Unscrambling the trail judgment of Matter v Zynga in the Court of Appeal - A Case Update | Fieldfisher
Skip to main content

Unscrambling the trail judgment of Matter v Zynga in the Court of Appeal - A Case Update



United Kingdom

Review of the Court of Appeal judgment in Mattel v Zynga regarding the SCRABBLE and SCRAMBLE trade marks

This article follows Leighton's previous post (from January 2014) which summarised and commented upon Mr Justice Peter Smith's High Court judgment which held that Zynga did not infringe Mattel's SCRABBLE trade marks.


By way of reminder, JW Spears & Son own the rights and produce and market the board game Scrabble and is itself owned by US toy manufacturing company Mattel  (collectively, "Mattel").  Zynga develops apps for mobile platforms.  

The dispute related to Zynga's mobile app Scramble with Friends, which requires its players to find as many words as possible within a jumbled grid of letters within a short space of time.  Mattel alleged that Scramble with Friends infringed its registered rights:


Mattel's Registrations


Zynga Signs

CTM registration no. 401737 SCRABBLE in Classes 9, 28 and 41







 Scramble with friends

CTM registration no. 401489 SCRABBLE Device in Classes 9, 28 and 41


UK trade mark registration no. 2154349 for a three dimensional tile mark in Classes 9, 28 and 41

 Three dimensional tile mark

CTM registration no. 6223077 SCRAMBLE in Classes 9, 28 and 41



Smith J found that Mattel's CTM SCRAMBLE had been infringed by Zynga's use of SCRAMBLE but was invalid and/or liable for revocation in any event by virtue of its descriptive nature (Article 7(1)(b)) and the fact that it had become a common name in the trade for a product in respect of which it was registered (Article 51(1)(b)).  

As regards Mattel's claims of infringement of its SCRABBLE CTMs, these were dismissed by Smith J who somewhat controversially concluded that they were not similar to the signs SCRAMBLE or SCRAMBLE WITH FRIENDS and as such Mattel were unable to successful make out a case of likelihood of confusion under Article 9(1)(b). The only exception to this finding related to Zynga's use of a stylised "M" in its SWF logo which the judge considered liable to confuse the relevant public when viewed "on a quick glance."

In reaching this decision, Smith J relied heavily on the actions and inactions of Mattel and its employees; in particular the fact that Mattel allowed several earlier versions of the SCRAMBLE app to exist before deciding to take action and that Mattel had pitched to Zynga in relation to its wider portfolio which included the SCRAMBLE app. 

Smith J also brutally dismissed Mattel's survey evidence in relation to which Mattel's request to appeal was declined. 

The Grounds of Appeal

Both sides appealed Smith J's judgment. Mattel appealed against the finding of non-infringement (and passing off) of its CTMs.  It also appealed against the Judge's finding of invalidity/liability for revocation of the SCRAMBLE mark.   Mattel's appeal was primarily based on the following:

  • That the judge's finding as regards no similarity between the SCRABBLE marks and the SCRAMBLE and SCRAMBLE WITH FRIENDS signs were unsupportable. 
  • More particularly, that the judge erroneously applied a "minimum threshold of similarity" test and as such did not conduct the objective assessment of whether the marks were sufficiently similar to give rise to a likelihood of confusion. This included a failure to consider the way that Mattel had used the mark in terms of contextual matter extraneous to the mark itself (relying on Specsavers International Healthcare Ltd v Asda). 
  • That the judge allowed too much weight to be given to the fact that Mattel (i) failed to take action as regards the earlier versions of the allegedly infringing game; and (ii) had previously pitched to Zynga to make physical versions of Zynga's games.
  • That the judge failed to deal with its claim under Article 9(1)(c) at all and should have done given its non-reliance on a finding of likelihood of confusion.

Zynga cross-appealed against the judge's finding of infringement and passing off as regards Zynga's use of the stylised "M" in the SWF logo. 

The Appeal was heard on 18-20 November 2014 by Lord Justices Patten, Tomlinson and Floyd.  Floyd LJ delivered the leading judgment with which Patten LJ and Tomlinson LJ agreed.  Judgment was handed down on 27 March 2015. 

General observations on the Trial Judgment

Floyd LJ did not go to any great lengths to reveal his dissatisfaction with the way in which Smith J reached his final decision in the lower court.  He openly criticised the way that Smith J had focussed, "almost to the exclusion of everything else, on the attitudes and actions of the protagonists to the dispute". Floyd LJ recognised some force in the submissions made on behalf of Mattel that Smith J had "lost sight of the underlying exercise which the jurisprudence of the CJEU required him to undertake." (Para 123).

Whilst accepting that there was certainly scope for it, it was decided that a re-trial would not be in the best interest of either parties.  Remittal was also considered but given that the relevant material was already before the Lord Justices, the Court of Appeal in the end elected to "consider it all as best we can" (para 125). 

Appeal Findings

(i) Smith J's finding that SCRAMBLE CTM is invalid/ liable for revocation is reversed 

Floyd LJ began by reversing the trial judge's finding as regards the validity and/or liability for revocation of the SCRAMBLE CTM.  Again he criticised Smith J's assessment of this claim in that he failed to properly consider the arguments for invalidity as at the relevant date, i.e. in 2007 and his conclusion that the mark was "descriptive of the [Zynga] game" cannot be supported as at that point in time no such game existed. Smith J also failed to address the question of what the sign SCRAMBLE would immediately convey to the average consumer of the goods and services for which it was registered – "It is not enough to show that a particular game later accused of infringement has a characteristic which can be accurately described, with hindsight, by the sign." (Para 130). 

Considering both the rushed activity meaning and the jumbled letters meaning of SCRAMBLE in turn, Floyd J concluded that whilst there remained some degree of descriptive allusion in the mark SCRAMBLE he did not consider it far enough "towards the descriptive end of the spectrum" for it to be deemed unregistrable under 7(1)(c).  He also considered there to be nowhere near enough evidence to establish customary use or established practice under Article 7(1)(d).   

As regards Smith J's finding that SCRAMBLE had lost its distinctive character by being used commonly in the trade by different traders as a generic name for the goods concerned, Floyd LJ also took issue, and agreed with Mattel's submissions that the judge did not sufficiently scrutinise the evidence advanced in support of the claim.  Floyd LJ overturned this finding for, amongst other reasons, the fact that he himself remained unsure at the end of the appeal and even when writing the judgment what game or type of game the word SCRAMBLE is supposed to depict. 

(ii) Infringement of the SCRAMBLE CTM

Floyd LJ agreed with Smith J's findings that the CTM SCRAMBLE and the sign SCRAMBLE were not only liable to be confused but that confusion was inevitable. 

(iii) Infringement of SCRABBLE CTMS under 9(1)(b)

It was clear that by concluding that the self-evidently similar SCRABBLE and SCRAMBLE were dissimilar, Smith J had applied a threshold of similarity test. The Court of Appeal confirmed that, in accordance with the precedent handed down in L'Oreal v Bellure, this test is wrong in law.  Zynga's claims that subsequent CJEU decisions (Ferrero, Kaul, Lufthansa Airplus and Riha Weser-Gold) had cast doubt on the principle laid down in L'Oreal, fell on deaf ears.  Floyd LJ stressed that while these CJEU cases make the point that an overall finding of similarity of signs may well be precluded by virtue of significant dissimilarities as regards (a) visual; (b) aural; or (c) conceptual similarity, "it is the overall similarity that counts" (Para 59) and where there is a finding of overall similarity, however faint that similarity may be, a global assessment of similarity is required. 

Having found that the SCRABBLE and SCRAMBLE were moderately similar (visually and phonetically very similar, with only a lesser degree of conceptual similarity), Floyd LJ conducted the requisite global assessment that had erroneously been absent from Smith J's trial judgment. 

(a) The Global Assessment

In conducting this assessment, the Court identified the key factors in the context of the strength of each parties' respective claims. 

He identified the following as weighing in Mattel's favour:

  • Identity between the goods for which the mark is registered and the alleged infringing goods  (having legitimately disregarded known differences in gameplay);
  • The enhanced distinctive character of SCRABBLE and the fact that the average consumer is used to seeing/trained itself to see the mark in the context of games (although Floyd LJ rejected Mattel's submissions that contextual material extraneous to its SCRABBLE CTMs should be considered on the basis that it stretched the Specsavers principle that bit too far).
  • The fact that the product is free or relatively cheap and as such the consumer will pay less attention when purchasing it than if it were a more expensive item.   

Weighing in Zynga's favour were the following points:

  • That whilst Zynga's claims that the mark was descriptive in the trade had been rejected, there nevertheless remained a certain degree of descriptive allusiveness which would weigh in favour of Zynga's argument that a relevant consumer would not mistakenly link the sign SCRAMBLE with the well-known SCRABBLE mark.  
  • Equally, whilst SCRAMBLE sign had not become a common name in the trade for the game, the fact remained that users who searched for a game on the App Store or Google Play would be confronted with a variety of games using the name SCRAMBLE in some way. 
  • That the method of purchasing apps is such that a consumer is able to examine each app carefully before purchasing even where the app is cheap or free. 
  • The absence of weighty evidence showing actual confusion in the market despite the extent of contemporaneous, side by side use of the respective marks and signs.   

Floyd LJ noted that but for Zynga's arguments (particularly as regards an absence of evidential proof of actual confusion), the Court of Appeal would have been inclined to conclude that there was a risk of confusion without needing to go further.  However such a conclusion would not be supported without reference to the secondary evidence.  

 (b) Analysing the Evidence presented  

Although Mattel's survey evidence was not before the Court of Appeal (having been rubbished by Smith J and Mattel being refused permission to appeal this particular finding), Floyd LJ still took the opportunity to make his views known on the difficulty of deploying such evidence successfully; namely, that it is extremely difficult, if not impossible, to reproduce a real life situation and that questions are inevitably posed in a way that is liable to distort the natural conclusions that an average consumer would ordinarily reach.   

The remaining evidence before the Court of Appeal principally comprised third party reviews (from iTunes and Google Play) and other social media commentary.  Floyd J disagreed with Smith J's outright dismissal of this evidence on that basis that it was mere "hearsay" and as such incapable of satisfactory levels of scrutiny.   Whilst accepting the evidence's inherent limitations (significantly, the inability to cross-examine the numerous authors), Floyd J found a vast number of the numerous tweets (which inadvertently confused SCRABBLE and SCRAMBLE) to be compelling by virtue of their "unsolicited" nature.   

Floyd LJ took little interest in the evidence that went to the actions or inactions of Mattel and its employees.  He criticised Smith J's approach in erroneously equating one witness' view with that of the relevant consumer.   He also failed to see any significance in Mattel's pitches for a licence from Zynga – whilst these pitches may have been inconsistent to Mattel's claims of infringement, they did little to advance Zynga's case as regards likelihood of confusion.  In any event, Zynga had not pleaded acquiescence.   

Despite Floyd LJ's very different (and unquestionably more objective) assessment of the evidence, he ultimately reached the same conclusion as Smith J as regards the infringement of the SCRABBLE CTMs:-  "..although, as I have explained, I disagree with the judge's approach to the secondary material, looking at it as a whole, I am not persuaded that it gives sufficient support to Mattel's case to justify allowing the appeal on infringement of the SCRABBLE mark.  The principal reason for that conclusion is that the extent of side by side use would be expected to have produced for more abundant evidence of confusion by the date of trial.  The factors relied on by Zynga must, therefore, be sufficient in combination with the small difference between the mark and sign, to eliminate the risk of real confusion occurring amongst a significant proportion of the public."   [Emphasis added]. 

As regards passing off, on the basis that neither side sought to distinguish this from their respective submissions as regards CTM infringement under 9(1)(b), Floyd LJ treated them the same and accordingly held that this claim had not been made out. 

(iv) Infringement of SCRABBLE CTMS under 9(1)(c)

Floyd LJ dealt with Mattel's claim under 9(1)(c) quickly – he was not persuaded that there was sufficient evidence presented as to the requisite unfair advantage or detriment. 


Floyd LJ's choice of words; that he lacked sufficient evidence to justify overturning the High Court's decision is perhaps enlightening and may suggest that had the same assessment been conducted in the High Court (without the weight of a previous judgment) the outcome may have been different. This is particularly so given the objective similarity (as regards signs and goods covered) and the unquestionable enhanced distinctiveness of the SCRABBLE marks.  

Nevertheless, this judgment is a warning to trade mark owners seeking to claim infringement that, where there has been extended co-existence between their earlier rights and the allegedly infringing signs, abundant and compelling evidence of actual confusion will be required even where marks are on the face of it, self-evidently similar. 

The judgment also serves as a helpful reminder of the proper assessment of similarity under 9(1)(b) and dispels any doubt that may have been cast on L'Oreal's rejection of a threshold of similarity approach. 

Finally, the judgment suggests that whilst insufficient in this case, third party evidence extracted from online platforms and in particular from social media, can be treated as compelling (by virtue of its unsolicited nature) where it demonstrates actual evidence of confusion.