The Federal Court recently ruled against an attempt to introduce into evidence a poll in Tokai of Canada Ltd. v. Kingsford Products Company, LLC (2021 FC 782). In this decision, the court rejects the proposed evidence, finding it inadequate both because of the content of the questions and because the survey of members of the public was conducted online rather than in person. In his eyes, obtaining the impression of consumers sitting in front of a screen is simply not a good way to simulate how they would perceive a brand that they would come across in a store, for example.
The decision in question stems from an opposition to the registration of the trade-mark KING, particularly in relation to BBQs and devices allowing them to be lit. On appeal of the decision of the opposition court (before the Federal Court), we wanted to introduce into evidence the results of an online survey of a few hundred consumers, relating to the issue of confusion between the marks involved.
Unfortunately for Tokai, the court finally concludes that the new evidence we wanted to present is simply inadmissible. Admittedly, this type of evidence (surveys) can sometimes prove useful, for example, to help the judge decide on the question of possible confusion between two trade-marks. However, not all surveys are necessarily admissible in evidence, nor sufficiently relevant. As is often stated in case law, the introduction of a survey into evidence involves satisfying several relatively strict criteria, which the online survey conducted here was unable to do.
In normal times, i.e. when producing a traditional survey (conducted with respondents interviewed in person), the sample of respondents and the precise nature of the questions are all reasons why a prospective court might refuse to consider the survey results. Also, even in the traditional format, the survey is a perilous form of evidence to present in a Canadian court.
This time, to add to these difficulties, Tokai chose to innovate by attempting to produce a survey that was conducted online, asking targeted Internet users to answer an on-screen questionnaire. All the same, right? No, precisely, answers the Federal Court.
According to the court, the results of the survey in question cannot constitute admissible evidence. In particular, it is concluded that this particular survey would simply not meet the required eligibility threshold, since it would be too unreliable and would reflect questions leading to results whose validity can certainly be questioned.
With respect to reliability, the court concludes that because of the way the survey was conducted, it is doubtful that, if repeated, it would arrive at the same results. In other words, the exercise strays too far from a scientific experiment, which we usually try to approach with such surveys, in order to provide the judge with what would (ideally) amount to concrete data. objectives. For this reason, the survey does not present the degree of reliability required to be admissible in evidence.
Worse still, in the court’s opinion, the poll results are also downright invalid because the poll did not ask the right questions in the right way to an adequate sample of respondents. Here, there was therefore a problem not only with the content of the survey, but also with the way in which the exercise was conducted with the targeted consumers. Given the way the survey was conducted, the answers obtained are legally invalid.
In this regard, the court also finds fault with the wording of several questions. In this case, even though the application for registration covered BBQs and lighters for lighting them, the survey questions were limited to talking about lighters, without specifying what types of lighters we were talking about, not even if we were ordinary lighters or lighters intended for lighting a BBQ. Another example of a deficiency in this chapter was that the subject trade-mark was presented in a way that did not reflect the way a consumer would normally encounter it (in practice) in commerce, for example on shelves.
The nature of the questions asked would therefore be problematic. Thus, according to the court, we were allowed to doubt the validity of the answers that had been obtained, since we had simply asked the wrong questions. “Garbage in, garbage out”, to say (in a sense) the Federal Court here.
To add to these problems relating to the content of the questions themselves, it is also relevant to note that Tokai conducted its survey by administering it through a simple questionnaire displayed on the respondent’s screen (online). Because of this formula, the survey did not adequately simulate the encounter of a brand X product by a consumer, in the normal course of things, when shopping.
Moreover, the survey results were also invalid because the results had been pruned before counting them, by removing the answers of certain respondents who were deemed to have completed the questionnaire too quickly. Given the applicable test for brand confusion (which involves a typical consumer pressed for time), the court seized on the contradiction here to find that the results obtained by this survey are invalid and, therefore, inadmissible. Conversely, moreover, he also takes a dim view of the fact that some respondents were allowed to take several hours (or even days) to complete the survey, which is problematic in an attempt to obtain the “first impression” of consumers surveyed and undermines the credibility that can be given to the results. The method of conducting the survey and the time given to respondents to respond were therefore also an issue for the court.
In short, without going so far as to say that online surveys will never be admissible in such cases (concerning the likelihood of confusion between two marks), the Federal Court reminds us that surveys remain a very specific type of evidence. expert. In particular, if one hopes to reveal the typical impression of consumers through a survey, it is still necessary to carry out the exercise in a way that can simulate the real conditions in which a consumer would find himself when he actually encounters the brand in question. It is one thing to want to simulate the experience of an Internet user shopping online, but to reproduce what happens when a consumer walks along the aisles of a store is difficult to do by asking questions an Internet user simply sitting in front of his screen.
Consequently, we can conclude from reading this decision that it may generally be difficult to conduct a (valid) online survey in such a case. When the evidence aims to eventually help the judge rule on the likelihood of confusion between two trade-marks appearing on store shelves, a survey conducted by means of a questionnaire simply displayed on the screen (online) can often be problematic.
Me Sebastien Lapointe founded and works for his firm Techtonik Legal inc. since 2020. He practices mainly in business law, intellectual property and technology law.