In trade mark infringement actions under s10(2) of the Trade Marks Act 1994, and in passing off actions, one of the key issues for the judge is whether or not the brand name, logo or get-up of the defendant’s product or service is likely to lead consumers to believe that the defendant’s product or service is the claimant’s, or is in some way associated with the claimant.
The question is one of fact and is said to be a ‘jury question’ for the judge. In other words, judges faced with this issue have to put themselves in the place of typical consumers in the relevant business area and decide whether a sufficiently significant number of those consumers would be confused or deceived. In areas where the judge is likely to have personal experience of the business area in question (eg ordinary retail shopping), strictly speaking, little or no evidence should be necessary. However, it is a brave (perhaps foolhardy) claimant who goes into court with no evidence at all. In specialist areas, of which the judge is unlikely to have any personal knowledge or experience, evidence of some kind may be crucial. It would also be necessary where the likelihood of confusion, while real, might appear counterintuitive to the judge hearing the case.
The best possible evidence is direct evidence; for example, consumers contacting the claimant (or others) in circumstances where they are clearly under the impression that the defendant’s product is the claimant’s product.1 However, direct evidence is notoriously difficult to come by and claimants commonly need to find other ways of demonstrating to the court that confusion and deception are likely as a result of the activities of the defendant.
Alternatives to direct evidence include trap orders, surveys and expert evidence. In IHL181 (p56) Lindsay Cook addressed the subject of trap order evidence. This article addresses the subject of survey evidence.
Commonly, the primary objective of the survey is to assist the judge in answering the ‘jury question’ as to the likelihood of confusion or deception. However, surveys for this purpose have received a mixed judicial response. If the survey simply confirms the judge’s first impression, it may be regarded as ‘helpful’.2 On the other hand it may be regarded as a waste of money as, if the survey does not confirm the judge’s first impression, it may be dismissed as unreliable for one or more of the multitude of reasons that survey evidence is normally discredited. Alternatively, albeit rarely, it may divert the judge’s mind away from the first impression.3
An additional benefit attached to surveys, which is also directed to assisting the judge in answering the ‘jury question’, is that they often uncover witnesses who can give valuable viva voceevidence from the witness box. In some cases, it is those witnesses who save the day when the survey is discredited. United Biscuits (UK) Ltd v Asda Stores Ltd  (also known asPenguin v Puffin) is an example. In Penguin the survey evidence had to be withdrawn in light of coding problems (as it did in Anheuser-Busch Inc v Budejovicky Budvar NP  (see p45), but the judge nonetheless found the evidence of witnesses identified from the survey as being of assistance:
‘That judicial first impression was largely confirmed by the oral evidence.’
In Neutrogena Corp v Golden Ltd  an in-store survey was conducted with no intention of making any evidential use of the survey as such, but simply to gather witnesses for the witness box. As seen below, it proved to be very successful.
IMPERIAL TOBACCO GUIDELINES
For a survey to have any chance of being given any weight by a judge, it must be properly planned and executed. The guidelines set out by Whitford J in his decision in Imperial Tobacco Group plc & anor v Philip Morris Ltd & anor  still hold well today:
- the questions must be appropriate;
- the interviewees must be representative of the market for the goods or services concerned, and in sufficient numbers to render the survey statistically significant;
- the interviewers must be properly instructed and follow the instructions given to them;
- the exercise must be conducted fairly; and
- the survey results must be properly codified.
Crucially, full disclosure of all the underlying documentation must be made to the other side, including the various sets of instructions given to the interviewers and those responsible for coding the answers, and, of course, the completed questionnaires.
In Imperial Tobacco it turned out that the claimants had carried out six different surveys, but little was known of the detail of some of them:
‘If survey evidence is to be of any weight at all it can only be of weight if in any case where, as here, a number of surveys have been carried out the plaintiffs give – and they must give this to the defendants before ever the action comes on – 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.’
Problems of this kind are less likely to arise in the future as a result of decisions of the late Pumfrey J and Rimer J (as he then was).4 Previously, parties conducted surveys at will, and sought and obtained, as a matter of course, leave to adduce expert evidence as to the conclusions drawn from them. Now, as a result of these decisions, the normal rule is that, before conducting a survey, the permission of the court must be obtained. The overriding objective of the Civil Procedure Rules, enabling the court to deal with cases justly, obliges judges to exercise case management, and the rationale behind these cases is that surveys, which are often found to be both unhelpful (or irrelevant) and very expensive, should not be undertaken without first having been judicially vetted.5 Even where they have been judicially vetted, it is not always the case that the trial judge will find them to be of any assistance. This is particularly likely where the judge vetting the survey is not a specialist in the area of law in question.6
Ordinarily, the planning and supervision of the survey will be undertaken by an expert in the field who will make a witness statement confirming what was done. Expert evidence will then be given as to the conclusions that can properly be drawn from the results of the survey.
A fundamental problem that is applicable to the vast majority of surveys and that will not be assisted by judicial vetting at the outset (even if that vetting is properly undertaken) is the manner in which the survey is conducted and the results coded. Will the interviewers follow their instructions to the letter? Will they record the answers verbatim? Will those responsible for the coding of the answers do their job properly?
When selecting experts for these tasks it is important that they are made aware of the Imperial Tobacco guidelines and, indeed, Laddie J’s cross-examination (more precisely, demolition) of the survey evidence in Anheuser-Busch (also known as Budweiser), extracts of which are quoted in Whitford J’s decision. In Budweiser, as is often the case when surveys go awry, the problem lay in the coding of the answers. The survey was intended to assess consumer perception as to the geographical origin of the beer. However, on checking the questionnaires it was found that the code for ‘made in USA’, for example, included answers such as, ‘I should say America, if somebody asked me I don’t know’, even though there was a code for ‘Don’t knows’.
However, even before becoming aware of potential coding problems, other issues commonly come to light: answers are recorded in shorthand form, or answers appear in such similar language it seems either the interviewers have substituted their own language for that of the interviewees or the interviewers have been lazy and, instead of asking people individually, have put the questions to a group of people.
The authors have experience of this first hand. In a recent case, a survey was formulated and organised by a very reputable market research agency. However, the agency had to subcontract the field work in some areas to other agencies. Passers-by were stopped in the street and, having cleared the opening filter questions, were brought into a hall for an interview. Respondents who gave the answers most helpful to the client were asked if they were prepared to be interviewed by solicitors, with a view to making witness statements. The authors were present for that purpose and were struck by the number of interviewees who expressed themselves in materially different terms when they were interviewed by solicitors. While this might have been due to the fact that they were talking to solicitors rather than market researchers, many made statements contradicting the answers recorded as having been given by them only a few minutes earlier to the market researchers. Whatever the reason, unless the evidence is rigorously tested in advance, there has to be a reasonably high level of risk that the witnesses, if called to give evidence, will come up with some unwelcome surprises in the witness box. If the survey evidence is the core of the case and the survey witnesses fail to give the anticipated answers, the case is likely to collapse like a house of cards.
Interestingly, this issue came up in the very recent A & E Television Networks LLC & anor v Discovery Communications Europe Ltd . The party seeking permission to conduct the survey wished to have it conducted by trainee solicitors rather than a firm of market researchers for reasons of reliability, even though it would be significantly more expensive to use that method. The judge permitted the survey to go ahead as planned, but restricted the recoverable to what would have been the price had it been conducted by market researchers.
FORMS OF SURVEY
Surveys come in all shapes and sizes, and are used for various purposes. This article is concerned solely with surveys designed to address the ‘jury question’ as to the likelihood of confusion or deception.
The most basic form of survey is a casual, statistically insignificant, straw poll of the kind regularly conducted by solicitors on receipt of instructions in a passing off or trade mark infringement action. It may be conducted simply to check whether there is a case or it may be with a view to testing questions for a more formal survey. In Diageo North America Inc & anor v Intercontinental Brands (ICB) Ltd & ors  (also known as Vodkat), a case in which Rouse acted for the claimants, a vacation student was sent out into the street with a bottle of Vodkat and told to stop passers-by to ask them what the product was. Twenty people were asked the question and all 20 answered ‘vodka’. Vodkat was not actually a vodka, and Rouse was reassured that its immediate impression was confirmed, an impression ultimately supported by the trial judge. However, there was never any intention of using the straw poll in evidence. When such informal surveys are sought to be used in evidence, they are normally discredited.7 The surveys fail the Imperial Tobacco guidelines on most (if not all) counts.
Even where the Imperial Tobacco guidelines have been followed as to the size of the survey and interviewee selection, the scope for variety is enormous. There are street surveys, doorstep surveys, telephone surveys, omnibus surveys, in-home interviews, consumer group discussions and tachistoscope tests.
Any and all of these can work in the appropriate circumstances and with the required evidential support. However, the writers are not aware of a tachistoscope test yet having succeeded in convincing a judge of its reliability. One such test was put before Whitford J, the architect of the Imperial Tobacco guidelines, in Laura Ashley Ltd v Coloroll Ltd . He described the test in the following terms:
‘It is a little bit of market research, a tool which may no doubt have great uses in some fields, but has not, in my experience, proved particularly fruitful in litigation in this sense that, by and large, the results have been of but little use to the court… there is a viewing cabinet, it has got an eye-piece and material is exposed on an illuminated screen by an operator. The operator, he or she, exposes the material in question first subliminally and then by increased intervals of milliseconds until the subject reaches a point where he or she can identify either shapes or colours or words, as the report puts it “supraliminally”.’
Whitford J was not impressed; nor was Vinelott J in United Biscuits (UK) Ltd v Burtons Biscuits Ltd . In United Biscuits the judge was exposed to three forms of survey:
- the straw poll;
- the tachistoscope test; and
- a professionally devised version of the straw poll on a larger scale.
The judge found all of them to be unreliable.
The essential problem in those days was that those ‘expert’ in the field were mainly expert in commercial market research, not market research for the purpose of legal proceedings. For commercial purposes, assessing consumer perception, there was apparently little need to discriminate between those firmly articulating a view and those guessing along the same lines. A rough and ready answer would suffice.
In Imperial Tobacco Whitford J commented:
‘However satisfactory market research may be in assisting commercial organisations as to how they can best conduct their business, they are by and large, as experience in other cases has indicated, an unsatisfactory way of trying to establish questions of fact which are likely to be matters of dispute.’
In The European Ltd v The Economist Newspaper Ltd , a case where the claimants were seeking to establish that the defendants’ European Voice was likely to be confused with their newspaper, The European, Rattee J expressed himself as follows:
‘The persons interviewed, most of whom had never seen a copy of European Voice, were asked, simply by looking at the top half of the front page of one issue, to speculate as to who published it – a question which they may well never have asked themselves or been concerned to know the answer to. This characteristic of the evidence is symptomatic of the artificiality of a survey of this type, in which members of the public are asked to speculate on the provenance of goods in a context quite outside the normal operation of the relevant market.’
In The European the relevant market was the purchase of newspapers. The judge was questioning the extent to which the ordinary purchaser of a newspaper is concerned as to its trade source.
A consequence of this understandable judicial reluctance to give weight to survey evidence was that market research ‘experts’ willing to put their necks on the line became more difficult to find.8 Solicitors tended only to proffer survey evidence if:
- that was the only available evidence; and
- there were ‘live’ witnesses to be obtained from the survey to support the conclusions drawn by the ‘expert’.
Survey evidence is usually the only option where the defendant’s product has not yet been fully launched, but then particular difficulties can arise in the formulation of the survey questions. The obvious temptation is to put the defendant’s proposed product (or a mock-up) to the interviewees and ask them if they have seen or purchased the product in the past – the hope being that they will say ‘yes’, thereby arguably demonstrating that they have mistaken the defendant’s product for the claimant’s product. This style of survey was castigated both at first instance and in the Court of Appeal in Scott Ltd v Nice-Pak Products Ltd :
‘The judge also said, however, that question 11 suffered from an even more serious defect, in that it proceeded upon a false basis.9 The false basis was that the make of baby wipes was available in this country for viewing or purchase when, of course, that was absolutely untrue. The question therefore… which is the most important question of the survey is, it seems to me, fundamentally flawed and is central to the effectiveness of the survey itself.’ (Per Fox LJ.)
Despite the large number of cases in which surveys have been found to be unsatisfactory, there have been cases where judges have found them useful. That said, they are usually cases where the survey results match the judge’s first impression, and where witnesses who have participated in the survey appear in the witness box and impress the judge.10
In GE Trade Mark  the survey was severely criticised by the judge at first instance, but in the Court of Appeal Salmon LJ, while accepting certain of the criticisms, nevertheless found it to be of probative value:11
‘Allowing for every criticism that was made of the survey evidence by the American Company’s statistical experts, the fact remained that at least 12% to 35% of the adult population would think that the GE mark on domestic electrical goods signified that the goods were made by the English company. This minimum of 12% of the adult population is, on any view, a very substantial number of persons.’
In Lego Systems A/S v Lego M Lemelstrich Ltd  the evidence before the court satisfied the judge that consumers would be likely to believe that Lego irrigation equipment was made by the toy brick company. The survey, which the judge found to be satisfactory, was a survey of 96 interviewees. Question 1 was:
‘Please will you examine this product [hand to respondent] and tell me, if you can, what firm or company you think manufactured it?’
Question 2 was:
‘What other goods are made by the same person or company?’
Despite a spirited attack by the defendants’ counsel on this survey, which must surely have failed the Imperial Tobacco guidelines on:
- too small a size to be statistically significant;
- asking the respondents a question that they might well not have asked themselves in the ordinary course of purchase; and
- inviting them to speculate;
the judge nonetheless found the survey both satisfactory in form and persuasive. Clearly, it confirmed the judge’s first impression.
In White Horse Distillers Ltd & ors v Gregson Associates Ltd & ors  Nourse J was inclined to discount the claimants’ survey in Uruguay that was presented to the court to support the contention that the defendant’s product was being passed off as Scotch whisky. For him, this view of the get up was counterintuitive. However, having heard from some of the survey participants, who were brought over from Uruguay for the trial, Nourse J was persuaded that the survey findings had merit:
‘I find that the evidence of confusion on the plaintiffs’ side, although not overwhelming and qualitatively far from perfect, is nevertheless substantial and cannot be disregarded.’
Kimberly Clark Ltd v Fort Sterling Ltd  involved the ‘Andrex’ and ‘Nouvelle’ brands of toilet tissue. The claimants said that the defendants’ (Nouvelle) packaging, which featured an offer reading, ‘Softness guaranteed (or we’ll exchange it for Andrex®),’ would lead consumers to believe that the two brands were associated. They introduced survey evidence to support the claim. Laddie J found the evidence to be of assistance, despite certain shortcomings:
‘It may well be that many of the interviewees were simply guessing an answer to a question they would never have asked themselves. This undoubtedly limits the value of the results. But, even subject to that caveat, it appears to me that the second survey illustrates the strong attractive power of the Andrex name, particularly when compared with Nouvelle. Although it does not prove that all the interviewees would have been deceived had they seen the second “offer-back” packaging in a supermarket, it is consistent with some of them being so deceived. Qualitatively it has some value.’
In Penguin, notwithstanding withdrawal of the claimants’ survey evidence, the judge found that the evidence of the witnesses derived from the survey was helpful.
In the authors’ view, the best form of survey is one where neither the interviewer nor the interviewee can have the faintest idea of the purpose behind the questions and where the bulk of the work is done by the interviewee and not the interviewer.
A few years ago a client introduced a new logo in the form of a stylised geometrically shaped ‘L’. The design featured the shape of a red triangle within the arms of the ‘L’. Bass, the owner of the famous red triangle trade mark, objected, claiming confusing similarity. It was decided to conduct a survey designed to determine whether the interviewees saw ‘L’s or triangles. The survey, the brainchild of Geoffrey Hobbs QC, involved a card featuring the ‘offending’ logo. Interviewees were asked to study it for a few seconds. They were then provided with a pen and told to turn the card over and draw a picture of what they had seen on the other side of the card. Happily for the client, they all drew ‘L’s. The case never reached court.
For one of Rouse’s clients, the firm is frequently required to investigate products made to look like whiskies to see if they are thought by consumers to be whiskies. A standard form of questionnaire could be devised to elicit the answer to the question, eg ‘What can you tell me about this product?’, but a survey along those lines is always going to be vulnerable to the usual criticisms. Rouse have found that a less vulnerable method is to stop people in the street, tell them that we are conducting some research into the pricing of spirits, give them a piece of paper and a pen, and ask them to list all the brands they can find on display in a shop nearby, identifying the type of spirit and the price. This involves finding a co-operative off-license stocking the target product. It is important, too, to ensure that the number of spirits on display is not so great as to render the exercise impractical.
If the shop does not stock the target product, it will be necessary to try and replicate a typical display. It is, of course, important that the display can be demonstrated to be typical (eg by comparing it with photographs of retail displays in other shops).
Another form of survey is one that was described by the trial judge in Neutrogena as a ‘witness collection programme’. It was conducted in 22 chemists’ shops around the country. Anyone picking up either of the parties’ products was given a questionnaire and some were interviewed by a solicitor on the spot. There was no intention of relying on any statistics. The sole purpose was to identify witnesses for the witness box. The judge commented that:
‘The court in a passing off case is not concerned with statistical precision. What it wants to know is whether or not there is a substantial degree of deception or confusion.’
The judge concluded:
‘The result of all this was that a significant degree of deception and confusion amongst some of the 131 instances was indeed shown.’
Survey evidence has generally always been viewed with suspicion by the judiciary. This scepticism, combined with the emphasis that is now given to efficient case management and control of costs, has made it difficult for survey evidence to be admissible. Furthermore, the newly required transparency with regards to survey evidence plans (the other side should also be provided with detailed plans, as well as the court) means that if the results of the survey are not deployed, inevitably, adverse inferences will be drawn by the court.
Survey evidence can still be of worth where a party believes the court may have insufficient evidence of confusion and, often, the oral evidence of witnesses who took part in the survey can be given weight by the court. However, if a party wishes to use survey evidence then they should be alive to the difficulties and risks involved. There are certain basic rules that a party should follow. The Imperial Tobacco guidelines remain a useful starting point, but any party proposing to undertake a survey must be warned that, however well it is constructed, executed and analysed, there will always be scope for attack. They should also be warned that the exercise of seeking permission to conduct a survey can itself, in certain circumstances, result in absurd levels of cost.12
Certainly, it would be prudent to undertake a pilot survey in advance of applying for leave to conduct a full survey just to check whether the exercise is likely to prove fruitful. When formulating the survey itself a party should obtain the assistance of an acknowledged expert to oversee the exercise, and provide expert evidence as to the survey method, its conduct and the conclusions that can be drawn from the results. Crucially, the format of the exercise should incorporate a means of identifying potential witnesses; in other words, a concluding question asking if the respondent would be prepared to be interviewed by solicitors in relation to the answers given and seeking the interviewee’s contact details for that purpose. For the survey to have any statistical significance, it is likely to have to involve at least 500 questionnaires and it would be remarkable if the opposing party, picking over those questionnaires, could not find some defect somewhere. Above all else the exercise should be conducted fairly.
If the judiciary exercise their case management powers properly, the risks associated with a poorly constructed survey, namely wasted cost and the opprobrium of the trial judge, should be reduced. Such surveys should not now be possible. However, even the best constructed survey will carry little weight with a judge if it is not properly executed.
- See Wilkinson Sword Ltd v Cripps & Lee Ltd & ors  FSR 16.
- ‘The relevance of this evidence is that it confirms the view which I had formed before seeing the market research documents,’ per Morritt LJ in Neutrogena Corp v Golden Ltd  RPC 473 at 504.
- See White Horse Distillers Ltd & ors v Gregson Associates Ltd & ors  RPC 61.
- See O2 Ltd & anor v Hutchison 3G UK Ltd  EWHC 2571 (Ch) and UK Channel Management Ltd v E! Entertainment Television Inc & anor  FSR 5.
- See Arden LJ in esure Insurance Ltd v Direct Line Insurance plc  EWCA Civ 842.
- See Diageo North America Inc & anor v Intercontinental Brands (ICB) Ltd & ors  EWCA Civ 920.
- See United Biscuits (UK) Ltd v Burton’s Biscuits Ltd  FSR 14.
- One of the leading market research agencies refused to get involved in this line of work ever again.
- ‘Have you ever seen or bought this make of baby wipe?’
- See Neutrogena.
-  FSR 418.
- See A & E Television Networks LLC & anor v Discovery Communications Europe Ltd  EWHC 1038 (Ch).