun to provide services that create an IoT ecosystem by, inter alia, facilitating M2M communication and centralised data management (eg, the Delious Cloud software at [200(a)] above). Such services are highly similar, in my view, to the appellant’s goods and services which “[provide] the essential layer of [M2M] communications, and the remote monitoring and management of devices that critical applications depend on.”

Additionally, while the appellant claims not to directly compete with the respondent in the retail, food, logistics and hospitality industries, I find this hard to accept. The statutory declaration of Mr Mauer disclosed brochures featuring products by the appellant which are catered for these industries. Although I subsequently observe at [221(b)] that there is no evidence that these brochures have been viewed or distributed in Singapore, or of sales in Singapore of the goods featured in the brochures, the point on my difficulty in accepting the appellant’s position remains. The goods featured in these brochures include:

(a) Edgeport, a USB-to-serial converter that makes it easy to add serial and USB ports to a PC, server, or thin client. It is designed for “peripheral device connectivity in retail stations, self check-out systems, …, kiosks, ATMs and more” [emphasis added]. Possible applications include connecting a PC to credit card readers, bar-code scanners, receipt printers, touch screen monitors, pole displays and voucher checkers. While not explicitly stated in the appellant’s product brochure, I can envisage Edgeport being relevant in the hospitality industry as well. The Appellant’s Singapore Subsidiary is named at the bottom right of the last page of the Edgeport product brochure;

(b) DIGI Connect SP, a serial server that enables plug and play device network as well as custom networking applications. Possible applications include connecting an information management system, via DIGI Connect SP, to a deli scale. While not explicitly stated in the appellant’s product brochure, I can envisage this product having applications in the food industry. The Appellant’s Singapore Subsidiary is named at the bottom right of the last page of the DIGI Connect SP product brochure; and

(c) Portserver TS connects multiple serial devices over industry standard Ethernet. “Target applications include retail/point-of-sale, building control and medical point-of-care.” Possible applications include connecting a PC server to bar-code scanners and credit card readers. The Appellant’s Singapore Subsidiary is named at the bottom right of the last page of the Portserver TS product brochure.

Evidently, there is both (a) a high degree of similarity in the nature of the parties’ goods and services; and, (b) an overlap in the industries they target. It also bears emphasising that the parties’ goods and/or services are distributed through highly similar trade channels, as is evidenced by several real-life examples found in Mr Kishi’s First Affidavit (see [151], [164] above). Therefore, even though the relevant public comprises sophisticated consumers, given the extremely close proximity of the parties’ fields of business, compounded by the high degree of similarity of the competing marks (see [177] above), and low inherent distinctiveness but moderate acquired distinctiveness of the Respondent’s Mark (Singapore) (see [118] above), I find that there is a likelihood of confusion amongst the relevant public. The element of misrepresentation is made out. I should also state that the very fact that the parties’ entered into the 2002 Agreement and 2003 Canadian Settlement Agreement to demarcate a scope of goods and services for each party is telling. They must have recognised the likelihood of confusion in the market if both traded under “DIGI” related marks.

Relevance of registration and use of the Appellant’s DIGI-related Marks (Singapore) in Singapore

The appellant submits that “the specifications of the [Appellant’s DIGI-related Marks (Singapore)] are relevant to disprove [the respondent’s] claims concerning goodwill, misrepresentation and damage.” If the respondent did not oppose the registration of these marks, then it must mean that the use of these earlier marks would not constitute passing off. It is therefore inconsistent for the respondent to say that the use of the Application Mark is passing off as it seeks registration for similar specifications and the Application Mark and Appellant’s DIGI-related Marks (Singapore) contain the word “DIGI”.

One issue is whether the existence of the Appellant’s DIGI-related Marks (Singapore) on the register is relevant to this ground of opposition. I need only refer to Jacob J’s pronouncement in British Sugar (see [193] above) to conclude that reliance on the register in the manner described in the preceding paragraph is generally irrelevant. These marks’ registrations leave us none the wiser as to “what the circumstances were which led the registrar to put the marks concerned on the register” and “what is actually happening out in the market”. Therefore, even if the appellant is right that the specifications of the Application Mark are, in substance, similar to specifications of the Appellant’s DIGI-related Marks (Singapore) (see [51] above), this does not assist its case.

However, the question of whether the appellant’s prior use (if any) of the Appellant’s DIGI-related Marks (Singapore) is relevant in the passing off inquiry is an interesting one. The respondent makes two points. First, the respondent cites the EU Court of First Instance (“CFI”) decision Grupo Sada, pa, SA v OHIM Case T-31/03, ECLI:EU:T:2005:169 (“Grupo”) for the proposition that for such co-existence of earlier marks on the market to reduce the likelihood of confusion, the appellant here must demonstrate that such co-existence was “based upon the absence of any likelihood of confusion ….and… that the earlier marks concerned and the marks at issue are identical” (Grupo at [86]). The appellant is unable to satisfy either of these requirements. In particular, the respondent highlights that there are not “inconsequential” differences between the Appellant’s DIGI-related Marks (Singapore) and the Application Mark. Alternatively, if the use of the Appellant’s DIGI-related Marks (Singapore) is relevant, the respondent submits that the appellant “failed to provide any supporting evidence relating to sales and advertising of its products and services” in Singapore. Such insufficiency of evidence precludes the appellant from claiming concurrent goodwill.

With regards reducing the likelihood of confusion, the two requirements laid down in Grupo are that (a) the co-existence of the earlier marks and the Respondent’s Mark (Singapore) must be based upon the absence of any likelihood of confusion; and (b) that the earlier marks concerned and the marks at issue are identical. In a similar vein (although I make no finding as to whether both Grupo requirements should be followed in Singapore), the Court of Appeal in Super Coffeemix Manufacturing Ltd v Unico Trading Pte Ltd and another and another appeal [2000] 2 SLR(R) 214 (“Super Coffeemix”) made the following observation. In that case, the Court of Appeal upheld the judge’s dismissal of the plaintiff’s claim for passing off as the plaintiff was found to have no goodwill in Singapore or acquired distinctiveness in respect of the mark “COFFEEMIX” (at [58], [60]–[61]). The Court of Appeal then went on to state (at [64]):

While absence of evidence of actual confusion is not fatal to a claim for passing off (Singsung at [40]), Super Coffeemix demonstrates that the prior use of the Appellant’s DIGI-related Marks (Singapore) in Singapore may, in this case, preclude the finding that there is likelihood of confusion when no confusion in the market has been shown.

Nevertheless, in my judgment, the appellant’s reliance on the prior use of the Appellant’s DIGI-related Marks (Singapore) in Singapore fails on a more fundamental level. Namely, there is insufficient evidence to establish prior use of these marks in Singapore.

The respondent stresses that the appellant “failed to provide any supporting evidence relating to sales and advertising of its products and services” in Singapore under the Appellant’s DIGI-related Marks (Singapore).

Having carefully considered the appellant’s evidence alongside the respondent’s submissions, I agree with the respondent. Generally, the appellant’s evidence of use of the Appellant’s DIGI-related Marks (Singapore) or advertisement of goods and services offered under these marks in Singapore is deficient in these respects:

(a) The evidence relates to use of these marks outside of Singapore, or does not clearly evidence use in Singapore (eg, the location of tradeshow or conference is unspecified and unknown).

(b) The appellant disclosed brochures of its goods and/or services bearing the Appellant’s DIGI-related Marks (Singapore). However, although these brochures identify the Appellant’s Singapore Subsidiary, there is no evidence that these brochures have been viewed or distributed in Singapore, or of sales in Singapore of the goods and/or services featured in the brochures (“no evidence of Advertisement or Sales in Singapore”).

In fact, the appellant also adduced similar brochures featuring goods and/or services bearing the Application Mark. These latter brochures identify the Appellant’s Singapore Subsidiary. However, there is similarly no evidence of Advertisement or Sales in Singapore in connection with these brochures.

(c) The appellant’s evidence of having advertised goods and services under the Appellant’s DIGI-related Marks (Singapore) does not clearly show that any advertising expenditure was incurred on advertisements in Singapore, or which were targeted at and viewed by the relevant public in Singapore.

I note that Mr Mauer’s statutory declaration states that the appellant’s annual revenue in Singapore for goods and services offered under the Appellant’s DIGI-related Marks (Singapore) rose from $4,955,244 in 2011 to $8,244,566.40 in 2016. However, as no supporting documents (eg, invoices) were exhibited to substantiate these figures, I am unable to ascribe them any weight.

In these premises, the appellant’s argument, that prior use of its DIGI-related Marks (Singapore) in Singapore disproves the elements of passing off, fails. In addition, for the reasons stated at [221] and [222], the appellant cannot rely on the defence of concurrent goodwill to resist the ground of opposition under s 8(7)(a) TMA.

Damage

A misrepresentation is actionable only if it has caused or is likely to cause damage to a plaintiff’s goodwill (Tuitiongenius at [101]; AMC at [94]; Novelty at [94]). Examples of cases where damage in the form of diversion of sales or custom has been recognised include:

(a) Hai Tong, where the Court of Appeal at [118] found that there was real likelihood of damage to the appellant’s goodwill arising from the diversion of sales of “Lady Rose” products if consumers of the relevant segment of the public were led to believe that “Rose Lady” cosmetic products were the goods of the appellant or were otherwise associated with or connected with the appellant. Such a likelihood of damage arose given that the parties were in direct competition, the parties’ mark and sign were similar and there was a likelihood of confusion in relation to the trade mark infringement claim.

(b) In Singsung at [87], the Court of Appeal found that the appellant’s goodwill would be adversely affected through a diversion of custom, given that the appellant and respondent were competing in exactly the same line of products and in the same export jurisdictions.

The Adjudicator held that there was a real tangible risk of the respondent’s goodwill being adversely affected through a diversion of custom. Her conclusion was based on her earlier findings that the competing marks were similar (aurally, visually and conceptually), the products and services covered by the competing marks were similar and there was a likelihood of confusion (GD at [75]).

In light of the authorities cited at [224], I likewise find that there is a likelihood of damage to the appellant’s goodwill in the form of diversion of custom. Most pertinently, I have held that the parties’ businesses are in competition (see [211] above), that their marks share a high degree of overall similarity (see [177] above), and that there is a likelihood of confusion among the relevant public that the appellant’s goods and services emanate from the same source as the respondent’s goods.

In conclusion, I uphold the Adjudicator’s finding that the ground of opposition under s 8(7)(a) of the TMA succeeds.

Whether the ground of opposition in s 8(4)(b)(i) of the TMA, read with s 8(4)(a), is made out?

Tay Yong Kwang J (as he then was) in Formula One Licensing BV v Idea Marketing SA [2015] 5 SLR 1349 at [107] summarised the elements that must be proven in order to establish the ground of opposition under s 8(4)(b)(i) TMA:

(a) the whole or an essential part of the application mark is identical with or similar to an earlier trade mark;

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

(c) use of the application mark on goods or services that it is to be registered for will indicate a connection between those goods or services and the proprietor of the earlier trade mark; and

(d) there is likely to be damage to the interests of the proprietor of the earlier trade mark.

The first two elements are uncontroversial. The court in Rovio at [146] clarified that despite the different language used to describe the marks similarity requirement in ss 8(4) and 8(2)(b) of the TMA, there is no difference in substance between the marks similarity analysis in these two provisions. I repeat my finding, under s 8(2)(b) TMA, that the competing marks are similar to a high degree (see [177] above). As for the requirement that the Respondent’s Mark (Singapore) is well known in Singapore, the appellant is not challenging this element on appeal. The SUM 2178/2020 evidence, including portions which pertain to goods and services falling outside of the Respondent’s Mark’s (Singapore) Class 9 Specification, also strengthens the reputation of the Respondent’s Mark (Singapore). This further evidence shows that the respondent used the Respondent’s Mark (Singapore) on its Wireless Goods and IoT services in Singapore prior to the filing of the Application Mark, thereby increasing the extent of its use here. Having regard to the principles and factors in ss 2(1), 2(7)–2(9) TMA (see [33] above), and the evidence of the use of the Respondent’s Mark (Singapore) in Singapore since 1987, including the relevant sales figures in Singapore (see [105] above), I see no reason to disturb the Adjudicator’s finding (see [33] above).

As for the elements at [228(c)] and [228(d)], the Court of Appeal in Novelty at [234] held that the tests for the “connection” requirement and “likely to damage the interests of the proprietor of the earlier trade mark” requirement in s 55(3)(a) of the TMA would yield the same results as those obtained from applying the corresponding tests under the elements of misrepresentation and damage in passing off. As s 8(4)(b)(i) is the analogue of s 55(3)(a) in the context of opposing registrations of trade marks, this proposition in Novelty applies to s 8(4)(b)(i) TMA. I am therefore satisfied that the elements at [228(c)] and [228(d)] are made out given my earlier findings that the elements of misrepresentation and damage are established in the ground of passing off (see [213] and [226] above).

For these reasons, I uphold the Adjudicator’s finding that the ground of opposition under s 8(4)(b)(i) TMA succeeds.

Whether the ground of opposition under s 7(6) of the TMA is made out?

This is the respondent’s cross-appeal against the Adjudicator’s dismissal of the ground of opposition under s 7(6) of the TMA. s 7(6) states that: “[a] trade mark should not be registered if or to the extent that the application is made in bad faith.”

The general principles governing the law on bad faith are not in dispute and are summarised in Tomy Inc v Dentsply Sirona Inc [2020] 5 SLR 424 at [32]. Relevant to the present appeal are the following principles:

(a) Bad faith “is to be determined as at the date of application and matters which occurred after the date of application which may assist in determining the applicant’s state of mind as at the date of application can be taken into consideration”: Festina Lotus SA v Romanson Co Ltd [2010] 4 SLR 552 (“Festina Lotus”) at [100]; see also Hotel Cipriani SRL and others v Cipriani (Grosvenor Street) Ltd and others [2008] EWHC 3032 (Ch) (“Cipriani”) at [166]–[167].

(b) Bad faith “includes dishonesty and…includes also some dealings which fall short of the standards of acceptable commercial behaviour observed by reasonable and experienced men in the particular area being examined”: Gromax Plasticulture Ltd v Don & Low Nonwovens Ltd [1999] RPC 367 (“Gromax”) at 379, affirmed in Valentino at [25]. The test is a “combined” one and “contains both a subjective element (viz, what the particular applicant knows) and an objective element (viz, what ordinary persons adopting proper standards would think)”: Wing Joo Loong Ginseng Hong (Singapore) Co Pte Ltd v Qinghai Xinyuan Foreign Trade Co Ltd and another and another appeal [2009] 2 SLR(R) 814 at [105]; affirmed in Valentino at [29]; see also Cipriani at [166]–[167].

(c) Bad faith requires a holistic assessment. As summarised in Lionel Bently et al, Intellectual Property Law (Oxford University Press, 5th Ed, 2018) at p 1019, the court ought to take the following into account:

This was in reference to Chocoladefabriken Lindt & Sprüngli AG v Franz Hauswirth GmbH (Case C-529/07) [2009] ETMR 56 at [37]–[38], which has been affirmed in Philip Morris Products S A v PT Perusahaan Dagang Dan Industri Tresno [2010] SGIPOS 8 at [219].

(d) Where allegations of bad faith are concerned, the “standard of proof is on the balance of probabilities but cogent evidence is required due to the seriousness of the allegation”: Cipriani at [177]; see also Valentino at [30]. Nevertheless, a conclusion of bad faith is “largely, if not invariably, based on circumstantial evidence”: see Festina Lotus at [115]; see also Brutt Trade Marks [2007] RPC 19 at [29]–[30] and Ng-Loy at para 21.4.1.

The Adjudicator dismissed the allegation of bad faith. She held that there is no requirement under the 2002 Agreement for parties to inform or consult with each other if a new “DIGI” related mark is developed (GD at [28]). She also rejected the suggestion that the 2002 Agreement prohibited parties from adopting a new mark which was more similar to the other’s mark for goods and services beyond the demarcated limits of the 2002 Agreement (GD at [28]).

On appeal, the respondent argues that the appellant acted in a manner which is contrary to the spirit of the 2002 Agreement, even if not in breach of its terms:

(a) the appellant “should have known” that the respondent would likely have objected to the Application Mark and “should have broached the change with the [r]espondent with the view to updating the 2002 Agreement to avoid any ambiguity”. This is because: (a) the Application Mark is more similar to the Respondent’s Mark (Singapore) than the Appellant’s Old Mark (Singapore); and (b) the appellant is applying for registration in respect of specifications beyond the limits permitted by the 2002 Agreement ; and

(b) bad faith is evidenced by the appellant’s Class 9 Amendment, which was done “late”, and the appellant has begun trading in contravention of this amendment. In this regard, the respondent points to the appellant’s computer hardware and software relating to weighing and/or measuring purposes or transfer of data connected to weighing and/or measuring purposes.

In response, the appellant agrees with the Adjudicator that there is no obligation on either party to consult with the other before making changes to their respective marks. It adds that Clauses 4 and 5 of the 2002 Agreement envision that parties may apply for “new registrations” of trade marks. However, this latter point fails given my earlier finding that the 2002 Agreement does not govern the registration of the Application Mark (see [83] above).

First, I do not accept that the appellant should have known that the respondent would likely have objected to the Application Mark. This is because the appellant has already registered a variety of trade marks incorporating the word “DIGI” in Singapore, that are different from the Appellant’s Old Mark (Singapore), without any opposition from the respondent. Such trade marks include:

In these circumstances, the following pronouncement by Arnold J in Cipriani at [189] (re-iterated in Och-Ziff Management Europe Ltd and another v OCH Capital LLP and others [2011] Bus LR 632 at [37]) is highly relevant:

A similar proposition is advanced in Tan Tee Jim, Law of Trade Marks and Passing Off in Singapore vol I (Sweet & Maxwell,, 3rd Ed, 2014) at para 7.170, which was affirmed by Chan Seng Onn J in Harvard Club of Singapore v President and Fellows of Harvard College [2020] 4 SLR 1378 at [117].

Considering the appellant’s evidence, this appears to be a case where the appellant, albeit mistakenly, honestly believed it had a right to register the Application Mark. The respondent has not proven otherwise. For instance, Mr Mauer’s view is that the green triangle device in the Application Mark is unlikely to inflict any additional damage on the interests of the respondent over and above what the Appellant’s Old Mark (Singapore) would. This is because the respondent does not claim exclusivity over the small green triangle device, and there are marks on Singapore’s register with a triangle device, including those set out at [101] above.

Even further still, Mr Mauer’s view is that: (a) the Application Mark’s Specifications fall within the Appellant’s Specified Goods and Services in the 2002 Agreement, viz, the broad umbrella of “data processing devices and computer”; and (b) the Application Mark’s Specifications do not include goods and services concerning weighing and measuring devices and thus do not encroach on the Respondent’s Specified Goods and Services.

It bears emphasising that cogent evidence is required before a finding of bad faith is made (see [233(d)] above). Having regard to all the circumstances of this case, I am not satisfied that the appellant knew that the respondent was likely to oppose the registration or that it lacked a right to register the Application Mark. The subjective limb of the test for bad faith is not made out.

Second, the point that the appellant is trading in contravention of the appellant’s Class 9 Amendment does not take the respondent’s case much further. As I pointed out at [221(b)], there is no evidence that the brochures featuring goods and/or services bearing the Application Mark have been viewed or distributed in Singapore, or of sales in Singapore of the goods and/or services featured in those brochures.

Based on the foregoing, I uphold the Adjudicator’s finding that the ground of opposition under s 7(6) of the TMA fails.

Conclusion

I dismiss the appeal and cross-appeal against the Decision of the Adjudicator and summarise my findings as follows:

(a) I decline to register the Application Mark under s 8(9) of the TMA.

(b) I uphold the Adjudicator’s finding that the grounds of opposition under ss 8(2)(b), 8(7)(a) and 8(4)(b)(i) of the TMA are made out.

(c) I uphold the Adjudicator’s finding that the ground of opposition under s 7(6) of the TMA fails.

I take this opportunity to express my view that this is a case in which both parties’ commercial interests would have been best served by arms-length negotiations out of court. The parties know their businesses best and should have worked towards updating the demarcation of goods and services that each may operate in, to best balance their interests. Doing so would have saved parties the time and expense of prosecuting proceedings relating to the Application Mark in multiple jurisdictions and facing the spectre of inconsistent outcomes.

Annex 1: Specification of goods and services in each Nice Class for the Application Mark