Consumer Uncertainty in Trademark Law: an Experimental Investigation

• Publication year: 2023

Consumer Uncertainty in Trademark Law: An Experimental Investigation

Barton Beebe

Roy Germano

Christopher Jon Sprigman

Joel H. Steckel

CONSUMER UNCERTAINTY IN TRADEMARK LAW: AN EXPERIMENTAL INVESTIGATION Barton Beebe Roy Germano Christopher Jon Sprigman Joel H. Steckel^* Abstract

Nearly every important issue in trademark litigation turns on the question of what consumers in the marketplace believe to be true. To address this question, litigants frequently present consumer survey evidence, which can play a decisive role in driving the outcomes of trademark disputes. But trademark survey evidence has often proven to be highly controversial, not least because it has sometimes been perceived as open to expert manipulation. In this Article, we identify and present empirical evidence of a fundamental problem with trademark survey evidence: while the leading survey formats in trademark law test for whether consumers hold a particular belief, they do not examine the strength or the varying degrees of certainty with which consumers hold that belief. Yet as the social science literature has long recognized, the strength with which consumers hold particular beliefs shapes their behavior in the marketplace, and thus it should also shape how trademark disputes play out in the courtroom. Through a series of experiments using the three leading trademark survey formats (the so-called Teflon, Eveready, and Squirt formats), we show the remarkable degree to which these formats as conventionally designed overlook—or suppress—crucial information about consumer

[Page 490]

uncertainty. We further demonstrate how low-cost, easily administered, and relatively simple modifications to these formats can reveal that information.

We explain both the practical and theoretical implications of our findings. As a practical matter, trademark survey evidence that shows only weakly held beliefs (or that does not even test for belief strength) should not, without more, satisfy a litigant's burden of persuasion on the issue addressed by the survey. Furthermore, in line with courts' growing efforts in intellectual property cases to tailor injunctive relief, survey evidence showing only weakly held mistaken beliefs may provide courts with the opportunity to fashion more limited forms of relief short of an outright injunction. As a theoretical matter, trademark survey formats that reveal the true extent of consumer uncertainty in the marketplace may finally force trademark law and policy to confront normative questions it has long left unanswered going to exactly what kind of harm trademark law is meant to prevent.

Table of Contents

Introduction..........................................................................................491

I. Background: Framing Consumer Uncertainty in Trademark Law...............................................................................................497

A. Trademark Law and the Probabilism of Consumer Beliefs ...... 497
B. The Social Science of Consumer Beliefs .................................. 501
C. The Treatment of Consumer Uncertainty in the Early History of Trademark Survey Evidence.................................................... 506

II. Experiments Assessing the Effect of Consumer Uncertainty on Trademark Survey Methods .................................................... 510

A. The Teflon Genericism Survey Format.................................... 511

1. The Origins of the Teflon Survey Format .......................... 511
2. Measuring Uncertainty in Teflon Surveys ......................... 513

a. Alternative Teflon Method A: Following Up on the Three-Answer Forced Choice Question ................................. 514
b. Alternative Teflon Method B: Measuring Uncertainty on a Seven-Point Likert Scale ............................................. 519
c. Alternative Teflon Method C: Measuring Uncertainty Using a Continuous Scale ..................................................... 521
d. Evaluating the Relative Merits of Alternative Teflon Methods A, B, and C ................................................... 522

B. The Squirt Confusion Survey Format ...................................... 523

1. The Origins of the Squirt Survey Format........................... 524
2. Measuring Uncertainty in Squirt Surveys .......................... 526

[Page 491]

a. Alternative Squirt Method A: Following Up on the Three-Answer Forced Choice Question................................. 527
b. Alternative Squirt Method B: Measuring Uncertainty on a Seven-Point Likert Scale............................................. 530

C. The Eveready Confusion Survey Format ................................. 533

1. The Origins of the Eveready Survey Format...................... 533
2. Measuring Uncertainty in the Eveready Surveys................ 534

III. Implications................................................................................538

A. Consumer Uncertainty and the Expansion of Trademark Rights ..................................................................................... 539
B. Incorporating Consumer Uncertainty into Trademark Practice .................................................................................. 543

Conclusion.............................................................................................546

Introduction

What do consumers believe?¹ That is the deceptively simple question on which nearly every important issue in trademark litigation turns. For example, the basic test for trademark infringement asks: Is it likely that a substantial proportion of consumers believes that goods bearing the defendant's trademark originate from the plaintiff?² The basic test to determine if an asserted mark is "generic"³ and thus unprotectable asks: Is it likely that most consumers perceive the mark as referring to an entire category of goods rather than to a specific producer of goods within that category?⁴

[Page 492]

Trademark law focuses so intently on the question of what consumers believe because consumer beliefs—or courts' estimates of those beliefs—so often determine whether there is a trademark property right, whether that right has been infringed, and what the remedies for that infringement may be. At every stage of trademark litigation, consumer perception is at the center of the inquiry.⁵

To understand what consumers believe, courts in trademark disputes often consider consumer survey evidence.⁶ Litigants hire survey experts to survey a sample of the relevant consumer population and then testify about their findings. These surveys can play decisive roles in the outcomes of trademark disputes. Consider the recent, closely watched case United States Patent & Trademark Office v. Booking.com B.V.⁷ There, the Supreme Court held that adding ".com" to a generic word could result in a distinctive, protectable mark if consumers perceived it as indicating the source of products.⁸ The Court's ruling in the case

[Page 493]

relied heavily—too heavily, Justice Breyer suggested in dissent⁹ —on the respondent's survey showing that 74.8% of the survey's respondents perceived BOOKING.COM as a brand name.¹⁰

Though its use is widespread, consumer survey evidence has long been a controversial form of proof in trademark litigation, and for good reason. First, it has sometimes been attacked for being susceptible to expert manipulation. Judge Posner repeatedly questioned the value of "tendentious expert testimony"¹¹ offered by trademark survey experts who are "prone to bias"¹² and took pains in his opinions to expose what he called the "tricks of the survey researcher's black arts."¹³ Other judges have lamented that "as is so often the case in high-stakes litigation, highly qualified experts have presented dueling [survey] reports that reach significantly different results."¹⁴

Second, and more importantly for our purposes here, trademark surveys tend to elide complex normative and empirical questions that underlie trademark law and policy. Among those questions is how trademark law should understand consumer uncertainty, i.e., the varying degrees of confidence consumers have in their beliefs in the marketplace. As currently constituted, the leading survey formats provide no way for respondents to indicate the strength with which they hold a particular belief—what the social science literature calls "belief strength."¹⁵ While it is true that each format now conventionally provides respondents with the option to respond "Don't know/No opinion," decades of experience show that typically only very low proportions of respondents resort to that response.¹⁶ The remaining response options reduce to a starkly binary

[Page 494]

choice: respondents must indicate that they perceive the asserted mark as a designation of source or not; they must indicate that they are confused as to source or not. The result is that trademark surveys fail to measure potentially valuable empirical information about consumer uncertainty. That is a real shortcoming because social science evidence strongly suggests that the likelihood that a person will take some marketplace action on the basis of a particular belief is tied to the strength with which that belief is held.¹⁷ For example, consumers who strongly believe that a product comes from a company whose products they like are more likely to purchase based on that belief, when compared to a consumer who holds that same belief but only weakly. Thus, the relevance of belief strength to the goals of trademark law should be obvious. And yet had trademark survey experts deliberately set out to conceal consumer uncertainty and dodge the important theoretical questions that uncertainty raises, it is not clear they could have come up with better formats than those currently in use.

In this Article, we argue that trademark law and trademark consumer surveys in particular must finally acknowledge consumer uncertainty—the reality that consumer beliefs are not binary but held at...