Zee Entertainment Enterprises Ltd & Ors v Zeebox Ltd

This is an appeal from the judgment and order of Birss J in a passing off and trade mark infringement action. By his order of 7th June 2013, the judge refused permission to the appellants to carry out a market research survey in support of their allegation of passing off.

The appeal involves the application of the principles and guidance given by this court about the approach to the evidence of surveys, and evidence derived from surveys, in two recent cases Marks and Spencer v Interflora Incorporated [2012] EWCA (Civ) 1501 and Marks and Spencer v Interflora Incorporated [2013] EWCA (Civ) 319:(“Interflora 1” and “Interflora 2”).

Zee TV is a satellite and cable television channel. Broadcasting from India, it provides a wide range of programming in Hindi and other regional Indian languages. It launched its services in the United Kingdom in 1995, targeting its service at the United Kingdom's Asian population. Its current channels are Zee TV, Zee Cine, Zee Cafe and Zee Punjabi. In addition to making use of the word "Zee", all these channels make use of a large stylised "Z" device with the words "Zee TV" beneath.

Information about the appellants' services is provided on its Indian and UK websites. The websites contain video clips, details about programmes, a schedule and various downloadable items. The UK website includes the ability to interact via Facebook and Twitter. The appellants claim to have built up a significant reputation and goodwill in the name Zee TV and its device.

In November 2011, the respondent launched an app. An app (which is short for application) is a small computer program which can be installed on smartphones and tablet devices. The app is marketed under the name "Zeebox". The icon which appears on a smartphone is a "Z" with the word Zeebox underneath. The Zeebox app is designed to be used while the user is watching television on another device. It provides a means for discovering more about programmes while they are being watched. It includes a TV guide, the ability to remote-control a connected television or other device from a phone or tablet, and the ability to chat live during TV programmes. It is available free of charge, but makes money by advertising.

The app does not itself enable users to watch or record TV. The aim of the app is to make watching TV a more social experience. One example of the Zeebox branding appears in the judgment of the learned judge at paragraph 6 and is reproduced below:

The appellants claim that the respondent, by use of the word "Zeebox" and the "Z" device, are committing the tort of passing off. The respondent denies that the word or device is distinctive of the appellants and denies that a substantial number of members of the public will believe that their products and services are connected in the course of trade with the appellants or their services. The respondent relies heavily on the fact that here has not been any recorded instance of actual confusion or deception to date.

In September 2012, the appellants carried out a pilot survey, Pilot 1, which was objected to by the respondent on various grounds. In response to the criticisms, on 23rd May 2013, the appellants commissioned a further pilot, Pilot 2, intended to deal with the respondent's objections to Pilot 1. It was carried out on 27th and 28th May, only a few days before the hearing before Birss J.

Both surveys were carried out by an independent market research organisation, BDRC Continental. They were intended to target what were described as "British Asian people". Members of the public were interviewed in the street at locations in London for the first pilot and in Birmingham, Leicester and Manchester for the second pilot. Only 60 or so interviewees were included in the two pilots, it being intended to expand the number to around 500 if a full survey were to be permitted. The respondent was provided with all the survey materials generated by the pilots, including the questionnaires.

The respondents in the survey were first filtered in various ways. Questions 5 and 6 then restricted respondents to those who regularly use a smartphone or tablet. The key questions were questions 7-9. These were: "This is the name of an app relating to TV [or website relating to TV]. What can you tell me about it?" The precise question asked depended upon the results of the filtering. When being asked this question, the respondent was shown a card with the Zeebox logo and the word "Zeebox" underneath. The show card appears at page 159 of the appeal bundle and is reproduced below:

Question 10 read: "Who do you think is responsible for Zeebox?" Later questions investigated, with prompting, the types of TV the household received, which Asian TV channels they were aware of, their preparedness to be contacted and their ethnicity.

The appellants considered the results to be encouraging. By their original application, they envisaged calling not only evidence about the survey itself, supported by expert evidence, but in due course witnesses of fact from amongst the persons interviewed. In the course of his submissions before us, Mr. Malynicz, who appeared on behalf of the appellants, has indicated that the appellants do not now propose to call any such factual witnesses if the survey is admitted. This is a difference from the basis on which the judge considered the application.

The judge summarised the law as to the grant of permission to adduce survey evidence, in terms which neither side disputed, at paragraphs 14-28 of his judgment:

"The law.

"14.

Imperial Group PLC v Phillip Morris [1984] RPC 293 was the starting point for the modern approach to the admission of survey evidence in cases of trade mark infringement and passing off. In that case Whitford J laid down guidelines for the future conduct of surveys, which have since become known as the Whitford Guidelines. The Whitford Guidelines were recently summarised by Lewison LJ in Marks & Spencer v Interflora Incorporated [2012] EWCA Civ 1501, which I will refer to as Interflora 1. At paragraph 61 Lewison LJ said the following:

"(i)

If a survey is to have any validity at all, the way in which the interviewees are selected must be established as being done by a method, such that a relevant cross section of the public is interviewed;

"(ii)

any survey must be of a size which is sufficient to produce some relevant result viewed on a statistical basis;

"(iii)

the party relying on the survey must give the fullest possible disclosure of exactly how many surveys they have carried out, exactly how those surveys were conducted and the totality of the number of persons involved, because otherwise it is impossible to draw any reliable inference from answers given by a few respondents;

"(iv)

the questions asked must not be leading and must not direct the person answering the question into a field of speculation upon which that person would never have embarked had the question not been put;

"(v)

exact answers and not some sort of abbreviation or digest of the exact answer must be recorded;

"(vi)

the totality of all answers given to all surveys should be disclosed;

"(vii)

the instructions given to the interviews must also be disclosed."

"15.

Lewison LJ, in Interflora 1, also said:

"Evidence that is not useful should not be allowed to distract the focus of a trial, even if it is technically admissible."

"16.

He indicated that surveys that are not fully compliant with the Whitford Guidelines are one example. At paragraph 21 he said:

"The value of the evidence is severely diminished if not eliminated by a failure to follow the Whitford Guidelines."

"17.

In Interflora 1 Lewison LJ summarised that the standard practice in trade mark and passing off cases in future should be as follows:

"(i)

A party may conduct a true pilot survey without permission, but at its own risk as to costs;

"(ii)

no further survey may be conducted or adduced in evidence without the court's permission;
"(iii) no party may adduce evidence from respondents to any survey without the court's permission."

18.

Then, at paragraph 151:

"When applying for permission, as referred to in subparagraph 2 above, the applicant should provide the court with:
"(i) the results of any pilot survey;
"(ii) evidence that any further survey will comply with the Whitford guidelines;
"(iii) the cost of carrying out the pilot survey and the estimated cost of carrying out the further survey."

"19.

At paragraph 147 of Interflora 1 Lewison LJ explained that such a procedure provides the court with the material it needs to rule definitively on whether the cost of the full survey should be undertaken. It would be a false kindness simply to permit the survey to go ahead on the basis that the trial judge can always subsequently rule it inadmissible. He said:

"The objective of such an application is to have a definitive ruling one way or the other. It is a natural temptation for a judge who is not immersed in the case to leave questions of admissibility to trial. It is the temptation to which I succumbed in UK Channel Management Limited v E! Entertainment Television Incorporated. But balancing the cost of a survey (or witness collection exercise) against its likely utility, this temptation should be resisted."

"20.

The test to be applied is a cost benefit test (see paragraph 150 of Interflora 1):

"In deciding whether to give permission, the court must evaluate the results of whatever material is placed before it. Only if the court is satisfied that the evidence is likely to be of real value should permission be given. The reliability of the survey is likely to play an important part in that evaluation. Even then the court must be satisfied that the value justifies the cost. As Mr. Hobbs said, this requires the court to conduct a cost benefit analysis. In a case of trade mark infringement, in which the issue is one of deception in relation to the provision of ordinary consumer goods or services, these criteria are likely to be satisfied only in a special or unusual case."

"21.

Having dealt with paragraph 150, I will note that in this action it is the passing off case and not a trade mark case that I am being asked to adduce the survey in relation to.

"22.

The question of survey evidence returned to the Court of Appeal in a case referred to as Interflora 2, which is Interflora Inc. V Marks & Spencer [2013] EWCA Civ 319. Lewison LJ, at paragraphs 26 to 28, said as follows:

"26.

With the benefit of hindsight, perhaps I did not make my message clear enough in Interflora 1. Let me say it again but more loudly. A Judge should not let in evidence of this kind unless the party seeking to call that evidence satisfies him (a) that it is likely to be of REAL value; and (b) that the likely value of the evidence justifies the cost.
"27. Thus, in my judgment, the judge should have asked himself whether the evidence was likely to be of real value. The negative way of framing the test means that evidence which was not of real value will be admitted in a case which, after all, is about the provision of ordinary consumer services.
"28. I am unable to tell from the judge's statement that he considered that the evidence would be likely to have some value, whether he simply thought that the value would be more than minimal or whether he thought it would carry real weight. If the latter, I find it hard to identify why he thought that. In my judgment, therefore, the judge conducted a flawed analysis of the nature of the application, the quality of the raw data that supported the witness statement and ultimately applied too lax a test."

The judge also reminded himself that, as from 1st April 2013, applications of this kind had to take account of the overriding objective, as amended in the light of Jackson LJ's reforms. These include the objective of enabling the court to deal with cases justly and at proportionate cost.

In his judgment, Birss J considered that the case turned essentially on two points. The first was whether it was legitimate in law to look at a distinct group of the public, namely, British Asians. He considered that the survey was irrelevant to that issue. He pointed out that if the appellants were wrong on this issue, the survey would be of no value.

The second point was whether the Zeebox app, as a television app, was confusingly similar to the appellant's TV channels and website. On this issue, the judge pointed out that watching television was not an esoteric exercise and the judge would be well-able to make an assessment. The use of apps might be more esoteric, but could be readily explained.

The judge declined to decide the criticisms made by the respondent's expert, Mr. Malivoire, of the questions asked in the survey. He thought the court might come to either view as to whether there was substance in those criticisms. The judge did reject two other criticisms of the survey, one which went to Pilot 1 alone and another which complained about the representative nature of the survey.

The judge held that the survey evidence would add some £150,000 to the costs, roughly half on either side. This was in the context of total costs of £1 million, with £600,000 on the respondent's side and £400,000 on the appellant's side. He said this at paragraphs 53 and 54 of the judgment:

"53.

As for a point about the lack of evidence of actual confusion, a court dealing with trade mark and passing off matters is very familiar with this question. There are all sorts of reasons why there may be a lack of confusion. The law is clear, that a lack of actual confusion is not an absolute requirement by any means. It is a relevant factor to be weighed up, but it will always be relevant, irrespective of the results of a survey.

"54.

It is obvious that the survey is not going to be all the evidence that the claimant relies on by any means and Mr. Malynicz did not say that it was. The claimant will be highly likely to be able to call proper evidence making out its case about a likelihood of misrepresentation on the British Asian community. Of course, the defendant will also be able to call evidence intending to try and show the opposite. However in my judgment if the evidence that the claimant is going to call anyway is not sufficient to persuade the trial judge in the claimant's favour then I doubt this survey will help. If I thought the passing off was weak then things might be different. Then things would be like the Jif Lemon case, in which an intrinsically weak passing off case was made good by survey evidence. But, in my judgment, in this case the results of a full survey based on this pilot may be confirmatory of a conclusion the judge has reached on the evidence already, but I think the survey is not likely to be of real value. It will be of some value, but I am not satisfied that the value justifies the substantial cost of carrying it out and the increase in the length of the trial which it will inevitably cause. Accordingly, I will dismiss the claimant's application."
The grounds of appeal

The grounds of appeal are five in number. They are: (1) that the judge failed to differentiate, in applying the real value test, between passing off and trade mark infringement cases; (2) that he failed to appreciate that the relevant public was a cultural and linguistic subsection of the UK public and not the population at large; (3) that he took too strict an approach to the real value test; (4) that he failed to take account of the fact that he had not found any criticism of the surveys to be proved; and (5) that he made errors in the assessment of the likely costs.

It is convenient to deal with those two grounds together. Mr. Malynicz pointed to the different legal question which the court is required to answer in a passing off case as distinct from a trade mark infringement case. This was made clear in Interflora 1: e.g. at [26]-[36] and [137].

Mr. Malynicz recognises that the judge had pointed out at paragraph 21 of his judgment that he was dealing here with a passing off case and not a trade mark infringement case as in Interflora, but submits that he failed to appreciate the significance of this in the present case. He says that the judge applied too strict a test, either requiring it to be the only evidence adduced or equating it with an examination of whether the survey evidence needed to be decisive of a finally-balanced case. He submits that in paragraph 54 of his judgment the judge wrongly took account of an interim assessment of the strength of the appellants’ case.

He points out that in the present case, and unlike Interflora, the evidence proposed to be adduced is the survey itself and not a carefully unrepresentative selection of witnesses from a survey conducted as a witness collection programme. He draws attention to the judge's reliance on his assessment at the interim stage of the strength of the appellants' case. He says that although this is comforting, a different judge might not take such a rosy prima facie view of the case. The judge erred in thinking, if he did, that evidence of the kind of which a survey could elicit would be available from elsewhere and that the non-esoteric nature of the goods and services rendered such evidence unnecessary.

Mr. Roberts, who appeared for the respondents, accepts that he cannot sidestep the indications in Interflora that different considerations may come into play in passing off cases. He submits, however, that the judge made it absolutely clear that he had this distinction well in mind and he submits that however much lower one sets the test for admission of a survey in passing off, as compared to trade mark infringement, the threshold was not crossed in the present case.

Mr. Roberts meets the judge's reference to his view of the strength of the appellant's case by referring to paragraph 45 of Interflora 1 where Lewison LJ said that it would be a compelling reason for admitting evidence of witnesses in Interflora and could not prove his case by any other means.

The question of whether the defendants are making a misrepresentation calculated to deceive is ultimately a question for the trial judge. In Spalding Brothers v AW Gamage Limited [1915] 32 RPC 273 on page 286, Lord Parker said:

"There can be no doubt that in a passing off action the question whether the matter complained of is calculated to deceive, in other words whether it amounts to a misrepresentation, is a matter for the judge who, looking at the documents and evidence before him, comes to his own conclusion and, to use the words of Lord Macnaghten in Payton v Snelling, 'must not surrender his own independent judgment to any witness whatsoever'."

However, as Jacob LJ recognised in Interflora 2, it is possible to conduct fair surveys and even lead evidence from witnesses obtained thereby.

I think it is clear from Interflora 1 that the real value test is primarily directed to the value of the evidence, its weight in the scales. Thus the particular question at issue in Interflora was the calling of witnesses obtained by means of a survey, without reliance on the survey itself. See paragraph 1 and see, for example, paragraph 54 which contrasts a properly conducted poll with the practice of calling a few hand-picked members of the public.

At paragraph 150 Lewison LJ emphasises that the reliability of the survey forms an important part of the real value determination and goes on to point out that that is not the end of the inquiry because the court must be still be satisfied that the value justifies the cost.

I do not think that the real value test is intended to invite the court to evaluate the likely outcome of the case at the interim stage. This would be a dangerous exercise. Suppose that the judge, at the interim stage, thinks that the case will probably succeed and refuses permission to conduct a proper survey, but the evidence adduced at the trial shows the case to be a marginal one and the claimant fails. The claimant would have a legitimate sense of grievance if his properly conducted survey had been excluded at the interim stage. I do not think that the passage relied from the judgment of Lewison LJ in Interflora supports the proposition that, in general, the apparent strength of the case is relevant. He was merely suggesting that a party who cannot prove their case by any other means would have a stronger case for the omission of a properly conducted survey.

In my judgment, although the real value test compounded in the Interflora cases does not make it necessary to scrutinise the strength of the claimant's case, it does require the judge to scrutinise the value of the survey as evidence. I will have to return to this topic later in the judgment.

I should add that Mr. Malynicz made detailed oral submissions to us as to the potential value of survey evidence in passing off cases. First, a survey could, for example, provide reassurance for a judge that his or her view was not idiosyncratic. Secondly, it could ensure that the judge is aware of the views of the relevant public, in the present case, watchers of the appellant's channel and users of apps, tablets and smartphones. Thirdly, this was particularly so where linguistic or cultural factors to come into play.

As was pointed out in the course of argument, however, many of these factors would be present in all passing off cases. In my judgment, one factor which might be of importance is where there is some special factor about the goods or services, or their consumers, which makes evidence of the kind likely to be of real value.

The appellants’ survey was targeted at members of the British Asian community. The appellants say that given the nature of the services offered by the claimant, which are also targeted at that community, the judge will be assisted by evidence of the reactions of the interviewees. They say that the trial judge is likely to have little familiarity with the appellant's channel or the reactions of the ethnically distinct market to which it is directed.

Mr. Malynicz cited in his skeleton argument the case of White Horse Distillers Ltd v Gregson Associates Ltd [1984] RPC 61, where the issue was whether the defendant whiskey exporters had sold scotch whisky in bulk to Uruguay for admixture with local spirits in circumstances that rendered them liable for passing off. The claimants put in evidence of a street survey carried out in Argentina where respondents were shown a colour photograph of an advertising hoarding showing a bottle of the product with its name in large letters and the slogan "Ecosces por crianza", meaning bred in Scotland. Nourse J said:

"In many, perhaps the majority of passing off actions, which depend on the get-up of an allegedly deceptive product, it is inevitable that judges will in the end make their own assessment of the likelihood of deception, even in cases where there is extensive evidence of confusion."

However, as regards the survey evidence, he said that he accepted the submission that:

"... it would be very dangerous for me not to give full weight to the evidence as a whole in a case where what is in issue is a question of passing off in a distant country with whose people and culture the court is unfamiliar."

Although Mr. Malynicz accepts that the passing off in the present case is not taking place in a distant country, he submits that it is taking place amongst a people and culture with whom the trial judge may be unfamiliar. He submits that in saying that the passing off case was concerned with matters which were not esoteric, the judge lost sight of this aspect of the case.

I do not think there is anything in these points. First, whilst of course the trial judge may not be a watcher of Zee TV, or indeed of any television channel at all, I have real difficulty in accepting that, when equipped with knowledge of the extent and popularly of the appellant's channel, he will not be able to assess the susceptibility of watchers of that channel to confusion or deception when confronted with the signs complained of. Judges in passing off cases are familiar with taking into account any relevant characteristics of the purchasing public and making allowances for it.

Secondly, the attempted analogy with White Horse v Gregson seems to me to be misplaced. In that case the impact of the defendant’s advertising in a foreign language was relevant, as were the ways in which whisky was sold in a foreign country.

This case, in contrast, is about a very short English trade mark and logo delivered to the public in the way that television services are the delivered to the public in this country. There is nothing in that which is likely to be unfamiliar or not readily understandable to the trial judge.

Thirdly, if the justification of the survey is the need to target speakers of Hindi because the appellant's channels are in Hindi, then the survey has failed to do that. As Mr. Roberts has demonstrated, significant numbers of interviewees speak nothing but English.

I turn to the specific criticisms of the survey. Mr. Malynicz has submitted that the judge made no criticism of the reliability of the survey and that this should be weighed in his favour. By his respondent's notice, Mr. Roberts maintains that the judge should have concluded that this survey was flawed and likely to be of no real value. As Lewison LJ pointed out in Interflora, the reliability of the survey is likely to be an important part of the evaluation of the real value of the survey.

Unless it is for some reason impossible to do so, I think that it is necessary, on an application of this kind, to assess on the available material whether the survey is likely to give rise to evidence of real value. That was made clear in Interflora, not only in paragraph 150 as quoted above in the judge's judgment, but also at paragraph 147, when explaining the purpose of the application procedure:

“The objective of such an application is to have a definitive ruling one way or the other. It is a natural temptation for a judge, who is not immersed in the case, to leave questions of admissibility to trial. … But balancing the cost of a survey (or witness collection exercise) against its likely utility, this temptation should be resisted.”

The judge's approach, concluding that it was foreseeable that the trial judge could come to either view, did not in my judgment go far enough in this particular case. It is convenient therefore, at this point, to consider the main criticisms of the survey.

The surveys were conducted under artificial circumstances. This is, of course, one of the underlying reasons why survey evidence of this kind is treated with special caution. However, the survey should be designed so that the artificiality of the survey situation does not introduce factors which would not be present in normal use of the disputed mark. Passing off is, after all, as Mr. Roberts submitted, a real world cause of action and the circumstances of the real world need to be present, so far as possible, in the survey's circumstances.

In the present case the activities of the parties are different. The appellants provide a TV channel and a website with information about that channel. The respondents provide an app which does not allow you to watch television but is for use while you are watching. It is not channel-specific. This is an important distinction which, to put it at is lowest, might be apparent to someone deciding to download or use the respondent's app. I do not think it is unreasonable to suppose that someone proposing to download the app would at least have some idea of what it is for.

Mr. Malivoire makes the obvious point that pilot questions 7 to 9, in using the words “relating to TV”, merely to invite the interviewee's attention to things within the world of TV but give no information as to the nature of the Zeebox service. The interviewee would therefore not be aware that Zeebox was a cross-channel service and unlikely to be associated with an individual TV channel. He says the questions give just the right amount of information to prompt speculation and maximise the number of interviewees who might associate Zeebox with Zee TV.

On the other side, the appellant's solicitor argues that the question is not objectionable because Zeebox is, in fact, an app relating to TV.

Mr. Malivoire's objection is, to my mind, a powerful point which immediately undermines the validity of the pilot survey results. Quite apart from the inappropriate stimulus given to the interviewee, the circumstances of the survey were very far removed from any real context in which a purchaser might encounter and download the respondent's app. Anyone who downloads an app, as was pointed out in evidence, would encounter information about what the app actually does. Interviewees, however, who are not even given the strapline "the best thing to happen to TV since TV", which might have given them a clue that the app did not have TV watching functionality, let alone any more detailed information about its cross-channel capability or specific functions.

It is no answer to any of this to say that Zeebox is in fact accurately but incompletely described as an app relating to TV, nor is it an answer that the interviewees were later asked about who they thought was responsible for the Zeebox. Any answer to this question would be preconditioned by the unfair nature of the earlier questions.

Mr. Malynicz showed us some of the answers from the interviewees which he considered helpful. Those, I am afraid tended to support the view that the respondents considered that they were being invited to speculate. Some of them, at least, were uncertain as to whether they could watch TV on the defendant's app, contrary to the fact. Respondent 12, for example, responded, in relation to question 9: "I haven't seen it before. I have never seen or heard of it. All I know is it is an app for watching TV, I think. I don't know anything else." Then, in response to the question, "Who do you is responsible?", the answer is: "Zee TV. I have heard of the Zee TV channel and it sounds similar. Maybe Apple. I don't know." In the absence of this particular respondent being called, it will remain a matter for speculation as to whether he or she thought they were being told that the defendant's app was something with which you could watch television, or whether that notion is something that occurred to them independently.

Mr. Malynicz accepted that the answers were not directly probative of deception. He suggested they showed a propensity to make a connection between the signs. I cannot accept that argument. If the questions are flawed and invite artificial speculation stripped of context, they do not show anything of value, whether described as a propensity to make a connection or deception.

It is also important to note the point already made, that the justification for the survey was the supposed special characteristics of the appellant's target audience. The survey itself has not been effectively targeted at that group. If the point was going to be that speakers of Hindi or other regional languages were going to be confused, then one would have thought some steps would have been taken to ask respondents if they spoke one of those languages, otherwise the survey will include many who, on the appellant's own case, are of marginal or no relevance. As Mr. Roberts submitted, inclusion of the base population of respondents who are not part of the relevant public may undermine the statistical basis of the survey. Whether this consideration on its own would be sufficient to undermine the validity of the survey is not something which we have to decide, although I am inclined to think that it might be.

It is true, as Mr. Malynicz submitted before the judge and before us, that no survey is perfect and it is hard to design any survey that does not to some extent invite speculation. Those points do not, however, justify the introduction of a survey with such obvious flaws. The present survey would leave the judge with the impossible task of having to discount them results by some measure, to allow for the fact that the interviewees were or might have been led to speculate that Zeebox was a channel, not a cross-channel app. I have no idea on what basis he would be able to do so.

This ground of appeal is based on the fact that the judge relied on a figure of £150,000 for the survey. This figure included the costs of calling witnesses derived from the survey when the application before him was, at least at the first stage, only for permission to adduce the survey supported by expert evidence.

I do not think this carries the appellants any further forward. Before the judge there was indeed the possibility of a later application for permission to call witness evidence derived from this survey. But even on the more limited figures for the survey evidence alone, one reaches a total of £100,000 or so. Mr. Malynicz submitted that this was a small proportion of the overall costs.

In my judgment the amounts of money to be spent on this survey are significant, particularly when viewed from the point of view of the respondents, who are a start-up company and of lesser financial strength than the appellants. This was a factor the judge could and should properly have borne in mind.

Because the judge may have been influenced by his assessment of the appellant's case on the merits and because I think he should have gone further and decided the issue of the value of the survey, I considerer that we are entitled to exercise his discretion afresh.

By far the most significant factor is that the survey is flawed in the way I have described. Any marginal value it might have is outweighed by the significant and, in my view, disproportionate cost. The appellants have other avenues open of them to prove their case. Against that, one must of course bear in mind that one is preventing the appellants from adducing some potentially relevant evidence. I consider that in this case the balance comes down heavily in favour of refusing permission for the survey evidence.

For my part I would not grant permission for this survey. If follows that I would dismiss the appeal.

I agree with Floyd LJ that the appeal should be dismissed, for the reasons that he gives. I add a few short observations of my own, not least because I gave the leading judgments in the two Interflora cases.

In Interflora 1 the issue was not whether Interflora should be entitled to rely on a survey, it was whether they were entitled to call evidence from internet users. It was in that context in which I said that if the case could not be proved without such evidence that would be a powerful reason for allowing it to be called. What I had in mind was that if the goods or services in question are specialised or esoteric, a judge might need evidence from users of the goods or services in order to be able to form his own view about whether a mark had been infringed. That is why I went on to say that no such evidence was necessary in the case of ordinary consumer goods or services. I was certainly not inviting a mini trial of the strength of the parties' cases at the stage of deciding whether or not to allow a survey to be conducted.

The second point relates to the survey itself. In the present case the interviewees were shown the mark and the sign devoid of all context. Even in the case of alleged trade mark infringement the acontextual comparison of mark and sign has been abandoned.

In the case of alleged passing off, the context in which the sign is used can be of critical importance. As Mr. Popat accepts in the course of his evidence, the show card which was used in conducting the survey is not how real world users would encounter Zeebox's app. Thus, as Floyd LJ points out, the survey got off on the wrong foot and also missed its target market by failing to ask anyone about what languages they understood.

Mr. Malynicz submitted that speculation is inherent in all surveys. Quite so. That is precisely why surveys should be admitted into evidence only in special cases. It may be that Mr. Roberts goes a little too far in saying that it is no longer a question of whether a survey will be held to be probative but a question of how long it will last before it is demonstrated to be worthless, but the general run of cases, both before and after Interflora, does tend to show that such surveys have very limited value.

Finally, Mr. Malynicz does not submit that the answers go to the question of deception. The survey goes to the question of whether the public or the relevant public has a propensity to make a connection between the mark and the sign. But, as Floyd LJ has said, that is not the real issue in cases of passing off, where the real issue is deception, so the fact that the survey only goes to a peripheral or subsidiary issue is itself a reason for degrading its value and not allowing it into evidence. I therefore agree that the appeal should be dismissed.

I agree with both judgments and dismiss the appeal also.