¶ 1         Evidence of actual confusion can present one of the most formidable challenges to an attorney representing a trademark defendant. In a case where all other circumstances point to a finding of non-infringement, significant evidence of actual confusion dramatically alters the equation. Such evidence can shift the focus away from other factors impacting the infringement analysis and drive litigation into discovery and evidentiary hearings, in contrast to the increasingly frequent tendency to determine infringement before discovery has commenced— from a paper record on a preliminary injunction.

¶ 2         Yet, plaintiffs' attorneys often draw a false sense of security from such evidence, which is often not what it appears, at first glance, to be. Indeed, such evidence is typically characterized by vagueness, ambiguity, untrustworthy sources, and a host of unknown circumstances when sources are not identified or do not testify. Furthermore, such evidence is frequently characterized by circumstances that do not reflect the commercial realities of how marks are used and how consumers react to them, as well as by widely varying quantities of incidents that are difficult to assess when viewed in light of various factors that may lead one to expect more or less of such evidence to have surfaced.

¶ 3         From the perspective of courts, such evidence has a powerful allure. In particular, in trademark litigation, judges (and sometimes juries) are given the difficult task of determining whether one mark is so like another that its use should be enjoined—based on predictions of whether the coexistence of those two marks is likely to confuse the consuming public rather than on subjective impressions about the mere similarity of the marks and whether the senior user should be afforded protection for the sheer creativity of the mark. This task requires suspending one's instinctive feelings about similarity of the marks and fairness to the creator and substituting a judgment about a likely commercial impact that, in reality, is frequently unclear and requires drawing inferences that inevitably involve highly subjective judgments.¹ Evidence of actual confusion holds the allure of solving that problem by giving courts a window into what would appear to be actual commercial effects of the competing marks.² Accordingly, courts frequently issue pronouncements that give preferred or favored status to evidence of actual confusion as a method of proving likelihood of confusion.³

¶ 4         Nevertheless, commentators have recognized that evidence of actual confusion comes in many different varieties with widely varying degrees of probative value. Furthermore, judicial treatment of many types of such evidence has been inconsistent.⁴ Thus, however powerful the allure of an objective measure for determining infringement, evidence of actual confusion is often not such a clear indicator of anything, and the interpretation of such evidence often requires highly subjective judgments. Indeed, judicial pronouncements on actual confusion are often sweeping and, yet, diametrically opposed to each other, while failing to articulate sufficient rationale to provide meaningful guidance for future courts. Accordingly, one commentator has concluded that many types of evidence accepted by courts as proof of actual confusion are not objective indicators of actual injury to the trademark owner.⁵

¶ 5         It is the thesis of this article that greater consistency, coherence, and predictability and, ultimately, greater objectivity can be attained in cases involving actual confusion by articulating a set of factors that distill the many ways in which different types of evidence have produced different results under varying circumstances. This approach is consistent with trademark law's overall factor-based approach to determining the effect of different marks under varying commercial circumstances.⁶ Distilling such factors requires reviewing a vast body of caselaw, and this article undertakes such a review. Section I introduces the place of actual confusion in the context of various courts' standards for determining infringement. Section II discusses the need for evidence of actual confusion, including the extent of any such need and the quantity of evidence necessary to establish actual confusion. Section III reviews a host of circumstances that may impact the value of evidence of actual confusion. Section IV appraises various tests that courts have articulated for assessing evidence of actual confusion. Finally, in an effort to steer a course between the extreme, vague, and conflicting rhetoric that courts use to describe how such evidence should be assessed, Section V proposes a multi-factor test to provide clearer guidance.

I.    The Status of Actual Confusion as a Factor in Determining Trademark Infringement

¶ 6         Trademark infringement is determined by the "likelihood of confusion" test.⁷ This test examines whether consumers are likely to be confused between products identified with the respective parties' marks. This likelihood of confusion is determined based upon the commercial realities of how the products are sold. To determine whether consumers are likely to be confused by the respective parties' marks on their products under the proper commercial circumstances, courts weigh a variety of non-exclusive factors.

¶ 7         Although the factors typically enunciated in the test for infringement vary from jurisdiction to jurisdiction, they tend to include:

(1)    The similarity of the respective parties' marks;

(2)    The similarity of the parties' marketing methods;

(3)    The similarity of the parties' channels of distribution for their goods or services;

(4)    The level of sophistication of the prospective purchasers for the respective parties' goods or services, and the degree of care used in purchasing such goods or services;

(5)    The source-designating strength of the mark sought to be protected;

(6)    Where the second-comer's goods or services differ from the first-comer's, the likelihood that prospective purchasers of the second-comer's goods or services would expect the first-comer to have expanded its marketing or sponsorship into the second-comer's field;

(7)    The extent of overlap in the parties' geographic markets and whether the prior user is known by its mark in geographic markets in which it does not actually sell its goods or services;

(8)    Whether the second-comer intended to copy the first-comer's mark in order to cause confusion or deceive; and

(9)    The degree of actual confusion that has surfaced as a result of the two parties' respective marks.⁸

¶ 8         Among other things, these factors are designed to ensure that the determination of whether one mark infringes another is not merely a subjective decision about whether the second-comer's mark so closely resembles the first-comer's mark that it would be unfair to allow the later entrant to continue. Trademark law does not confer a "right in gross" to a mark but, rather, confers rights limited to what is necessary to prevent consumer confusion and protect the good will of the trademark owner.⁹ Thus, trademark infringement typically is not determined based solely upon a side-by-side comparison of the two marks.¹⁰

¶ 9         In determining whether confusion is likely, evidence of actual confusion is often given considerable weight and preferred status. Some courts view evidence of actual confusion as "persuasive" evidence of likelihood of confusion.¹¹ Other courts have used a variety of adjectival phrases to describe the utility of the presence or lack of evidence of actual confusion.¹² Some decisions even give evidence of actual confusion status as a preferred method of proving infringement.¹³ Courts are receptive to inferring likely confusion from instances of actual confusion because instances of actual confusion are seen as representative examples of how others in the marketplace are likely to react.¹⁴ As put by one court: "[a]ctual confusion is... extremely probative of the likelihood of confusion, in that proof of confusion by actual customers is strong evidence that a hypothetical reasonable customer would also be deceived."¹⁵ As another court explained: "[i]f enough people have been actually confused, then a likelihood that people are confused is established."¹⁶ Not all evidence of actual confusion, however, is entitled to preferred or favored status. Rather, such evidence must first survive the particular level of scrutiny given to evidence of the types at issue by the jurisdiction in question.¹⁷

¶ 10         Notwithstanding this seeming judicial preference, statements giving undue weight to evidence of actual confusion are not consistent with other judicial pronouncements regarding the infringement test as a multi-factor test or with other judicial pronouncements giving preferred status to other factors in that test. For example, a number of courts have stated that the similarity of the parties' marks is the most important factor in determining likely confusion.¹⁸ Intuitively, that approach makes sense. Where the parties are competitors for the same type of goods, at least one court has gone one step further to opine that, in such cases, courts generally need not look beyond the marks at issue.¹⁹ Indeed, try as the courts may to escape the appearance of subjective comparisons between the parties' marks by using a multi-factor test, nearly all judicial recitations of factors place the similarity of the marks first in the list of factors. Judges reason that, ultimately, it is the similarity of the marks that will drive consumer confusion. Yet, in instances where the defendant intended to trade on the plaintiff's good will, that factor has also been recognized to have potential preeminent status.²⁰ Courts in the Seventh Circuit have deemed three factors—similarity of the marks, the defendant's intent, and actual confusion—to be the most important.²¹ In other cases, courts have found different subsets of the likelihood of confusion factors to be of paramount importance, with emphasis on the similarity of marks.²² Two commentators have deemed the most important factors to be the strength of the plaintiff's mark, the similarity between the marks, and the defendant's intent.²³ In any event, actual confusion is but one of several factors relevant to determining the likelihood of confusion.²⁴

¶ 11         As previously noted, any list of factors to be weighed in determining the likelihood of confusion will necessarily be non-exhaustive.²⁵ The non-exhaustive nature of such factors flows from the fact that the likelihood of confusion standard is designed to assess what impression consumers are likely to draw in actual marketplace settings, which are infinitely varied. Thus, any circumstance that would impact a consumer's impression may be relevant to determining whether that consumer is likely to be confused between two parties' marks.²⁶

II.    The Need for Evidence of Actual Confusion and Quantum of Proof Necessary to Establish Actual Confusion

¶ 12         This section explores two related areas of tension in cases involving actual confusion. The first area of tension concerns whether evidence of actual confusion is needed to establish infringement. Paradoxically, evidence of actual confusion is not needed, as a matter of law, but its absence can have great significance—in some cases, so much so that evidence of actual confusion is, in reality, needed. The second and related area of tension concerns the extent of proof needed to establish actual confusion—a point on which cases are widely inconsistent. First, this section will explore the reasons why such evidence generally is not needed. Second, this section will discuss why evidence of actual confusion, in some circumstances, may be needed. Third, this section will explore the extent of the plaintiff's need to demonstrate actual confusion and the quantum of proof needed to establish that factor.

¶ 13         It is generally accepted as established law that a party need not produce evidence of actual confusion in order to obtain injunctive relief.²⁷ The reasons for this rule are many. First, the Lanham Act and state laws governing trademarks and unfair competition only require the plaintiff to prove a likelihood of confusion. Second, courts will prevent harm that is likely before such harm actually occurs. Third, it is often less harsh to the infringer to nip the infringing activity in the bud before harm actually occurs. Fourth, for a variety of reasons, it is difficult to prove actual confusion—often, so much so that requiring such proof would gut the protections afforded by trademark law.

¶ 14         The Lanham Act establishes a mere likelihood of confusion as the threshold for proving that a second-comer has violated the statute. Specifically, a trademark owner may establish infringement of a registered mark by proving that the second-comer's designation "is likely to cause confusion, or to cause mistake, or to deceive...."²⁸ Similarly, with respect to unregistered marks, the first party to obtain rights in one of the marks at issue can establish unfair competition by showing that the second-comer's designation "is likely to cause confusion, or to cause mistake, or to deceive...."²⁹ As discussed above, actual confusion is but one factor in the multi-factor test used to determine the likelihood of confusion. Accordingly, requiring evidence of actual confusion would be inconsistent with the statutory language requiring only likely confusion.³⁰ Requiring evidence of actual confusion would likewise conflict with common law and state statutory standards, under which only a likelihood of confusion is required to prove infringement or unfair competition.³¹ Similarly, requiring evidence of likely confusion would elevate one of many factors potentially probative of actual confusion over others that may be more significant in the circumstances of particular cases.³²

¶ 15         A principal reason for not requiring evidence of actual confusion to prove likely confusion—and one of the principal policy reasons for requiring only proof of likely confusion to prevail—is the injustice that would result from allowing likely confusion to come to fruition before the infringement can be stopped. As stated by the First Circuit:

There is ample reason, at the incipiency of an alleged infringement, in a suit seeking injunctive relief, for a plaintiff to argue and a court to rule that the similarity of marks is such that confusion is all too likely to ensue. Plaintiff should not be expected to stand by and await the dismal proof.³³

¶ 16         Other courts have agreed that a trademark plaintiff should not be penalized for acting to protect its rights before it has been damaged by actual confusion.³⁴

¶ 17         One commentator has argued that placing greater emphasis on actual confusion will not cause trademark owners to stand by and suffer irreparable harm while they wait for such evidence to surface because equity affords relief in advance of such harm.³⁵ The problem with this argument is that, in trademark cases, a likelihood of success on the merits establishes the irreparable harm that is critical to preliminary injunctive relief.³⁶ Actual confusion is one of the factors in determining the merits of the claim. Accordingly, ignoring the factor at the preliminary injunction phase will likely affect the substantive standard on the merits by reducing the degree to which published decisions addressing the infringement standard reflect the need for such evidence. More importantly, trademark cases are increasingly resolved in the context of motions for preliminary injunctions. Therefore, ignoring actual confusion at that stage—while giving greater emphasis to the factor on what is supposedly the final merits—will reduce the need for such evidence in the context in which it most often counts while introducing dissonance between the infringement standards at different phases of the litigation and, thus, further inconsistency in judicial treatment of this factor.

¶ 18         Both substantive trademark law and the law of equity governing injunctions encourage a trademark owner to act early so as not to prejudice the defendant. In applying equitable standards governing injunctions, courts will balance the harm to the defendant from granting the injunction requested with the irreparable nature of the harm to the plaintiff that will occur if the injunction is not granted.³⁷ In applying this standard, courts have recognized that it may be less harsh to enjoin a defendant that has invested fewer resources to date in an allegedly infringing brand than to wait until the defendant has invested more substantial resources in a brand and then enjoin its use.³⁸

¶ 19         As a matter of substantive law, similar concerns underlie the laches defense.³⁹ Although delay for the purpose of obtaining evidence of actual confusion may be asserted as a counter to a laches defense, such a delay may introduce uncertainty into the proceedings by prompting the defendant to raise such a defense.⁴⁰

¶ 20         Numerous courts have recognized that it is often difficult for a trademark plaintiff to obtain evidence of actual confusion.⁴¹ This difficulty is attributable to a number of circumstances. First, the infringement may be in its incipient stage. Second, the relevant markets may be sufficiently separated or dispersed, and the market penetration of the infringer may be sufficiently low, as to make confused consumers difficult to locate. Third, consumers may lack the incentive to contact the trademark plaintiff to report confusion. Fourth, consumers may be so successfully deceived that they do not realize their confusion, and thus have nothing to report. Fifth, evidence of confusion may not be sufficiently documented by the parties or others. The difficulties associated with obtaining evidence of actual confusion have led one commentator to conclude that most markets simply do not generate real evidence of actual confusion.⁴²

¶ 21         In order for an infringing mark to cause confusion among consumers, that mark must be presented to a sufficient number of consumers who are familiar with the plaintiff's mark and who are sufficiently interested in the defendant's presentation that they either make a mistaken purchase or form a mistaken impression. In many instances, at the time an infringement action is adjudicated, there will not have been sufficient opportunity for evidence of actual confusion to surface. For example, if an infringer has only recently commenced sales of goods with the mark in question or has only recently begun efforts to promote the sale of goods with the mark, the infringer may not have reached a sufficient audience to cause actual confusion.⁴³ This problem is particularly pronounced where the demand for the goods or services in question is seasonal and any coexistence between the parties has not occurred during the season of peak demand.⁴⁴ The problem is further compounded where the plaintiff moves quickly to obtain a preliminary injunction or where the court must rule on the issue before discovery has been conducted.⁴⁵ Consequently, a number of courts have recognized that the absence of evidence of actual confusion will have little or no probative value where the extent of infringement has not become sufficiently widespread for actual confusion to manifest itself.⁴⁶

¶ 22         One commentator has argued that placing greater emphasis on the actual confusion factor will not make it unfairly difficult for the plaintiff to prove infringement. According to this argument, a plaintiff should be able to present such evidence with ease, given the varieties of anecdotal evidence cited by courts and the ability of parties to conduct surveys.⁴⁷ As will be discussed at length in this article, the many varieties of anecdotal evidence regarding actual confusion lend themselves to inconsistent judicial treatment. Anecdotal evidence is often unreliable and open to highly subjective judgments. As for surveys, they are not truly evidence of actual confusion, but, in any event, can be extremely expensive and time-consuming. Furthermore, surveys may be impractical where there is no easy way to get widespread access to an appropriate universe of people.⁴⁸

¶ 23         The marketplace conditions under which the respective parties are using the designations at issue may also make it unlikely that evidence of actual confusion would have surfaced. For example, if the parties' past operations have been centered in geographically distinct areas, one may not expect evidence of actual confusion to have surfaced.⁴⁹ Indeed, even within an overlapping geographic territory, where the parties' marketing efforts are focused on different groups of individuals, there may not have been sufficient exposure of both parties to a common customer group to yield evidence of actual confusion.⁵⁰ Similarly, if the parties are not direct competitors, or if there are differences in the parties' channels of trade, actual confusion may take longer to surface.⁵¹

¶ 24         In many cases, it will be difficult to gather evidence of actual confusion because consumers will lack the incentive to report mistaken purchases.⁵² For example, many will not do so where the price of the goods in question is low and consumers lack ready means of contacting a corporate plaintiff.⁵³ They are particularly unlikely to do so where there are no discernible differences in the quality of the respective parties' goods.⁵⁴ Some consumers may even be willing to accept the infringing item in place of the genuine article sought after learning of their mistakes.⁵⁵ Indeed, even if a confused consumer is dissatisfied with the goods in question, if the price of those goods is sufficiently low, the most efficient way to register that dissatisfaction may be simply to avoid the product in the future.⁵⁶ In some instances, the consumer may not learn of the mistake until long after the transaction.⁵⁷ Some consumers may also feel too embarrassed to report their mistakes.⁵⁸

¶ 25         Consumers may also lack incentive to report complaints if they believe either that the problem is already being addressed or that they will become embroiled in legal proceedings to address the problem. For example, the Fifth Circuit held that a lack of consumer complaints indicating actual confusion should not be construed against the plaintiff where the plaintiff filed a much publicized infringement action shortly after the alleged infringement began. In such instances, the court held that a consumer aware of the action would have no reason to complain.⁵⁹ Of course, such a well publicized infringement action would probably reduce the likelihood of consumer confusion following publicity regarding the action.⁶⁰ In addition, the possibility of becoming a witness in litigation may deter many from coming forward to report confusion or to testify concerning their reports.⁶¹

¶ 26         Notwithstanding these issues, a lack of consumer incentive to report confusion is not ubiquitous. For example, a consumer who mistakenly purchases a subscription to the wrong magazine is likely to complain about that mistake after receiving the first issue.⁶² Nevertheless, it has been recognized that, in many retail situations, evidence of actual confusion is unlikely to surface.⁶³

¶ 27         Where trademark infringement is so successful that consumers are deceived into believing that the defendant's goods are those of the plaintiff and never realize the mistake, the deceived consumers may have widely different reactions, depending upon the circumstances. Where the consumer has a reason to communicate with the company thought to be the producer of the goods in question, the consumer may mistakenly communicate with the wrong company.⁶⁴ Similarly, where consumers realize that they have made mistaken purchases, they may still be confused about where to report such mistakes.⁶⁵ In contrast, where the consumer has no reason to communicate with the mistaken producer of the goods, the consumer cannot be expected to report to either company a mistake unknown to that consumer.⁶⁶ Instances where confused consumers have no incentive to communicate with the producer of the goods are particularly likely where there are no discernible differences between the qualities or characteristics of the plaintiff'goods and those of goods sold by the defendant.⁶⁷ Furthermore, courts have held that businesses that purchase low-priced goods for use in their establishments cannot be held to a more discriminating standard of actual confusion than an ordinary purchaser.⁶⁸ In some cases, the consumer may not have a meaningful opportunity to inspect the goods or discover any mislabeling.⁶⁹

¶ 28         Even where confusion is reported, it may have been insufficiently documented such that evidence of actual confusion may be unavailable for use in court. For example, in Wynn Oil Co. v. American Way Service Corp.,⁷⁰ the defendants argued that marketplace conditions in that case should have caused actual confusion to become manifest if, in fact, any confusion was likely to occur. Because actual confusion did not surface, the defendants maintained that their mark was unlikely to cause confusion. The Sixth Circuit rejected this argument as incongruous with the defendants' own recordkeeping practices.⁷¹ Indeed, where instances of confusion are likely to be reported to the defendant and not the plaintiff, it would be particularly unfair to place the plaintiff's prospects of success at the mercy of the defendant's recordkeeping practices.⁷² Nevertheless, assuming that personnel employed by the defendant testify truthfully, it should be possible to determine whether such evidence once existed and was not kept. Where relevant evidence was destroyed, such destruction may be remedied by laws governing spoliation of evidence.⁷³ However, it is also possible that consumers may report their confusion to non-party retailers who may not provide such information to the parties in litigation.⁷⁴

¶ 29         Although evidence of actual confusion is not required as a general rule, the lack of such evidence can play an important role in the adjudication of many cases. Indeed, the lack of such evidence can significantly detract from—or be fatal to—the plaintiff's claim for at least two reasons. First, if other circumstances weigh against the plaintiff, a court may look for evidence of actual confusion to establish a likelihood of confusion that other factors have not. Second, the circumstances under which the parties have been operating may indicate that evidence of actual confusion should have surfaced.

¶ 30         The first reason why courts may require evidence of actual confusion to rule in favor of the plaintiff is relatively straightforward: a court may not be persuaded by the other factors that confusion is likely. In such instances, only evidence of actual confusion can save the plaintiff's case—assuming that such evidence is powerful enough and that other factors do not undercut such evidence.⁷⁵ Similarly, if the court assesses the other likelihood of confusion factors with respect to a new technological practice, a new medium for communicating or a new context for determining infringement, it may find that the other factors do not provide a sufficient basis for predicting likelihood of confusion in this new context and may look for evidence of actual confusion to guide its determination.⁷⁶

¶ 31         A second reason why the lack of actual confusion evidence may have significance is the logical outgrowth of rationales for not requiring evidence of actual confusion. Although there are a variety of circumstances that make evidence of actual confusion difficult to obtain, there are cases where those circumstances do not exist. Indeed, cases exist where such evidence should have surfaced if, in fact, it is likely that consumers will be confused.

¶ 32         Notably, where the defendant has achieved a meaningful commercial presence for a sufficient duration and there have been no, or few, reported instances of actual confusion, that silence may indicate that confusion between the respective parties' marks is unlikely.⁷⁷ The lack of such evidence may be particularly telling where the marks are so similar and the plaintiff's mark is so strong that one would expect consumer confusion, but for some other countervailing factor (such as use of the respective marks on different products).⁷⁸ In recognition of the importance of the length of time in which the marks have had overlapping usage in valuing evidence of actual confusion, at least one court has recast the actual confusion factor as "duration of use without actual confusion."⁷⁹ Another court has gone so far as to articulate the length of time during which there has been no actual confusion as a stand-alone factor in the likelihood of confusion test, separate from the actual confusion factor.⁸⁰

¶ 33         By contrast, where the defendant has achieved a meaningful commercial presence for a sufficient duration and where there have been many reported instances of actual confusion, those instances will be highly probative of a likelihood of confusion.⁸¹ Indeed, some courts have afforded greater weight to relatively few reported instances of actual confusion where alleged infringement is in its incipient stages.⁸² Other courts have held that incipiency merely reduces the weight of the inference to be drawn from the lack of evidence of actual confusion.⁸³

¶ 34         Bound up in the question of what effect should be given to the lack of evidence of actual confusion is the question of how much evidence of actual confusion is necessary to establish actual confusion. This subsection introduces the issue of "quantum of proof" and then reviews various circumstances that can affect how much proof of actual confusion may be needed.

¶ 35         Courts follow widely divergent paths in assessing how much evidence is necessary to establish actual confusion. On the one hand, many courts frequently discount evidence of actual confusion as insubstantial or de minimis.⁸⁴ The likelihood of confusion standard is, itself, only satisfied by a showing with respect to an appreciable number of consumers, and that showing must be a showing of probable confusion rather than merely possible confusion.⁸⁵ On the other hand, a very influential decision by the Fifth Circuit established long ago that "very little proof of actual confusion would be necessary to prove the likelihood of confusion."⁸⁶ Consistent with this language, some courts appear willing to accept any amount of evidence of actual confusion as sufficient to establish the factor in the plaintiff's favor.⁸⁷

¶ 36         It is hazardous to extrapolate from the cases a particular number of instances that must be present for actual confusion to be established; what constitutes the necessary critical mass will vary widely from case to case. The Eleventh Circuit observed: "There is no absolute scale as to how many instances of actual confusion establish the existence of that factor."⁸⁸

¶ 37         Determining how much evidence is necessary is particularly complicated by the fact that, where circumstances indicate that evidence of actual confusion should have surfaced in a meaningful way, the plaintiff may need to produce a higher quantum of proof not merely to demonstrate that the actual confusion factor favors the plaintiff, but also to maintain the viability of its infringement claim.⁸⁹ For example, where the alleged infringement occurred over a period of years, a dearth of incidents of actual confusion may mean that the plaintiff cannot establish the factor in its favor and also become a critical factor leading the court to find confusion not likely.⁹⁰ As an extreme example, where an alleged infringement involved a highly successful television show spanning 225 episodes, court held the plaintiff's inability to produce more than six witnesses claiming to have been confused over the course of eleven years to be dispositive that confusion was not likely.⁹¹

¶ 38         Similarly, if a long period of time has elapsed between commencement of the alleged infringement and adjudication, or there has been significant market penetration for a sufficient period of time, the critical mass required to establish the actual confusion factor will be higher.⁹² For example, where a large number of transactions and widespread advertising had taken place over a period of three years, and where the purportedly confused persons could not be identified as potential customers, nineteen instances of reported confusion were deemed "negligible" and insufficient to establish actual confusion.⁹³ Yet, courts generally do not distinguish between how much evidence is necessary to establish the factor and how much evidence is necessary to overcome circumstances that may make it necessary to maintain viability of the claim.

¶ 39         There are numerous and varied factors that impact how much evidence of actual confusion may be needed, including: whether there is a relative or statistical benchmark for measuring the quantity of incidents of confusion; circumstances showing the duration, extent, and nature of any coexistence between the parties; the ability and incentive of consumers to report confusion; the stage of the proceedings; the parties' recordkeeping practices; the types of persons confused and the degree of their confusion; the quality of the evidence of confusion; and what inferences can be drawn from the other likelihood of confusion factors.

¶ 40         Courts are unlikely to require that evidence of actual confusion surpass any "statistical significance" threshold before being credited.⁹⁴ Indeed, one court held it erroneous to discount evidence of actual confusion as de minimis (in the absence of a full survey) because principles of statistical significance do not apply outside the context of a formal survey.⁹⁵ Nevertheless, courts may be persuaded that a particular quantity of incidents can be assessed by comparing the number of incidents to a numerical measure showing the extent of opportunity for confusion to surface. One yardstick that has been used to measure whether actual confusion is appreciable is the extent of the plaintiff's sales.⁹⁶ If the number of instances of actual confusion appears small relative to the total number of consumer transactions or communications during the same time period, such evidence may be discounted.⁹⁷ However, any such comparison between incidents and market presence will necessarily involve a highly subjective judgment.

¶ 41         As the foregoing discussion suggests, whether and to what extent evidence of actual confusion will be required to establish infringement may turn on the duration, extent, and nature of the parties' coexistence. For example, the absence of such evidence may assume particular importance where the products have been quickly sold in the same channels of trade and extensively promoted in the same advertising media.⁹⁸ Even within a narrow geographic market, meaningful evidence of actual confusion may be expected where there has been extensive publicity and deep market penetration.⁹⁹ Another factor lending greater importance to the absence of evidence of actual confusion is the defendant's engagement in comparable conduct with respect to other brands in the past, without generating evidence of actual confusion in those contexts.¹⁰⁰ Courts may give enhanced weight to an absence of evidence of actual confusion where the plaintiff has actively engaged in efforts to try to gather evidence without success.¹⁰¹

¶ 42         By contrast, where the parties have not had the opportunity to compete head-to-head so as to expose a significant number of actual or potential customers to the competing marks and products, little or no confusion would be expected, and, therefore, relatively few instances would be given greater weight.¹⁰² For instance, in Daddy's Junky Music Stores, Inc. v. Big Daddy's Family Music Center,¹⁰³ the Sixth Circuit held that the lower court erred in finding that a single instance of actual confusion could not increase the likelihood of confusion. In that case, when a consumer requested a catalog from the plaintiff, the consumer mentioned that the consumer was familiar with the plaintiff's store at a location that was, in fact, the defendant's store. Although the parties had been in business for years and the plaintiff had an extensive customer list in the location of the defendant's store, the Sixth Circuit nonetheless held that the case lacked the circumstances that would lead one to expect more incidents of actual confusion. The court found such circumstances lacking because there was no evidence of overlap in the audience exposed to the parties' marketing. The court observed:

¶ 43         Bearing in mind that a successful Lanham Act plaintiff only must show a sufficient potential of confusion, not actual confusion, the fact that some confusion already has occurred favors plaintiff at least to some extent. The District Court therefore should reevaluate upon remand the significance of the single instance of actual confusion.¹⁰⁴

¶ 44         Courts must be careful to look past superficial evidence of the extent of the parties' presence to determine whether there has truly been competition or other meaningful coexistence in a setting that could generate consumer confusion. For example, parties may coexist for a substantial period of time without direct competition. Thus, in CAE, Inc. v. Clean Air Engineering, Inc.,¹⁰⁵ the Seventh Circuit held that twenty-five years in which no confusion was reported was not sufficient to demonstrate that confusion was not likely as a matter of law. In CAE, the parties had some product overlap but were not direct competitors, and the court noted the difficulty of discovering instances of actual confusion.¹⁰⁶ Nevertheless, courts should consider direct competition in terms of competing sales or coexistence in the same channels of trade as well as advertising and publicity that reaches a common audience.¹⁰⁷

¶ 45         Courts should also recognize that parties may coexist for a substantial period of time under conditions not likely to produce confusion and then change practices in a way likely to cause confusion. For instance, in Keycorp v. Key Bank & Trust, the court determined that the plaintiff's case was enhanced by the fact that actual confusion only surfaced after the defendant recently modified its name in a way likely to cause consumer confusion between the parties.¹⁰⁸

¶ 46         An additional factor that may affect the quantum of proof a court may require is the likelihood in the particular industry at issue that consumers will come forward with such evidence. As discussed earlier, consumers may be unlikely to demonstrate their confusion for a variety of reasons.¹⁰⁹ In circumstances where few instances of actual confusion would be expected, as with inexpensive items or items with limited availability, courts have relied upon relatively few instances of actual confusion to find a likelihood of confusion.¹¹⁰ In some industries, for example, trademark owners may have evidence about the timing and nature of customer interactions that will explain a lack of reported confusion. For example, in CFM Majestic, Inc. v. NHC, Inc.,¹¹¹ where both parties manufactured iron stoves for residential heating, the court found the lack of actual confusion to be excused by a number of factors, including the fact that consumer feedback about stoves typically does not occur until after the stove has been installed through a burn-season.¹¹²

¶ 47         Factors that make it less likely that consumers will report confusion, however, must be weighed against other factors indicating that some reports of confusion should have surfaced. For example, one court has held that the passage of sufficient time with sufficient competition between the parties and no actual confusion can override any consideration that would otherwise be given to the difficulty of obtaining evidence from retail consumers.¹¹³

¶ 48         Another factor impacting the required quantum of evidence is the stage of the proceedings, which affects expectations as to how much evidence of actual confusion would ordinarily be available. For example, if a case is heard on a motion for a temporary restraining order or a preliminary injunction, frequently the plaintiff will not have had an opportunity to conduct discovery to uncover all of the actual confusion that can be located. Thus, in reversing a district court order denying a motion for preliminary injunction, the Seventh Circuit found just four instances of actual confusion to be probative and distinguished cases finding isolated instances of actual confusion to be insufficient on the ground that those cases "found that such isolated instances were insufficient to establish actual confusion in the context of a trial, not in a preliminary injunction hearing."¹¹⁴ Indeed, at least one court has determined that a lack of evidence of actual confusion may be excused by the fact that the case was adjudicated on a motion for summary judgment.¹¹⁵

¶ 49         As discussed above, poor recordkeeping practices may make it unlikely that reports of consumer confusion would be available. Accordingly, in assessing the quantum of proof of actual confusion, courts should examine the degree to which consumers communicate directly with the parties and the degree to which the parties typically preserve records of such communications.

¶ 50         Factors other than the opportunity for evidence of actual confusion to be generated and ascertained can also impact the requisite quantum of such evidence. As the Fifth Circuit explained: "Perhaps as important as, and helping to explain the various interpretations of the relevance of, the number of instances of confusion are the kinds of persons confused and degree of confusion."¹¹⁶ Incidents involving actual or potential purchasers are likely to receive greater weight, even if those incidents are few in number.¹¹⁷

¶ 51         However, when actual consumers mistakenly communicate with the wrong party when seeking to make a purchase, courts may afford greater significance to relatively few such incidents.¹¹⁸ For example, in Roto-Rooter Corp. v. O'Neal,¹¹⁹ where the trial court rejected four instances of mistaken purchases by the plaintiff's customers as statistically insignificant, the Fifth Circuit reversed, based largely on that evidence of actual confusion. The Fifth Circuit noted that, while such evidence was not needed, it was the best evidence of infringement.¹²⁰

¶ 52         By contrast, relatively few incidents of confusion have been discounted when they did not involve individuals responsible for purchasing decisions and reflected initial assumptions rather than sustained confusion.¹²¹ Similarly, several incidents of actual confusion were discounted as de minimis when they were among "random acquaintances."¹²² However, when actual consumers mistakenly communicate with the wrong party when seeking to make a purchase, courts may afford greater significance to relatively few such incidents.¹²³ Nevertheless, relatively few such incidents may be insufficient where they consist either of inattentiveness that does not reflect confusion engendered by the marks or of mistakes that accompany the introduction of the defendant's mark but do not persist through sustained competition.¹²⁴

¶ 53         If confusion is transitory and prompted by an event that has passed (as is the case with a vague impression left by a competing trademark or confusion that may be generated by the introduction to the market of a similar product), then courts are likely to give such incidents little weight.¹²⁵ One court distinguished between incidents of actual confusion that are "isolated" and "unrelated" from "an ongoing pattern which has occurred in the past and is likely to continue."¹²⁶

¶ 54         As will be discussed in Section III, evidence of actual confusion comes in widely varying degrees of quality and reliability. Courts have pointed to defects in quality of the evidence as a reason to require a greater quantum of proof. For example, where evidence of (relatively) few incidents consists of out-of-court statements that are subject to interpretive problems or pertain to individuals who entertained confusing impressions only briefly, the quantum of such evidence may not be sufficient, particularly if the marks in question are highly dissimilar.¹²⁷ In one case, nine letters and a memorandum of a telephone call, all of questionable probative value, were held insufficient in light of the testimony of a former manager at the plaintiff's only directly competing location that he did not observe a single instance of confusion over two years.¹²⁸

¶ 55         Methods for bolstering the credibility of witnesses testifying about actual confusion may enhance the quality of the evidence of actual confusion and reduce the quantum of such evidence. For example, small numbers of reported instances of actual confusion have been given much weight when recounted in testimony by an employee of the defendant.¹²⁹ Commercial circumstances showing the reasonableness of customer confusion may also bolster evidence of actual confusion.¹³⁰

¶ 56         Any circumstantial evidence of actual confusion may also give greater significance to relatively few instances of actual confusion that can be demonstrated by direct proof. For example, in Baker v. Simmons Co., in addition to four of the defendant's employees who testified about instances of confusion that they had observed, the plaintiff introduced literature the defendant had supplied to its employees instructing them in how to handle confused inquiries.¹³¹ The court found this literature to be highly probative of the "frequency" with which such confused inquiries arose.¹³²

¶ 57         Notwithstanding an absence of evidence of actual confusion—even in the face of sustained competition—any negative inference may be overcome by strong proof of likely confusion through other means. For example, where the parties' products were sold side-by-side for six years without any discernable actual confusion, despite the plaintiff's engagement of paid investigators, the close similarity in names, product types, and channels of trade combined to lead to a finding that confusion was likely.¹³³ Indeed, a strong showing as to other likelihood of confusion factors may warrant giving greater credence to types of actual confusion evidence that might otherwise be tenuous, such as inquiries regarding affiliation.¹³⁴ Moreover, survey evidence indicating confusion may overcome any negative inference arising from a lack of actual confusion or warrant giving greater significance to relatively few incidents of actual confusion.¹³⁵

¶ 58         Where, however, other likelihood of confusion factors point strongly to the conclusion that confusion is not likely, then a higher quantum of proof of actual confusion may be needed.¹³⁶ For example, with regard to highly descriptive marks, relatively few instances of actual confusion may be discounted as not probative.¹³⁷

III.    Assessing the Value of Evidence of Actual Confusion

¶ 59         This Section explores in more depth some of the problems surrounding evidence of actual confusion. First, this Section will review the various types of evidence of actual confusion—both circumstantial and direct. Second, this Section will explore the vexing question of whose confusion is relevant to establishing actual confusion. Third, this Section will review a host of shortcomings in many types of evidence of actual confusion.

¶ 60         Actual confusion need not be proven through direct evidence. Although rarely employed, circumstantial evidence may be used when available.¹³⁸ For example, circumstantial evidence can be found when the defendant implements business practices for addressing confusion if it appears the practice was designed to respond to past instances of reported confusion. In Baker, defendant furnished sales employees with a booklet designed to provide those employees with answers to anticipated questions from prospective customers, including whether the defendant was the plaintiff.¹³⁹ The court found "little question" that the inquiry was "based on past salesmen's experience" and observed that "[t]he fact that the very first question to be 'anticipated' by the [defendant's] salesman dealt with the relationship between [the plaintiff] and [the defendant] speaks volumes as to the doubtful state of the public mind on this question and on the frequency with which the question was raised."¹⁴⁰

¶ 61         It should be noted, however, that, although precautions taken by a defendant against future confusion may have been prompted by past instances of actual confusion, such precautions may also have been prompted by the defendant's concern about likely confusion based upon other factors, such as similarities in the marks and the parties' goods. Although circumstantial evidence is inherently subject to such ambiguity, under either interpretation of the previous example such evidence is strong proof of infringement. By contrast, some types of ambiguous circumstantial evidence may not be sufficiently probative of actual or even likely confusion. For instance, although a decline in the plaintiff's sales may be consistent with actual confusion, such a sales decline may not show actual confusion if other potential causes exist—even where the defendant's sales increase in the same proportion as the plaintiff's decline.¹⁴¹

¶ 62         Just as the defendant's business practices may give rise to circumstantial evidence of actual confusion, so too can its conduct with respect to the litigation. For example, if the defendant refuses to respond to discovery requests specifically targeting actual confusion, the court might infer that the defendant is hiding evidence of actual confusion.¹⁴² Circumstantial evidence can also be used to demonstrate that confusion is not likely. For example, if the parties exist amicably during a period of time and refer business back and forth between the two parties, a court might conclude from such coexistence that consumers were not confused.¹⁴³

¶ 63         Direct proof of confusion takes two forms: live testimony by a witness purporting to have been confused and out-of-court statements or conduct by persons indicating confusion. The latter category includes statements authorized by the persons who were purportedly confused (such as letters), instances of purported confusion documented by others (such as in a business record), and testimony by witnesses who heard confused statements of others or observed confused conduct of others.

¶ 64         Out-of-court statements or actions indicating confusion come in many varieties. At the most general level of categorization, such statements or actions include: inquiries about whether a relationship of identity, source, sponsorship, or affiliation exists between the parties;¹⁴⁴ misdirected communications or deliveries;¹⁴⁵ mistaken purchases;¹⁴⁶ misattribution in published sources (i.e., by newspapers and periodicals);¹⁴⁷ mistakes by advertisers;¹⁴⁸ mislabeling or other mistakes by distributors or retailers;¹⁴⁹ mistakes by less formal resellers;¹⁵⁰ mistakes by actual or potential investors;¹⁵¹ prospective and current employees who confuse the parties;¹⁵² current employees or agents who confuse parties' goods or services;¹⁵³ mistakes by parties' service suppliers;¹⁵⁴ misdirected website searches information about the parties;¹⁵⁵ and assumptions of mistaken connection in terms of identity, source, sponsorship, or affiliation.¹⁵⁶ Often, out-of-court statements or actions are the subject of live testimony. In such cases, the two forms of direct proof converge as the witness testifies first as to a previously confused state of mind and then as to contemporaneous statements or actions that flowed from that confusion.¹⁵⁷

¶ 65         Not all confusion is relevant. At the most obvious level, it is not relevant if a person is confused where that person has no interaction with the trademark owner's business and has no interest in purchasing the type of goods sold by that business and where that person's confusion cannot otherwise be viewed as indicative of the reactions of persons in such relevant categories.¹⁵⁸ However, categories of relevant confusion are more difficult to define. Many incidents of confusion are subject to interpretation in light of the credibility of the witness, the extent to which circumstances about the incident are known, and the nature of those circumstances. As a result, some of the decisions crediting and discounting various types of actual confusion have apparent inconsistencies. More fundamentally, the leading sources of guidance—the Lanham Act and the Restatement—do not make clear whose confusion is relevant.

¶ 66         Originally, the Lanham Act protected trademark owners from a narrow category of likely confusion among actual purchasers as to the origin of the trademark owner's goods and services. Specifically, the statute previously defined actionable infringement as the use of a designation that is "likely to cause confusion or mistake or to deceive purchasers as to the source of origin of such goods or services."¹⁵⁹ In 1962, the statute was amended to its current form, which eliminates any reference to purchasers or source in defining infringement. The current version defines infringement as a use that is "likely to cause confusion, or to cause mistake, or to deceive...."¹⁶⁰ The legislative history pertaining to this amendment makes clear that Congress intended to include "potential" as well as actual purchasers within the category of persons whose confusion is relevant.¹⁶¹ The statutory language, however, is far broader and appears to encompass any type of confusion.¹⁶² Indeed, a number of courts have cited the 1962 amendments as justification for recognizing confusion among classes of persons beyond actual and potential customers and types of confusion other than source confusion.¹⁶³ By contrast, other courts have made various pronouncements that have hewn more closely to the legislative history and that have indicated that the amended Lanham Act redresses only confusion among actual or potential purchasers.¹⁶⁴

¶ 67         The Restatement of Unfair Competition attempts to synthesize the authorities and update the types of confusion deemed actionable. Specifically, the Restatement provides that confusion will not be actionable unless it "threaten[s] the commercial interests of the owner of the mark."¹⁶⁵ This actionable confusion "is not limited to the confusion of persons doing business directly with the actor."¹⁶⁶ Nevertheless, the Restatement's discussion of harm to good will is limited to the plaintiff's reputational damage among potential purchasers. The Reporter's Note provides that only where there is likely harm to the trademark owner's commercial interests will confusion among non-purchasers be actionable.¹⁶⁷ Indeed, in a parenthetical reference, the Reporter's Note states that "trademark infringement" protects only against mistaken purchasing decisions and not against confusion generally."¹⁶⁸

¶ 68         A case summarized by the Reporter's Note is an influential Second Circuit decision on the matter of confusion, Lang v. Retirement Living Publishing Co.¹⁶⁹Lang, which followed an earlier draft of the Restatement, concerned misdirected telephone calls made by individuals who were neither "purchasers [n]or prospective purchasers."¹⁷⁰ Since the callers were not actually prospective purchasers, the court noted that "there is no reason to believe that any confusion represented by the phone calls could inflict commercial injury in the form of either a diversion of sales, damage to goodwill, or loss of control over reputation."¹⁷¹

¶ 69         Consistent with the language in Lang, a number of decisions in the Second Circuit and elsewhere have indicated that the scope of actionable confusion extends to persons whose views may impact the good will of the trademark owner but have differed in the degree to which they are willing to extend this category to classes of persons other than actual or potential purchasers.¹⁷² The category of persons whose views may impact the good will of the trademark owner remains largely unsettled. A number of courts have extended the types of persons whose confusion is deemed to be relevant because of harm to the trademark owner's good will to include investors, prospective employees, suppliers, distributors, and service companies.¹⁷³ Other categories of persons whose views may impact the good will of a company but who also may be seen as more akin to actual or potential customers include persons seeking service with respect to items already purchased and persons who return a product or direct complaints about a product to the wrong company.¹⁷⁴ Such persons may not have been induced to make the initial purchase out of confusion, but they may be confused with respect to potential future purchases. As such, any dissatisfaction they experience may lead them to give misleading views to other potential purchasers. Another type of confusion that may similarly result in misleading information being disseminated to prospective customers is mistaken identification of the parties in the media.¹⁷⁵ Confusion among those in the media may also be seen as relevant insofar as less sophisticated consumers might be confused more easily.¹⁷⁶

¶ 70         Another form of confusion that may be harmful to the good will of the trademark owner is so-called "post-sale" confusion, which may occur among members of the public who view the trademark on goods previously purchased—even if the purchaser of the goods was not confused. Even with respect to this type of confusion, courts are split between those that limit actionable post-sale confusion to people who are, themselves, likely purchasers of the goods or services at issue and those that recognize actionable harm to the trademark owner (i.e., people who knowingly purchase counterfeit goods bearing prestigious marks).¹⁷⁷

¶ 71         In justifying the expansion of trademark protection to prevent non-purchaser confusion that is harmful to the trademark owner's good will, one court explained:

A business reputation, earned over years of conscious effort, deserves protection from those who would tarnish it unfairly by using a confusingly similar trade name. Entities such as suppliers, distributors and service companies certainly consider the business reputation of a firm with which they do business; and especially for transactions based on assessment of good credit. If a likelihood of confusion among suppliers, distributors, service companies or other relevant classifications exists, it deserves consideration as a relevant factor in determining infringement.¹⁷⁸

¶ 72         Nevertheless, this rationale will not protect against confusion among members of the public at large who do not have actual, or potential, purchasing or other relationships with the trademark owner's business or do not have the potential to give the knowing purchaser of counterfeit goods the recognition associated with the trademark owner's prestige.¹⁷⁹ Indeed, some decisions employing language seeming to require confusion among actual or prospective purchasers to establish cognizable confusion may be seen as driven by alleged instances of confusion among members of the public at large—who obviously would not be in a position to affect the trademark owner's good will unless they could be shown to be actual or prospective purchasers.¹⁸⁰ Given the difficulty of identifying what segment of the public falls within those whose views may impact the good will of the trademark owner, one measure may be whether the alleged incidents of confusion reflect the type of mistaken impression that the plaintiff seeks to prevent.¹⁸¹

¶ 73         As courts have been slow to recognize actionable confusion among persons other than actual or prospective customers and the types of persons who may impact the trademark owner's good will remains relatively undefined, courts have divided on whether to recognize confusion among certain types of persons is relevant. For example, courts have had particular difficulty with the question of whether confusion among members of the trade is relevant. Some courts will credit such confusion.¹⁸² Others will not.¹⁸³

¶ 74         While a number of courts have indicated that the appropriate perspective for determining likely confusion is that of an actual or potential customer,¹⁸⁴ other courts have found confusion within the industry to be equally relevant.¹⁸⁵ One way in which confusion within the industry can have relevance is if confused sales persons unwittingly pass one party's products off as those of a competitor.¹⁸⁶ Under another view, there may be multiple "consumers" for a given business, including consumers within the trade.¹⁸⁷ For example, magazine publishers have multiple classes of consumers, including advertisers, subscribers, and newsstand purchasers.¹⁸⁸ Other courts have held that, where a product has a mixed buying class of professional buyers and retail consumers, testimony by professional buyers as to the absence of confusion should not be dispositive, and that the court should determine confusion from the perspective of the ordinary consumer.¹⁸⁹ In addition, some persons within the trade may not be actual or prospective purchasers. Nevertheless, under an approach that credits any confusion among those whose views could impact the trademark owner's good will—a category that could potentially include suppliers, shippers, distributors, providers of credit, investors, prospective employees, employment agencies, and industry rating groups—confusion of many different types of persons in the trade should be actionable. Indeed, some courts have held actual confusion among members of the trade must be credited because of the 1962 amendments to the Lanham Act, which expanded the likelihood of confusion inquiry to encompass non-purchasers.

¶ 75         There may be reason for caution in assessing evidence of confusion among members of the trade, particularly where it takes the form of out-of-court statements by those who should not be so easily confused. Casual questions by sophisticated persons asking when the plaintiff went into the business of producing a particular product that happens to be a product produced by the defendant and not the plaintiff may well be evidence of humor or sarcasm by one who knows better rather than evidence that the person asking such questions was actually confused.¹⁹⁰ Where the statement is made by a particularly knowledgeable and sophisticated person in the trade and the parties have been in competition for a sufficient amount of time,, it may be improbable that the person had acquired an insufficiently precise understanding of the differences between the two parties and their respective marks before making the statement at issue.¹⁹¹ Furthermore, allegedly confused communications by persons within the trade may be discounted when they do not arise from the context in which the marks are presented to an audience of potential consumers.¹⁹² In addition, persons within the trade may have particular biases due to relationships with one of the parties.¹⁹³

¶ 76         Nevertheless, each allegedly confused individual must be assessed in light of that person's credibility and the circumstances surrounding the incident in question. Indeed, under another view, actual confusion among particularly sophisticated and knowledgeable members of the trade may augur greater confusion among ordinary, less sophisticated and knowledgeable consumers.¹⁹⁴ Ultimately, determining whether such alleged confusion is relevant depends on assessing the credibility of the particular accounts in light of circumstances known and observed as to the particular witnesses and the circumstances that purportedly generated the witness's alleged confusion.¹⁹⁵

¶ 77         Although the infringement standard measures the likelihood of consumer confusion under marketplace conditions, actual confusion among actual or potential purchasers need not necessarily occur in the context of a sale. For example, courts have held that actual confusion is relevant when it flows from advertising not presented at the point of sale.¹⁹⁶ In addition, actual confusion by persons seeking service for products previously purchased is similarly relevant: confusion in that context may harm the plaintiff's good will and reputation, even where the services offered by the respective parties are purely ancillary to the product sold.¹⁹⁷ Finally, actual confusion may occur when persons arrive at one party's location in response to another's advertisements or attempt to exchange one party's merchandise at the location of another.¹⁹⁸

¶ 78         Although there may be some dispute as to whether certain types of persons should be deemed relevant in determining the presence or absence of actual confusion, there is relatively little dispute that the confusion of some persons is more probative than that of others. Those whose confusion generally carries the most weight tend to be actual or prospective purchasers.¹⁹⁹ The category of prospective purchasers, however, is extremely broad and can encompass those who are likely to make a transaction imminently and those who may make a transaction at some point in the future. Similarly, the category of persons whose views may affect the good will of the trademark owner encompasses persons who may have existing and prospective relationships with the trademark owner. Those who have existing relationships and have made serious mistakes based on confusion may be more relevant than prospective purchasers who are not proximate to a potential transaction.

¶ 79         Within a given category of persons, however, there is a hierarchy of incidents that are given differing degrees of weight by courts. At the most persuasive end of the spectrum are mistaken purchases.²⁰⁰ For those whose views may impact the good will of a company, the equivalent to a mistaken purchase would be a mistaken transaction. According to one commentator, mistaken purchases are afforded the most weight because they do not present any question about the relevance of the person or setting, because there is typically little vagueness, ambiguity, or doubt about such evidence, and because there are no issues as to the admissibility of such evidence if the confused purchaser testifies.²⁰¹ According to the same commentator, the next most probative types of evidence are those in which the confused person has affirmatively taken some action to manifest confusion, such as by making complaints to the wrong party, returning merchandise to the wrong party, or seeking product repairs or service from the wrong party.²⁰² Included in that category are instances in which the wrong party is named in a lawsuit²⁰³—although, unless the lawsuit is filed by a consumer, the relevance of the person and setting would be more attenuated and would seem to flow only from the high degree of care expected in such circumstances. Lesser in probative value are misdirected communications, which may be attributable to carelessness, to secretarial or bureaucratic errors, to other circumstances not related to confusion between the marks at issue, or to unknown causes.²⁰⁴ Still further down in this ranking of types of incidents of confusion are inquiries about affiliation, which may show some doubt but also tend to show some understanding as to the distinctions between the parties and could be motivated by mere curiosity—in marked contrast to a mistaken purchase by which the confused person acts to his or her detriment.²⁰⁵ Finally, of uncertain status are incidents of purported confusion by friends, family, and acquaintances of the trademark owner, who may be biased and who may not be representative of actual or prospective customers or of other persons whose views would affect the good will of the trademark owner.²⁰⁶ As one court has observed: "Attestations from persons in close association and intimate contact with its (the trade-mark claimant's) business do not reflect the views of the purchasing public."²⁰⁷

¶ 80         Both cases limiting relevant confusion to actual or prospective purchasers and cases expanding those limits to include persons whose views may impact the good will of the trademark owner appear to share a common feature—namely, that certain types of confusion are entitled to no weight at all. Nevertheless, the Fifth Circuit has expressed in dicta that trial courts cannot "total[ly] disregard... evidence of actual confusion..., regardless of the identity of the person confused."²⁰⁸ Yet, any probative value afforded to the confusion of persons not deemed to be relevant is likely to be minimal, at best.²⁰⁹ A comment in the Restatement appears to support that actual confusion in an irrelevant context is "entitled to little weight," as opposed to no weight.²¹⁰ Nevertheless, confusion among irrelevant members of the general public should be afforded weight, only when there is a finding that such confusion augurs confusion in other relevant persons by reason of some meaningful, shared characteristic.

¶ 81         Not only must the person be relevant, but the setting in which the person is confused must similarly be relevant. Thus, transactions that do not replicate actual marketplace conditions may not be probative of likely confusion. For example, if a product is usually purchased by sight on a self-serve basis, a mistake by a sales clerk in offering the wrong product when the product was requested verbally may not be indicative of likely consumer confusion.²¹¹

¶ 82         The status of the purportedly confused person may play an important role in instances where the type of evidence at issue is less reliable. For example, some courts have discounted or discredited mere inquiries about origin, sponsorship, or affiliation where the plaintiff could not establish that the persons making such inquiries were potential purchasers.²¹² The connection between type of person and type of evidence is particularly important with respect to inquiries, as it is often all too easy for a trademark plaintiff to produce friends and acquaintances who can attest to idle curiosity about a relationship between the parties. Thus, it has been held that less weight should be given to inquiries representing short-lived confusion among persons casually acquainted with a business and more weight to lasting confusion among actual customers.²¹³

¶ 83         Problems relating to the relationship between status of the purportedly confused person and the type of evidence of confusion can also arise with respect to misdirected communications. Specifically, where persons or entities doing business with the plaintiff direct communications to the wrong location, such mistakes may be attributable to clerical assistants, administrative personnel, or independent delivery services, rather than to relevant persons.²¹⁴ Such incidents are more likely to be credited as strong evidence of actual confusion where the plaintiff can demonstrate—from the face of the communication or otherwise—that the mistake was made by a relevant person who sent the communication rather than by an assistant or third party.²¹⁵

¶ 84         With any type of confusion that is subject to challenge or likely to receive less weight, such evidence can often be bolstered when accompanied by other forms of confusion.²¹⁶ By contrast, where such episodes are isolated, they are less likely to be credited.²¹⁷

¶ 85         Several problematic features of evidence of actual confusion have already been mentioned at various points. This Subsection explores those features more comprehensively. Roughly, the various problems that detract from the value of evidence of actual confusion can be categorized as those inherent in the procedural posture of the litigation, those relating to the law of evidence, and substantive issues that impact the value of such evidence.

¶ 86         There are at least two types of procedural issues that contribute to the inconsistencies and difficulties in cases involving evidence of actual confusion. The first type flows from the nature of preliminary injunctive proceedings. The second flows from the standard of review applicable to the likelihood of confusion finding.

¶ 87         A threshold issue that detracts from a clarity in treatment of evidence of actual confusion is how the preliminary injunction context (in which so many trademark decisions are reported) affects the evidentiary standard. Courts are afforded a wide degree of discretion in granting preliminary injunctions and in what types of procedures should attend their determinations.²¹⁸ This discretion can impact both the types of evidence that courts are willing to accept and the types of evidence that parties have an opportunity to present.

¶ 88         For example, a court need not be bound by the rules of evidence in ruling on a motion for a preliminary injunction.²¹⁹ Indeed, the court need not hold an evidentiary hearing.²²⁰ Thus, where an evidentiary hearing is not held, the only evidence will be by way of affidavits and exhibits. In view of the extraordinary nature of the preliminary injunctive remedy, however, some courts will conduct an evidentiary hearing and may preclude affidavits from witnesses that a party fails to produce at such a hearing.²²¹ Frequently, however, courts will relax evidentiary requirements in ruling on a motion for preliminary injunctive relief.²²² This circumstance is important because, as discussed below, much evidence of actual confusion is potentially subject to evidentiary objections.

¶ 89         Another procedural issue that severely constrains appellate courts' ability to provide guidance and ensure consistency in judicial treatment of evidence of actual confusion flows from the deference that appellate courts give to trial courts in this area. Determining such matters as credibility is inherently a function of the trial court. The weight that trial courts afford to evidence of actual confusion is given deference by appellate courts.²²³ Furthermore, findings of actual confusion are reviewed under the "clearly erroneous" standard.²²⁴ Indeed, the entire likelihood of confusion inquiry is so fact-intensive that some courts have held that each infringement decision has little precedential value.²²⁵

¶ 90         The discretion afforded to district courts on this matter is not unbridled, but it is broad. For example, actual confusion may not be inferred from factors tending to show a likelihood of confusion.²²⁶ It has also been held that judgment cannot be rendered in favor of the defendant based solely upon a side-by-side comparison of the marks without hearing witnesses proffered to show actual confusion.²²⁷ In addition, at least one appellate court has held it to be reversible error to disregard evidence of actual confusion on the part of distributors and trade show visitors.²²⁸ Conversely, however, a number of courts have held that a plaintiff cannot create a genuine issue of material fact sufficient to avoid summary judgment on the likelihood of confusion test merely by introducing de minimis evidence of actual confusion.²²⁹ Similarly, factual findings of actual confusion have been held clearly erroneous where there was no evidence that the persons purportedly confused had ever seen the allegedly confusing source identifier or where the findings were based on hearsay.²³⁰ It has also been held that evidence of actual confusion does not compel a finding of likely confusion as a matter of law.²³¹

¶ 91         Apart from unusual categories of instances in which appellate courts have found an abuse of discretion, most of the problematic features of evidence of actual confusion are not susceptible to resolution by articulating "bright line" tests to be applied as rules of law. Rare cases of reversible error tend to occur where the trial court incorrectly concludes that only one inference can be drawn from the evidence of actual confusion when other inferences are permissible or where the trial court misstates facts that impact the amount of confusion that would be expected during the period of infringement.²³² One Seventh Circuit decision held—over a vigorous dissenting opinion—that the trial court committed reversible error by finding a likelihood of confusion in the absence of evidence of actual confusion, given that the evidence offered in support of the other factors was only sufficient to establish a scenario under which confusion was hypothetically possible.²³³ A subsequent decision of that court interpreted the ruling to mean that the plaintiff must introduce some evidence of actual confusion in order for that factor to be weighed in favor of the plaintiff.²³⁴ Yet, if all of the likelihood of confusion factors are weighed together in light of the particular circumstances of the case, the actual confusion factor might be deemed to be less important than the other factors in a case where such evidence is not likely to have surfaced.²³⁵

¶ 92         A good example of the discretion afforded to trial courts can be seen in Lever Brothers Co. v. American Bakeries Co.²³⁶ In that case, the plaintiff used the mark AUTUMN for margarine, and the defendant used the mark AUTUMN GRAIN for bread. The parties largely sold their products in different geographic markets albeit with some area of sales overlap and with extensive advertising by both parties.²³⁷ The trial court found that the proximity of the products and their impulse purchase nature tended to point to a likelihood of confusion but held that confusion was not likely when such factors were considered in light of the fact that there had been no actual confusion over several years. On appeal, the plaintiff argued that it was unrealistic to expect evidence of actual confusion in light of the generally geographically separate nature of the parties' sales. However, because there was some evidence of sales overlap and because an inference could be drawn that confusion was unlikely where there were several years of sales in the face of the overlap without evidence of actual confusion, the Second Circuit held that it could not substitute the plaintiff's alternative interpretation of the evidence for that of the trial court:

Essentially, Lever would have us place the emphasis differently, by holding that minor sales overlap cannot be expected to result in actual confusion, and accordingly the proximity issue should have been decided in Lever's favor and without reference to the absence of confusion. Whether we find this alternative reading of the evidence attractive is not the point. We decline to substitute Lever's proposed analysis for the one employed by Judge Neaher.²³⁸

¶ 93         One consequence of such broad discretion is the potential for outcomes that may seem to be in tension with various maxims about actual confusion. For example, in some instances where the parties have had sustained head-to-head competition for sufficient time, some courts have required substantial evidence of actual confusion to maintain the viability of the claim.²³⁹ Yet, in Allied Marketing Group, Inc. v. CDL Marketing, Inc.,²⁴⁰ the Fifth Circuit refused to overturn a district court finding of likely confusion in such circumstances where the evidence of actual confusion was minimal. In that case, the parties each employed similar promotional postcard mailers for approximately one year.²⁴¹ During that year, the parties disseminated millions of these postcard mailers but less than ten examples of actual confusion surfaced.²⁴² The defendant argued that this evidence was inadequate in light of the degree of competition. Nonetheless, because evidence of actual confusion is typically not required and because the district court had relied on factors other than actual confusion, the Fifth Circuit refused to find the likelihood of confusion holding to be clearly erroneous.²⁴³ Thus, unless appellate courts articulate factors to guide the decision-making of trial courts, the trial courts will frequently be presented with competing sets of precedent that can support diametrically opposed outcomes on the questions of whether or not actual confusion is needed and whether or not actual confusion has been established.

¶ 94         Another source of inconsistency and lack of clarity in how courts treat evidence of actual confusion lies in the interface between trademark law and the rules of evidence. Specifically, much evidence of actual confusion is either inadmissible or, if admissible, presents many of the same concerns that underpin the rules for excluding certain types of evidence. The central issue here is the hearsay rule, which excludes, subject to exceptions and limitations, statements made by a declarant outside the courtroom when offered for the truth of the matters asserted in such statements.²⁴⁴ Underpinning this exclusion is the notion that courts should not trust out-of-court statements not made under oath where the declarant's credibility and demeanor cannot be assessed and where the declarant cannot be subject to cross examination.²⁴⁵

¶ 95         Some courts have rejected as hearsay out-of-court statements that are offered as evidence of actual confusion.²⁴⁶ This result is particularly likely where there is any question as to whether the statements are genuine or have been shaped in any way by one of the parties or its attorneys.²⁴⁷ Often, however, such statements will be admissible either by qualifying for the statement of then existing state of mind exception or as a non-hearsay statement that is offered for purposes other than its truth content.²⁴⁸

¶ 96         A number of courts have admitted misdirected communications and statements reflecting mistaken impressions as a statement of the declarant's then existing state of mind.²⁴⁹ Nevertheless, where the declarant transmits information about someone else's state of mind, as where a company's representative makes a telephone call at the instigation of someone else within the company who was allegedly confused, the "then existing state of mind" exception will not allow the proponent to introduce that statement as indicative of the other employee's state of mind.²⁵⁰ The same reasoning would exclude statements of then existing state of mind that are made to one company employee and passed to another who testifies about them.²⁵¹ At least one court has held that an affidavit by a plaintiff's employee recounting telephone conversations with purchasers will not qualify for the state of mind exception.²⁵² In addition, the Seventh Circuit has held that a statement purportedly showing the declarant's actually confused state of mind was properly excluded where the declarant could not be identified and the statement was paraphrased by the witness who testified about it.²⁵³ Another court has distinguished between out-of-court statements that are too vague to show that any purported confusion resulted from more than inattentiveness and out-of-court statements where the reason for confusion is provided with sufficient detail to determine that the statement is a statement of the declarant's then existing state of mind and is, therefore, admissible.²⁵⁴

¶ 97         Many courts hold that a variety of out-of-court statements indicating confusion are not hearsay at all. The types of communications subject to such holdings include misdirected communications and mistaken inquiries. These courts reason that such items are offered not to prove the truth of the matter asserted—for instance, that the declarant intended to call the wrong party—but, rather, to prove confusion. If the substance of the communication does not assert that the declarant was confused, then the statement is not offered for the truth of any matter asserted.²⁵⁵

¶ 98         Even though out-of-court statements offered as evidence of actual confusion will often be admissible as either "statements of then existing state of mind" or as non-hearsay statements that are "not offered for the truth of any matter asserted," the contents of any particular statement will need to be scrutinized carefully. For example, where the statements at issue consist of opinions on the ultimate issue of confusion rather than facts indicative of confusion, the statements will not be admissible.²⁵⁶

¶ 99         In determining that certain out-of-court statements either are not hearsay or are eligible for an exception to the hearsay rule, certain other factors may be influential as well. First, out-of-court statements that clearly and unambiguously indicate confusion are likely to be seen as reliable.²⁵⁷ For example, if sufficiently clear and unambiguous, misdirected communications by customers may "speak for themselves" without explanation by the customers.²⁵⁸ Second, many courts are used to relying on out-of-court statements to establish actual confusion.²⁵⁹ Nevertheless, particularly in light of the relaxed evidentiary requirements at the preliminary injunctive phase, as well as the fact that there are not well established rules delineating such matters as the necessity for identification of an out-of-court declarant to qualify for the then existing state of mind exception, much evidence of actual confusion is admitted despite its ambiguous quality. In such instances, the focus shifts to what weight must be given to such evidence in light of its substantive shortcomings.²⁶⁰

¶ 100         The mere fact that out-of-court statements indicating actual confusion may be admissible does not mean that such statements will necessarily establish actual confusion. Such statements are often subject to shortcomings that may be used in attacking the weight of such evidence.²⁶¹ The potential shortcomings of out-of-court statements are numerous. Some issues that limit the value of evidence of actual confusion have already been introduced­—namely, questions of whether the persons purportedly confused are deemed relevant persons and whether the amount of evidence of actual confusion is of a sufficient quantity. This Subsection reviews several additional factors that may diminish the quality of evidence of actual confusion in nearly all cases. These factors include interpretive problems presented by the evidence, the intelligence of the persons purportedly confused, the level of care or attentiveness exercised by such persons, the potential biases of such persons, and the credibility of actual confusion witnesses.

¶ 101         Evidence of actual confusion is often vague, ambiguous, or both.²⁶² This problem is particularly pronounced where the persons who were purportedly confused are not available to testify, as there may be insufficient evidence to determine what induced their purportedly confused state of mind.²⁶³ As one court observed, where such persons do not testify, there is no way of "know[ing] why they asked what they asked or said what they said."²⁶⁴ Even if the purportedly confused statement is admissible, essential context may be missing to the extent that additional out-of-court statements by the declarant indicating why that person formed the purportedly mistaken belief may not be admissible, inasmuch as such statements are offered for the truth of the matter asserted and go beyond the scope of a statement of contemporaneously existing state of mind.²⁶⁵ Thus, without subjecting the declarant to cross examination, it may be unfair to lend much weight to out-of-court statements.²⁶⁶ Indeed, at least one court has held that such out-of-court statements should be given less weight where the plaintiff had the opportunity but failed to present some of this evidence through in-court testimony.²⁶⁷ Where the individuals in question are not even identified, then the defendant cannot verify the instances or issue subpoenas for such persons to appear as witnesses for the defendant.²⁶⁸

¶ 102         Evidence of actual confusion that is vague or ambiguous is generally entitled to little or no weight.²⁶⁹ For example, in Mushroom Makers, Inc. v. R.G. Barry Corp.,²⁷⁰ where the parties were using identical "mushroom" marks on different apparel products, and both parties had extensive sales, the court discounted instances of "clerical errors and joint advertising involving the two companies' products" as there was no "testimony to explain these events...."²⁷¹ In Armstrong Cork Co. v . World Carpets, Inc.,²⁷² where the parties each used the term "world" in connection with their carpet businesses, the court found an article in a trade magazine referring to an affiliation between the parties to be insufficient to demonstrate actual confusion, where it was undetermined whether the author of the article was confused, merely speculating, or being facetious.²⁷³ In Black Dog Tavern Co. v. Hall,²⁷⁴ where both parties sold T-shirts featuring dogs in the same geographic area, requests made of the plaintiff's screen printer for the defendant's T-shirts were held not to be evidence of actual confusion since the screen printer was not exclusively working for the plaintiff and it would be reasonable to infer that the defendant's designs may have been in her product line.²⁷⁵

¶ 103         In some instances, purported evidence of actual confusion may really be evidence of something entirely different.²⁷⁶ For example, mere observations by witnesses as to the similarities between the respective parties' marks do not show that such witnesses were confused.²⁷⁷ Similarly, an assumption that one of the parties appropriated its name from the other is not evidence of actual confusion, absent some other indication that the person was confused with respect to source, sponsorship, affiliation, or connection.²⁷⁸ Actual confusion that is attributable to features not subject to trademark protection has been held irrelevant.²⁷⁹

¶ 104         Linguistic or spelling errors may not indicate confusion between the marks, particularly where there are other circumstances showing that the person in question understood the distinction between the two businesses.²⁸⁰ For example, because the mark "Alpa" is susceptible to the common misspelling "Alpha," and because there was evidence that consumers used that misspelling before the defendant began using the mark "Alpha," a misspelling that followed the defendant's adoption of that mark was held not to be evidence of actual confusion.²⁸¹ Mistakes between companies may also result from misdirection or miscommunication rather than similarities between the parties' marks.²⁸²

¶ 105         Other types of non-actionable errors may arise from mistaken assumptions or general curiosity resulting from a new entrant in the market. For example, where two parties use identical or nearly identical marks and the second comer has not yet obtained a telephone listing, telephone calls may be misdirected solely because of the single listing.²⁸³ Moreover, where only one business in an industry has existed within a geographic region, consumers may assume that representatives of a newcomer are working for the first business simply because the first business is the only one they are aware of within the industry.²⁸⁴ Furthermore, where a sales clerk offers one product instead of another and is not available to testify as to the context of the transaction, it is possible that the sales clerk was merely attempting to offer a functionally equivalent product without being confused as to source.²⁸⁵ In addition, confusion will not be established by a generalized inquiry about what products or services the plaintiff produces or by questions or erroneous assumptions about whether the plaintiff produces particular products or services.²⁸⁶ These issues help to explain why confusion that is fleeting or temporary may not be credited.²⁸⁷

¶ 106         In other instances, circumstances that would make confusion unlikely may indicate that purported incidents of actual confusion are caused by something other than the use of the marks at issue. For example, where the two parties did not have a coexisting presence in the same consumer audience during the time period in which purported incidents of confusion occurred, it is unlikely that those incidents are evidence of confusion as to source between the parties.²⁸⁸

¶ 107         Inquiries regarding affiliation are particularly prone to differing interpretations and involve inconsistent treatment by courts. Some courts disfavor evidence of inquiries on the grounds that the person making the inquiry is aware of the distinction.²⁸⁹ Nevertheless, a number of courts accept inquiries about relationship as evidence of actual confusion, albeit with limited weight.²⁹⁰ At least one court has held that inquiries should be accepted as evidence of actual confusion but should be given limited weight.²⁹¹ Although it is not always clear what leads courts to give credence to such evidence, courts may be inclined to accept it more readily where the other factors point strongly toward a likelihood of confusion, which may lead courts to interpret an inquiry as showing a confused state of mind.²⁹² In addition, courts may be more likely to credit such evidence where it is accompanied by other types of evidence of actual confusion.²⁹³ Indeed, according to one leading commentator, such inquiries may be relevant to the existence of confusion but are insufficient without other evidence to establish actual confusion.²⁹⁴ However, the value of inquiries can vary depending upon the context of the particular inquiry, and some such incidents may be sufficiently strong evidence of actual confusion as not to require other types of evidence to establish actual confusion.²⁹⁵ The key is whether sufficient context is available to determine the strength of the evidence.²⁹⁶

¶ 108         Courts may be less inclined to credit inquiries revealing a state of mind that appears uncertain rather than confused.²⁹⁷ By contrast, where the communication does not inquire about source, affiliation, sponsorship, or connection, but reflects a mistaken assumption about such a link, then the relevant part of the communication is not an inquiry at all and is more likely to be credited as actual confusion.²⁹⁸

¶ 109         One troubling aspect of inquiries is that any uncertainty in the mind of the person making the inquiry is likely to be dispelled with an honest answer.²⁹⁹ Thus, a number of courts have discounted or discredited instances of confusion that were temporary or fleeting.³⁰⁰ Other courts have given more credence to actual confusion that has been quickly dispelled.³⁰¹ The fact that such doubt is dispelled by the responses to such inquiries does not address the problem of persons who may harbor similar doubts, but may not ask for clarification.³⁰²

¶ 110         Temporary or fleeting instances of confusion may be given greater weight as courts increasingly recognize the doctrine of initial interest confusion. Under this doctrine, confusion is deemed relevant if it generates initial interest in a transaction with the defendant, even if that confusion is dispelled by further interaction with the defendant or its products.³⁰³ Such confusion is actionable because it appropriates the plaintiff's good will to draw consumer interest in a potential transaction with the defendant.³⁰⁴

¶ 111         Nevertheless, the fact that initial interest confusion is actionable does not automatically translate into a finding that inquiries about affiliation constitute initial interest confusion. For there to be initial interest confusion, it must be likely that the defendant's use of the mark in question would entice consumers into some action indicative of interest in a transaction—for example, by "reaching for the [defendant's] product on the shelf"³⁰⁵ or getting off a highway one exit too early because of a confusingly placed billboard.³⁰⁶ If a consumer is undeterred from seeking out the plaintiff after mistakenly coming across the defendant's presence or promotional materials, then no real initial interest has been generated.³⁰⁷ A mere inquiry about affiliation may not be sufficient for a second-comer to get its foot in the door to compete for a purchasing decision or any other meaningful action by the person temporarily confused.

¶ 112         Accordingly, it appears that instances of temporary or fleeting confusion are particularly likely to be discounted where relatively few in light of the opportunity for confusion to surface, where the goods or services at issue are purchased by sophisticated purchasers who are likely to make such purchases only after considering source, or where the persons purportedly confused are not persons who make purchasing decisions.³⁰⁸ Furthermore, initial interest is unlikely to harm a plaintiff's good will where the respective parties' markets are unrelated and unlikely to converge and where there is no evidence of intent to capitalize on any such initial interest confusion.³⁰⁹ Thus, for example, one court dismissed as de minimis twenty instances of communications by customers inquiring about affiliation and potential purchases where the parties had both coexisted for five years, in light of the fact that the very different nature of the parties' products and the level of sophistication of the consumers made confusion unlikely.³¹⁰ By contrast, a particularly large volume of inquiries may lead a court to conclude that the inquiries are indicative of real mass confusion unlikely to be easily dispelled.³¹¹ Similarly, inquiries may be interpreted as stronger evidence of actual confusion when combined with other items such as survey evidence that point to a likelihood of confusion.³¹²

¶ 113         One conceptual difficulty with discounting inquiries about affiliation or relationship is that a likelihood of confusion as to affiliation, sponsorship, or connection is actionable infringement.³¹³ Yet, an inquiry about affiliation or relationship does not necessarily mean confusion as to these items. For example, the mere fact that one mark causes consumers to be reminded of another mark is not evidence of actual confusion because such mental comparisons may be consistent with distinguishing between, rather than confusion among, sources.³¹⁴ As one court observed in the course of its holding that actual confusion was not established by two e-mails inquiring as to whether the plaintiff was associated with the defendant: "I think that anytime a person visits a wrong website, in which information on the site does not correspond to the presentation given, questions are going to be raised. I do not find that the above come close to supporting a finding of a likelihood of confusion."³¹⁵ However, where evidence of confusion as to affiliation or relationship indicates that consumers have altered their behavior based on such confusion, then actual confusion about affiliation or relationship has been deemed to be established.³¹⁶ Short of such a dramatic illustration of confusion, the contents of inquiries and the circumstances surrounding them may reveal genuine confusion.

¶ 114         Courts must attempt to resolve ambiguities or inconsistencies in evidence of alleged actual confusion by interpreting the entire statement or event in context. For example where a witness addressed a letter regarding the defendant's product to the correct address indicated on the defendant's packaging but then mailed the letter to the plaintiff after it was returned for having an insufficient address, the actions of correctly and then incorrectly addressing the letter are inconsistent and possibly ambiguous as to the witness's confusion.³¹⁷ Yet, where the contents of the letter referred to "your fine reputation" and the plaintiff had considerable fame among the consuming public, which the defendant did not, it becomes clear that the witness was mistakenly communicating with the plaintiff regarding the defendant's product.³¹⁸ Another example of how context can resolve ambiguity lies in a case in which employees' inquiries as to a union's sponsorship were deemed not to be evidence of actual confusion where the statements were made in the context of an organizing drive during which the employees must have known that the defendant had distributed the literature that prompted the inquiries.³¹⁹

¶ 115         Another contextual problem arises from the coexistence of protectable and unprotectable features. Although confusion attributable to unprotectable features of the plaintiff's mark or product is not actionable, to the extent that each party presents its mark in the context of common surrounding features that are not protectable, those surrounding features are relevant to the likelihood of confusion inquiry.³²⁰ In such cases, the challenge is to determine whether instances of actual confusion flow from the protectable features in light of the surrounding unprotectable features or whether the confusion flows only from the unprotectable features. Where there is a danger that confusion flows from unprotectable features that the court may discount the episode if the plaintiff does not submit the materials that caused the confusion.³²¹

¶ 116         Interpretive problems are particularly likely with respect to out-of-court statements by persons highly knowledgeable about the industry and, therefore, unlikely to be easily confused. An example can be seen in Fleischmann Distilling Corp. v. Maier Brewing Co.,³²² in which a distributor of whiskey under the mark BLACK & WHITE sued a brewer and grocer for using that mark in connection with the sale of beer. At trial, a regional vice president of the plaintiff was asked whether it was his contention that a customer would be confused in buying a can of beer sold by the defendant under the BLACK & WHITE name and think that the customer was getting a product of the plaintiff's producer. The witness responded as follows:

No. The only two very isolated instances, I was asked in Southern California by people who were in the business, 'When did you people start making beer?' And that is, as I said, only a couple of instances that has happened. No, my answer is no to your question.³²³

¶ 117         In discussing the import of this testimony, the Ninth Circuit noted that it had "some difficulty in interpreting" it.³²⁴ The Ninth Circuit observed that, because the inquiries were "made by those who were in the business and hence should know better, the trial judge may have been justified in believing that these inquiries were in the nature of jocular remarks used by way of banter."³²⁵ Ultimately, the court "attach[ed] little weight to his testimony."³²⁶

¶ 118         Interpretive problems may also exist with respect to circumstantial evidence. For instance, where a defendant instructs its employees on how to handle inquiries based on confusion,³²⁷ that instruction could be motivated by something other than previous instances of confused inquiries. As examples, such instructions could be motivated by intent to capitalize on anticipated confusion or by the recognition that confusion was likely or simply possible.

¶ 119         One commentator has argued that evidence of actual confusion should be assessed according to whether the proponent has taken all feasible steps to present such evidence in the most complete and documented form possible.³²⁸ While this circumstance is relevant, it should not necessarily be controlling in all instances. For example, where the trademark owner is not acting under advice of counsel, many reports of confusion may be accumulated but not in the best form possible. These reports may be more or less reliable in light of the various circumstances affecting the quality of the evidence and in light of the other factors for assessing evidence of actual confusion.

¶ 120         It has long been held that the law does not protect a trademark owner "from all of the inadequacies of human thought and memory."³²⁹ Courts examine the question of likely confusion from the standpoint of the ordinary consumer of average intelligence.³³⁰ In doing so, courts recognize that some people will always be confused and that such confusion is not actionable.³³¹ Such courts discount instances of actual confusion that cannot be attributed to anything other than cognitive failure.³³² Similarly, where wide disparities between the parties' marks make it highly unlikely that persons of normal intelligence could be confused, courts are likely to interpret inquiries as not indicative of actual confusion.³³³

¶ 121         Other courts have expressed solicitude for consumers who are ignorant, inexperienced, and gullible.³³⁴ However, solicitude for such persons is often expressed as part of the standard for identifying confusion at the level of the "ordinary purchaser".³³⁵ In other words, a person of ordinary intelligence can, nevertheless, be ignorant, inexperienced, and gullible in the commercial marketplace.

¶ 122         Evidence of actual confusion among particularly sophisticated persons—where credible—may be afforded greater weight.³³⁶ Nevertheless, where a sophisticated member of the trade has had sufficient opportunity to become acquainted with the respective parties, any purported confusion by that person may not be viewed as credible.³³⁷ In addition, sophisticated members of the trade might be expected to make inquiries about affiliation without becoming actually confused.³³⁸ Furthermore, "gossip" within the trade about whether a new company has anything to do with an already existing company has been discounted when not indicative of actual marketplace impact of the two companies' marks on the consuming public.³³⁹

¶ 123         Much confusion is attributable to carelessness or inattentiveness. How such confusion should be treated is a matter of some dispute. Typically, courts assume that an average degree of care will be used by the relevant populace in determining likelihood of confusion.³⁴⁰ Accordingly, a number of courts have discounted instances of confusion that appear to be driven by carelessness or inattentiveness.³⁴¹ If a consumer exercises less than ordinary care, any resulting confusion is irrelevant.³⁴² This issue tends to arise with respect to misdirected communications. Such incidents are particularly likely to be discounted when they are minimal in light of the total volume of communications processed by a party.³⁴³

¶ 124         Other courts, however, have given far more weight to confusion that is attributable to carelessness or inattentiveness or where the level of care is not certain.³⁴⁴ One rationale for giving weight to such confusion is that carelessness or inattentiveness can lead to sales for the infringer that might not otherwise be reaped.³⁴⁵ For example, in Harold F. Ritchie, Inc. v. Chesebrough-Pond's, Inc.,³⁴⁶ the plaintiff claimed that the defendant's "Valcream" mark for men's hair cream infringed the plaintiff's "Brylcreem" mark for the same product. The trial court found no likelihood of confusion and dismissed the action. The Second Circuit reversed, however, largely based upon the trial court's mischaracterization of evidence of actual confusion. This evidence consisted of: (1) a television commercial in which the defendant offered to refund purchasers of a particular size of the Valcream product with a check signed by a celebrity, which resulted in thirteen instances of Brylcreem cartons being mailed to the plaintiff and at least ten Brylcreem cartons being mailed to the defendant; (2) the testimony of a store manager and the depositions of two employees of another store about incidents whose nature was not described in the decision; (3) an instance in which a company mistakenly invoiced the plaintiff for the advertising the defendant's product in a newspaper; (4) an instance in which a company sent the plaintiff a letter relating to a price reduction offer made by the defendant for the defendant's products; and (5) a mistake in which Standard & Poor's transposed the names of the respective products and attributed each to the wrong corporation. The Second Circuit held that the trial court erred in discounting these instances as carelessness and inattention rather than as actual confusion and noted that such carelessness and inattentiveness can be the basis for an infringer to obtain wrongful sales. The Second Circuit also characterized the evidence as "impressive in view of its spontaneous character and difficulty of attainment."³⁴⁷

¶ 125         It may be possible to distinguish confusion that is based on carelessness from confusion that is based on similarities between the marks. Such a distinction is purportedly the basis for the decision in Roto-Rooter Corp. v. O'Neal.³⁴⁸ In that case, the plaintiff sold sewer, drain, and pipe cleaning services under the ROTO-ROOTER mark, and the defendant sold similar services under the mark ROTARY DE-ROOTING. Although four witnesses testified that they had mistakenly purchased the defendant's service when intending to purchase the plaintiff's service, the trial court held that confusion was not likely and, in doing so, discounted these instances of actual confusion as based on "carelessness" or "inadvertence." The Fifth Circuit reversed and held that the record did not support the finding that the witnesses' confusion was based on carelessness or inadvertence. Instead, the Fifth Circuit noted that cross examination of these witnesses had demonstrated that the mistaken purchases were made based upon reliance on the similarity of marks rather than carelessness or inadvertence.³⁴⁹

¶ 126         Perhaps, the Roto-Rooter court would have reached a different result if the witnesses' testimony had indicated that they paid scant attention to the marks in making the purchasing decision. However, because that case involved a mistaken purchase, the very act of making the purchase can be seen as evidence of the level of care expected in connection with such a purchase. By contrast, with other types of evidence of actual confusion, such as inquiries and misdirected communications, carelessness and inattentiveness are likely to play a greater role in limiting the value of the evidence.

¶ 127         Moreover, it may be possible to recognize degrees of care that are commensurate with expectations for how an ordinary consumer would react under the circumstances. Under such an approach, carelessness that is consistent with a reasonable consumer acting under similar circumstances would not discredit that person's confusion. For example, if reading the fine print on packaging, scrutinizing the defendant's mark, or comparing the parties' products directly would be necessary to avoid confusion and it would not be reasonable to expect a consumer to take such steps, any resulting confusion will not be discredited.³⁵⁰ Indeed, even a court that found a low level of confusion attributable to inattentiveness to be irrelevant nonetheless observed that infringement will be established where an appreciable number of persons are likely to be confused even if they could avoid the mistake by exercising greater caution.³⁵¹ Thus, a "casual purchaser" is held to the level of "ordinary care" that is appropriate for such persons.³⁵²

¶ 128         Other courts, although not clearly stating so, may accept evidence of misdirected communications without scrutiny based on a particularly large quantity of such misdirected communications or based on the fact that they are coupled with evidence of customers being enticed to the wrong location or making purchases from the wrong entity.³⁵³ In such cases, not only can a large volume of such communications seem difficult to ignore, but the volume may be high enough so that it will necessarily include consumers and will likely be attributable to more than mere carelessness with telephone directories and similar inattentiveness—for example, where persons identified as consumers express a mistaken belief that the parties are affiliated or connected or complain to the wrong party about goods or services sold by the other party.³⁵⁴ An additional reason that may persuade some courts to credit evidence of misdirected communications without scrutiny may be that other factors weigh heavily in favor of finding a likelihood of confusion.³⁵⁵ Nevertheless, some courts appear to give blanket approval to evidence of misdirected communications without regard for context,³⁵⁶ provided that the communications are from customers.³⁵⁷ However, at least one court refused to credit relatively few instances of misdirected communications even when the communications are from customers.³⁵⁸

¶ 129         One basis for distinguishing between confusion that is generated by inattentiveness from confusion generated by brand identity is to examine whether consumers had any reason to know of the entity to whom a mistaken communication was addressed. For example, as one court observed in discounting several e-mails that were directed to the defendant but intended for the plaintiff, "[t]he fact that [the defendant] does not advertise increases the likelihood that the people who sent the e-mail messages to [the defendant] were inattentive or careless when attempting to find the e-mail address for [the plaintiff], rather than confused about the source of the [products]."³⁵⁹

¶ 130         In some cases, courts have cited the service of legal documents or industry complaints on the wrong company as evidence of actual confusion.³⁶⁰ Such decisions could benefit from greater explanation of context. For example, where errors in service are attributable to carelessness or inattentiveness by a process server or clerical staff, such mistakes are analogous to other types of carelessness or inattentiveness in misdirecting communications that are not likely attributable to consumer confusion. However, if the mistake was made by a lawyer or by sophisticated support person who is knowledgeable about the party intended to be served and is likely to exercise a high degree of care, the mistake may augur confusion by less sophisticated and less attentive consumers. Despite this distinction, the prospect of a senior mark owner being tainted by any public association with the alleged legal wrongs of a junior user may lead courts to be particularly solicitous of misdirected service of process as an indicator of likely confusion. Some courts have gone so far as to find actual confusion in mistakes made by the attorneys litigating the case.³⁶¹

¶ 131         Any witness is potentially subject to impeachment based upon bias. Because so much evidence of actual confusion often takes the form of out-of-court statements made either contemporaneously with the events in question or through an affidavit where the witness cannot be cross-examined for bias, courts should be particularly sensitive to the potential for bias in assessing such evidence. Often, the relationship of the witness to one of the parties will be a source of bias. For example, a number of courts have discounted or discredited confusion that was claimed by employees or relatives of employees of the plaintiff.³⁶² Although not as inevitably biased as employees and relatives, some courts have discounted or discredited confusion that was claimed by acquaintances of the plaintiff's employees.³⁶³ Acquaintances are particularly likely to be biased when they have business relationships that give them an incentive to support the plaintiff.³⁶⁴ One court observed: "Trademark law is skeptical of the ability of an associate of a trademark holder to transcend personal biases to give an impartial account of the value of the holder's mark."³⁶⁵ Other forms of financial influence may also cause witnesses to be biased.³⁶⁶

¶ 132         Another potential source of bias is the desire of a sales clerk to make a sale by substituting an item in stock for a different item that was requested.³⁶⁷ In assessing whether this type of bias would have prompted a knowing substitution rather than a mistake, the court should consider the dollar value of the item and any other circumstances that would shed light on any such potential motive.³⁶⁸ Finally, animosity towards a party is yet another source of bias.³⁶⁹

¶ 133         The question of bias is particularly problematic where the declarant does not appear to testify and where circumstances that may indicate bias of that declarant are not known or where sources indicating bias cannot be explored on cross examination.³⁷⁰ This problem may be compounded further where the out-of-court declarant's statements are filtered through an additional level of bias when offered by the employees of one of the parties.³⁷¹

¶ 134         Nonetheless, the mere fact that a witness has a relationship with a party or an interest in the outcome of the matter may not always discredit that witness's testimony. Where the sources of the bias are explored and the trial judge determines the witness to be credible, that determination will not be disturbed by an appellate court.³⁷² Furthermore, because of the importance of confusion among actual purchasers, it has been held to be erroneous to discredit as "self-serving" an employee's testimony about confusion by an unidentified professional buyer where there was no finding that the employee's testimony was not credible.³⁷³ Indeed, business acquaintances who have an incentive to testify favorably to the plaintiff may also be the very class of persons whose state of mind is most relevant—the consumers of the parties' goods or services.³⁷⁴

¶ 135         Even where evidence of actual confusion survives the various hearsay obstacles to admission, the credibility of out-of-court statements can be a particular problem—quite apart from obvious sources of bias. In some instances, the circumstances surrounding a purported incident of confusion may simply render it implausible that the witness was actually confused by the parties' marks. One obvious example is where great differences exist between the respective marks at issue and between circumstances relevant to other factors that affect the likelihood of confusion analysis.³⁷⁵ In some instances, an out-of-court statement purportedly indicating confusion may be contradicted with live testimony that is more credible.³⁷⁶ Where witnesses are offered live in court to testify about actual confusion, however, it has been held to be reversible error to refuse to hear their testimony and, instead, render judgment for the defendant based solely on a side-by-side comparison of the marks.³⁷⁷ The court explained that "such comparison cannot conclusively negative the likelihood of confusion as a matter of law nor can it negative confusion as a matter of fact by the expedient of technically accepting the offered proof but summarily rejecting its weight or credibility."³⁷⁸ Nevertheless, it has also been held that, in some cases, a mere side-by-side comparison of the marks, without more, will be sufficient to establish a likelihood of confusion.³⁷⁹

¶ 136         Another circumstance that may render it improbable that a witness was actually confused would be where the witness had extensive knowledge and understanding of the respective parties. This circumstance may occur where the witness is a highly sophisticated and experienced member of the trade and where the respective parties have each achieved sufficient prominence in the trade so as to make it unlikely that a witness of such sophistication and experience would not understand the differences between the parties.³⁸⁰

¶ 137         Conversely, however, the purportedly confused persons may have insufficient knowledge of the parties to be actually confused. For example, where the parties did not have an overlapping presence among the same consumer audience during the period in which purportedly confused persons stepped forward, then it seems unlikely that those persons were confused as to the source of products or services offered by the respective parties.³⁸¹

¶ 138         Ways of attacking the credibility of a witness purporting to recount confusion by others include demonstrating that the witness is unable to recall specifics about the instances or showing that the number of specific instances the witness can recall conflicts with the witness's blanket generalization about the number of instances of confusion. For example, in Choice Hotels International, Inc. v. Kaushik,³⁸² the plaintiff offered two of its own employees as witnesses, and these employees recounted episodes of confusion by customers. However, the first employee was unable to identify names of customers or dates of the incidents and was only able to give specifics regarding two incidents, each of which involved police and fire department personnel arriving at the wrong location. The second employee had recounted "many" instances of confusion among customers in his deposition but could only recall one incident at trial.³⁸³ Given that the two establishments were located across the street from each other and that these employees would have been in a position to observe incidents of confusion, the court found it telling that the employees made no effort to document instances of confusion or express concern to the plaintiff.³⁸⁴ The court characterized the evidence as "vague and contradictory" and held that any actual confusion "was not substantial."³⁸⁵

¶ 139         Credibility of the witness may also be subject to challenge where the plaintiff instigated the reports of confusion. In some instances, courts have discounted purported confusion that was first reported in response to a communication by a plaintiff who sought such evidence.³⁸⁶ Similarly, courts have looked with a jaundiced eye at a party's effort to "manufacture" evidence by contriving transactions.³⁸⁷ Such contrived transactions are particularly likely to be discredited where the purchaser deliberately requests a product by using one party's mark but the product is only made by the second party.³⁸⁸ However, it has been held that a bald allegation of manufacturing evidence of actual confusion, without evidence to support the charge, will not be grounds for discrediting evidence of actual confusion.³⁸⁹ Nevertheless, witnesses who come forward on their own have been deemed more credible.³⁹⁰ A series of identical or nearly identical statements, however, may be more open to challenge.³⁹¹ Out-of-court statements that sound contrived or inconsistent with commercial reality will similarly be subject to challenge.³⁹² In one example, a court found it incredible that out-of-court witnesses could have associated the defendant's logo with the defendant sufficiently to be confused so soon after the defendant began using it.³⁹³ Similarly suspect are fact witnesses who have repeatedly testified for or against one of the parties in the past.³⁹⁴ Out-of-court statements may be subject to attack based on credibility where they do not include specificity as to the details of the incidents.³⁹⁵

¶ 140         Apart from circumstances that cast obvious doubt on a witness's purported confusion are unknown circumstances that surround so much evidence in trademark litigation. For example, one fundamental problem of credibility occurs where the persons who were purportedly confused cannot be identified.³⁹⁶ This problem is particularly pronounced when one cannot identify the individual in question as a potential customer for the goods in question.³⁹⁷ Where an out-of-court declarant can be identified, there is always the possibility that the declarant will come forward with testimony indicating that the statement attributed to that declarant is either incorrect or is not indicative of the declarant's true state of mind.³⁹⁸ One way to measure measure out-of-court declarants' credibility may be to compare those statements to the statements of witnesses who actually testify—credibly—that they were not confused.³⁹⁹

¶ 141         The credibility of witnesses purporting to be confused or reporting the confusion of others can be bolstered in a number of ways. Where such statements come from the mouths of an opposing party or its employees, the law of evidence deems such admissions to be inherently reliable.⁴⁰⁰ In addition, live testimony may assist in establishing the witness's credibility, depending upon how well the witness testifies.⁴⁰¹ Contemporaneous statements of confusion by the witness, if available, may bolster such testimony.⁴⁰² A consumer survey may also bolster the credibility of witnesses claiming to have been confused or reporting confusion.⁴⁰³ Some witnesses' credibility may be bolstered by consistent accounts of confusion that is the subject of testimony by other witnesses.⁴⁰⁴ The credibility of witnesses testifying as to their own confusion or reported instances of confusion can be bolstered by any circumstantial evidence of confusion that may be available, such any defendant practice of instructing its sales force in how to address confused inquiries.⁴⁰⁵ In addition, the plausibility of accounts of confusion and the fact that they are supported by individuals who are not subject to any credibility or bias challenge as well as those who are may persuade the court to credit all of the accounts.⁴⁰⁶

IV.    Approaches to Evidence of Actual Confusion

¶ 142         Discerning any formal standard by which to apply the actual confusion factor is difficult. Because so many decisions involving evidence of actual confusion are marked by differing treatment of the same or similar types of evidence, they provide limited guidance. In some instances, courts have focused on particular problems affecting such evidence—such as whether there has been sufficient opportunity for confusion to surface—and addressed such issues in a way that appears to formulate a standard for assessing such evidence.⁴⁰⁷ Piecemeal standards, however, do not provide comprehensive guidance. Other courts have articulated standards and tests that appear to be more comprehensive. Yet, closer examination reveals that even these tests often fail to address the full range of issues affecting the actual confusion factor—and, worse, either overemphasize the importance of the presence or lack of evidence of actual confusion or are too vague to provide meaningful guidance. This section reviews those standards for addressing actual confusion that are more comprehensive in nature and concludes that they reflect an insufficient appreciation for the multi-factor nature of the likelihood of confusion analysis and the different ways in which so many different circumstances impact the proper assessment of such evidence.

¶ 143         A number of jurisdictions apply a form of burden shifting whereby a showing of the plaintiff's actual confusion will place a more substantial burden on the defendant to prove that confusion is not likely. This approach was developed by the Fifth Circuit in World Carpets, Inc. v. Dick Littrell's New World Carpets.⁴⁰⁸ In that case, a manufacturer and wholesale distributor of carpets under the WORLD mark sued a carpet retailer for using the words NEW WORLD CARPETS in its trade name. The parties went to trial, and, after the close of evidence, the trial court directed a verdict in favor of the plaintiff on its infringement claim.⁴⁰⁹ During the trial, the plaintiff introduced "uncontradicted testimony that World Carpets had been called by retailers who [mistakenly] thought World Carpets had entered the retail market."⁴¹⁰ Because of this evidence of actual confusion and the defendant's failure to refute it, the Fifth Circuit shifted the burden to the defendant and upheld the directed verdict.⁴¹¹ In a frequently quoted passage, the court explained: "reason tells us that while very little proof of actual confusion would be necessary to prove the likelihood of confusion, an almost overwhelming amount of proof would be necessary to refute such proof."⁴¹² A number of courts have followed this approach.⁴¹³

¶ 144         In light of the preferred status given to evidence of actual confusion, World Carpets' form of burden-shifting appears to shift to the defendant the burden, not merely of disproving actual confusion, but also of proving that confusion is not likely. Indeed, the Fifth Circuit subsequently referred to this test as one that "shift[s] the burden of proof on the likelihood of confusion issue to the defendants."⁴¹⁴ According to one court, a defendant "may rebut the evidence of actual confusion with evidence, for example, that the confused customer was a rarity or drew an unreasonable conclusion and that there was no confusion regarding [the products'] history in the overwhelming majority of transactions."⁴¹⁵

¶ 145         As discussed above, a number of courts have held that the lack of actual confusion carries evidentiary weight of its own.⁴¹⁶ Not surprisingly, some have cast this approach as creating a presumption that confusion is not likely from the absence of evidence of actual confusion.⁴¹⁷ Others do not use the term "presumption" but find that the absence of such evidence weighs heavily against the plaintiff.⁴¹⁸ Still others take a different tack and simply hold that an inference may be drawn against the relevant party based on the presence or absence of evidence of actual confusion.⁴¹⁹ However, where circumstances indicate that no confusion would be expected, no negative inference should be drawn where no confusion has surfaced.⁴²⁰ Other courts reject any inference from the lack of evidence of actual confusion without making it clear whether that rejection is rooted in specific circumstances of the case.⁴²¹ Finally, another approach simply views the actual confusion factor as one of many factors relevant to the likelihood of confusion inquiry and views the presence or absence as establishing or not establishing the factor—although within the decisions taking this approach there is a difference as to whether the lack of such evidence means that the factor favors the defendant or is simply neutral.⁴²² One commentator has argued that courts should not draw any automatic presumption or inference from the absence of evidence of actual confusion but, instead, should consider whether such evidence should have surfaced in light of the types of products and companies, whether confused consumers would be likely to express their confusion, the volume of the parties' sales, and any other factors that would affect the accessibility of such evidence.⁴²³

¶ 146         These varying approaches are not all consistent with the same type of burden-shifting. For example, nothing short of presuming a likelihood of confusion from the presence of actual confusion could shift the burden of proving no likelihood of confusion to the defendant. By contrast, a permissive inference that confusion is likely based on actual confusion could conceivably shift the burden of coming forward with evidence—but solely with respect to the actual confusion factor itself and not with respect to the entire infringement standard.

¶ 147         A benefit of the burden-shifting approach is that it could provide guidance to trial courts on the amount of evidence required to establish actual confusion and set constraints on the trial court's discretion by making reversal much more likely where evidence of actual confusion is ignored—if the courts could settle on a consistent form of burden-shifting. An example is Meridian Mutual Insurance Co. v. Meridian Insurance Group, Inc., where an insurer using the MERIDIAN mark sued an insurance broker that was using the term "Meridian" as part of its name.⁴²⁴ The district court had denied a preliminary injunction and, in doing so, found four misdirected telephone calls to be irrelevant because they did not involve actual or prospective purchasers. The Seventh Circuit reversed and, in doing so, found that the district court committed clear error and concluded that the four persons who made the misdirected calls constituted the "very little evidence" necessary to shift to the defendant the burden of coming forward with "overwhelming proof" to rebut such evidence, at least for purposes of a preliminary injunction hearing.⁴²⁵ Moreover, steps that the defendant had taken to avoid confusion, such as discontinuing the telephone number that was the source of the misdirected calls, was held to be insufficient to rebut the evidence of actual confusion.⁴²⁶ Thus, the failure to give "substantial weight" to even a small number of instances of actual confusion among a relevant group, at least at the preliminary injunction phase, may constitute reversible error under the burden-shifting approach.⁴²⁷

¶ 148         The problem with the burden-shifting approach is that it does not differentiate between different types and quantities of evidence of actual confusion produced under different circumstances. Thus, in the Meridian example, misdirected calls are relatively weak evidence of actual confusion, and four is a relatively small number. By contrast, where such factors are used to discredit such evidence and avoid burden-shifting, that analytic step occurs before applying the test and, hence, without any guidance from it.

¶ 149         An example of how critical evaluation of evidence may precede and, thus, circumvent any burden-shifting test can be seen in Holiday Inns, Inc. v. Holiday Out in America.⁴²⁸ In that case, the plaintiff operated a chain of motels under the HOLIDAY INN mark, and the defendants offered campground services under marks containing the term "Holiday Out." After the district court found in favor of the defendant, the plaintiff argued on appeal that evidence of actual confusion shifted the burden to the defendant.⁴²⁹ Although the plaintiff's vice president testified to receiving inquiries about the defendant, and although the plaintiff introduced nine letters and a memorandum of a telephone call indicating confusion by the persons making such communications, the Fifth Circuit held that the burden was not shifted to the defendant. Without offering much description of the substance of these communications, the court questioned them in light of the fact that their authors were not available for cross examination, in light of the fact that two of the letters were suspiciously similar and generated after the litigation commenced, and in light of the testimony of a manager of one of the plaintiff's motels at the only location where the defendant was also based to the effect that no confusion was observed over the course of two years.⁴³⁰

¶ 150         Thus, in Holiday Inns, all of the evaluation of the evidence of actual confusion occurred prior to, and as a way of circumventing, the burden-shifting test; yet no real guidance was offered as to how to evaluate such evidence short of applying the burden-shifting test. The lack of such guidance is particularly striking with respect to the absence of evidence of actual confusion at the one location where the parties were in competition for a significant period of time. As previously discussed, a number of courts, including those in the Fifth Circuit, have held that the absence or relative lack of such evidence, where many instances of actual confusion would be expected, actually gives rise to a presumption that confusion is not likely.⁴³¹

¶ 151         As with the presumption that is drawn against the defendant when the burden is shifted, however, merely establishing such evidentiary rules does not address the distinctions in different quantities of evidence and different circumstances in which the absence or relative lack of such evidence may arise. Indeed, determining when a presumption from the lack of evidence of actual confusion would arise is particularly difficult, given that the lack of such evidence often will not demonstrate anything in light of the difficulty of obtaining such evidence.⁴³² A presumption in either direction presents the additional problem of elevating one of the likelihood of confusion factors over all others, including ones that are recognized as more accurate predictors of likely confusion.⁴³³

¶ 152         Some courts weigh the evidence of actual confusion by looking to the totality of circumstances surrounding the claimed confusion.⁴³⁴ "This examination may include consideration of the time period in question and how extensively the product is advertised or made known to the public..., as well as the type of confusion that exists and who suffers the confusion."⁴³⁵ The examination should encompass "countervailing circumstances which lessen the impact of asserted instances of confusion."⁴³⁶

¶ 153         Although the total quantity of incidents in light of the opportunity for confusion to develop is relevant to this analysis, other factors such as price, incentive to report confusion, the types of persons confused, and the quality of the evidence of confusion may lead a court to give greater emphasis to instances of confusion that are relatively few in number. For example, in AmBrit, Inc. v. Kraft, Inc.,⁴³⁷ the Eleventh Circuit upheld a finding that an ice cream bar wrapper infringed the plaintiff's trade dress and trademark used to distribute its competing ice cream product and upheld a finding of actual confusion based upon just four reported instances of actual confusion. Specifically, the Eleventh Circuit held that the trial court had properly found actual confusion based upon testimony of four consumers who reported being confused by letter or telephone, as well as a survey that was given little weight by the trial court. Although the defendant argued that four instances of confusion were too few in light of the high sales volumes at issue, the Eleventh Circuit concluded that this argument was not persuasive in light of the difficulties of obtaining evidence of consumer confusion - particularly, in cases involving such low-priced goods. Given that many consumers who are confused will not realize their confusion in order to report it and given that many consumers who come to realize their confusion will not bother to report that confusion to a "faceless corporation" when the dollar value of the purchase was so minimal, the court found that the four reported instances of confusion were sufficient to establish actual confusion. The Eleventh Circuit characterized this evidence as "far from overwhelming" but, nonetheless, sufficient to support the finding.⁴³⁸ Although the defendant argued that the four instances of confusion were inconsequential because the consumers were careless, the court held that the argument was misguided insofar as the items in question were inexpensive impulse items that are not typically purchased with great care.⁴³⁹ Finally, albeit not emphasized by the court, although the number of reported instances of actual confusion were few, the evidence of these instances was likely highly credible insofar as the incidents involved mistaken purchases and the witnesses testified in person and had contemporaneously reported their confusion by letter or telephone.⁴⁴⁰

¶ 154         One problem with the totality of the circumstances approach is that it often provides little or no guidance as to the manner in which various circumstances should affect how evidence of actual confusion or the lack thereof should be viewed. One example is as follows:

[A]ctual confusion or its absence is only one factor in the analysis of likelihood of confusion, and cases . . . have found its absence to be either important or insignificant, depending on the evaluation of all the other factors. The significance of this factor will therefore have to be judged in light of the evidence with respect to the other factors.⁴⁴¹

¶ 155         The solution proposed by this Article is to establish non-exhaustive factors to be weighed in assessing evidence of actual confusion.

¶ 156         One approach that has the potential to capture critical judgments that other approaches either elide or obscure is to focus on the value of the evidence of actual confusion. This approach recognizes that evidence of actual confusion, including its type, quantity, and/or lack can be more or less valuable in different circumstances.⁴⁴² The fact that value is assigned based upon particular circumstances brings to mind the totality of the circumstances test. However, focusing on the value of the evidence brings attention to those aspects of the case that can enhance or detract from various types of evidence of actual confusion. By recognizing the different types of circumstances that tend to give more or less value to evidence of actual confusion, this approach has more potential to channel the seemingly untethered discretion of lower court judges.

¶ 157         An example of judicial scrutiny of the value of evidence of actual confusion can be seen in the Seventh Circuit's decision in Union Carbide Corp. v. Ever-Ready Inc.⁴⁴³ In that case, the plaintiff owned trademark registrations on the term EVEREADY for use in connection with electric batteries, flashlights, and miniature bulbs for automobile and marine use. Between October 1965 and July 1967, the plaintiff sold bulbs under its EVEREADY mark in blister packages designated for use in high-intensity reading lamps. The defendants used the mark EVER-READY in connection with importing and distributing electrical supplies, stationary, gift items, and accessories, including lamps, light bulbs, light fixtures, and flashlights. In 1969, the defendants began importing miniature lamp bulbs with the EVER-READY mark stamped on the bases and distributed these bulbs in blister packages also bearing that mark and designated for use in high-intensity lamps. The defendants also imported high-intensity lamps stamped with the same mark.⁴⁴⁴

¶ 158         In reversing the district court's holding that there was no likelihood of confusion, the Seventh Circuit closely scrutinized the district court's rejection of all evidence of actual confusion. In framing this discussion, the Seventh Circuit noted the tension among cases holding actual confusion to be unnecessary, cases holding actual confusion to be the best evidence of likelihood of confusion, and cases holding isolated instances of actual confusion to be insufficient in establishing a likelihood of confusion. The Union Carbide court resolved this tension by focusing on the value of the evidence of actual confusion in light of the commercial realities:

¶ 159         The value of evidence of actual confusion is greater when the products involved are low value items because purchasers are unlikely to complain when dissatisfied, which would bring to light confusion; but rather they are likely simply to avoid all products produced by the company which they believe produced the product which caused them trouble.

¶ 160         With this focus on the value of the particular evidence of actual confusion, the Seventh Circuit then evaluated three incidents of actual confusion and emphasized additional circumstances that may impact the value of the evidence such as whether the statements of allegedly confused persons reflect assumptions about which company those persons were communicating with, whether instances of confusion appear credible in light of evidence that transactions were instigated by interested parties and in light of commercial circumstances that shed light on the motives of the parties involved in such transactions, and whether the level of care exercised by the purportedly confused persons was consistent with the nature of the products at issue.⁴⁴⁵

¶ 161         In addition to offering considerable guidance on how to value evidence of actual confusion, Union Carbide is also instructive in the limited control the court exerted over the discretion of trial judges. Specifically, notwithstanding several areas of disagreement with the district court's findings as to the value of the evidence presented on actual confusion, the Seventh Circuit noted that these points of disagreement, standing alone, would not be sufficient to hold that the trial court was clearly erroneous in holding that there was no likelihood of confusion.⁴⁴⁶ Nevertheless, when these areas of disagreement over the evidence of actual confusion were combined with the district court's rejection of survey evidence that the appellate court found probative of likely confusion, as well as other factors weighing in favor of likely confusion, the Seventh Circuit held that the district court committed clear error and reversed.⁴⁴⁷ In light of the wide degree of discretion afforded to district courts, it is not enough for appellate courts simply to provide examples of how they would value particular instances of confusion in light of particular circumstances if they were sitting as trial courts. Rather, explicit criteria are needed to guide trial courts valuing such evidence.

¶ 162         Even courts purporting to follow the burden-shifting test have showed signs of abandoning that approach in favor of an approach that affords different value to different types of evidence of actual confusion. Specifically, in Safeway Stores, Inc. v. Safeway Discount Drugs, Inc.,⁴⁴⁸ the then newly-formed Eleventh Circuit purported to follow its predecessor, the Fifth Circuit, but noted an inconsistency between the burden-shifting test, which is triggered by very little evidence of actual confusion, and decisions that have discounted relatively few instances of actual confusion that have occurred over a significant period of time.⁴⁴⁹ The court then focused on how much weight should be afforded to such evidence and pointed to factors influencing this determination other than simply the quantum of evidence and the relevant time period:

¶ 163         Perhaps as important as, and helping to explain the various interpretations of the relevance of, the number of instances of confusion are the kinds of persons confused and degree of confusion. Short-lived confusion or confusion of individuals casually acquainted with a business is worthy of little weight... while confusion of actual customers of a business is worthy of substantial weight.⁴⁵⁰

¶ 164         The court went on to find only two instances of confusion to be highly probative. One instance involved a misdirected letter from a supplier whose familiarity with the trade would presumably make confusion difficult. The second instance involved a customer—the precise class that must not be confused for a business to compete.⁴⁵¹

¶ 165         The value of the evidence approach improves upon the totality of the circumstances approach by recognizing that not all evidence deserves equal weight. Some evidence of confusion is highly probative. Other such evidence is less so. In each case, the value of the evidence must be determined by context.

¶ 166         Ultimately, however, like the totality of the circumstances approach, the value of the evidence approach fails to provide sufficient guidance. Missing are any general methods for ranking different types of evidence and any received understanding as to contextual factors that affect value. The solution proposed by this article is to articulate types of evidence that tend to have differing degrees of value and types of circumstances that tend to affect value in the form of a multi-factor test.

V.    A Multi-Factor Test for Valuing Evidence of Actual Confusion

¶ 167         Each of the approaches set forth above presents different difficulties. Either such approaches do not provide meaningful guidance or such approaches reflect overemphasis on the presence or absence of evidence of actual confusion with insufficient regard for context or the multi-factor nature of the likelihood of confusion analysis and the proper place of the actual confusion factor in that test. One solution to the difficulty in determining actual confusion is to take a lesson from the approach that courts have developed to establishing likelihood of confusion—namely, to use a multi-factor test that should be considered in valuing such evidence. These factors should give explicit recognition to the types of circumstances set forth in this article that are relevant to valuing evidence of actual confusion. In turn, each factor should be considered in light of sub-factors. As with the likelihood of confusion test itself, any such list of factors should be nonexclusive due to the infinitely varied circumstances that may affect any particular case.

¶ 168         The first factor to be assessed is the quality of the evidence of actual confusion. One component of this factor is admissibility of the evidence. If the evidence is not in admissible form, it should be excluded at trial. At a preliminary injunction hearing, the court must determine if there is reason to accept such evidence. This assessment must take into account the apparent reliability of the evidence—including its completeness and context—as well as the proponent's ability to present the evidence in a better format in light of the stage of the proceedings and the court's ability to conduct a full evidentiary hearing at the preliminary injunction phase.

¶ 169         If the court relaxes the evidentiary rules, or if the evidence satisfies an exception to the hearsay rule but, nonetheless, presents the very concerns that underlie the hearsay rule, then the court should discount such evidence, unless present in an overwhelming quantity or bolstered by evidence of other types of evidence of actual confusion, survey evidence, or a strong showing with respect to the other likelihood of confusion factors. Among other things, courts should determine whether the persons whose statements or actions at issue are identified, how many chains of declarants translate their statements or actions before they are offered into evidence, the opportunity of lawyers to shape the statements or actions that are offered into evidence, the completeness of any statements or actions, and context for any statements or actions, and the feasibility of gathering and presenting evidence with greater detail, given, among other circumstances, the nature of the incidents of confusion, the nature of the businesses, the ease with which such incidents can be documented, and the involvement of counsel during the time when incidents of confusion were surfacing.

¶ 170         Courts should recognize that different types of evidence of confusion are subject to differing degrees of reliability of evidence depending upon the degree to which such evidence consists of actual real-world reactions to the parties' marks as used in the commercial world, whether the actions or statements at issue require a degree of investment of thought, resources, or energy such that any confusion rises above the level of mere musings about the parties or careless or inattentive actions or statements, and whether such evidence takes the form of clear and unambiguous actions or statements and provides sufficient context to discern the nature and context of confusion. Thus, evidence of actual confusion will vary in descending order of quality roughly from mistaken purchases, other mistaken commercial interactions such as provision of financing, credit, or supplies, obtaining service or repairs from the wrong party, returning products to the wrong party, redeeming coupons with the wrong party, mistakes by advertisers, mistakes in the press that are likely to mislead others, misdirected communications that reflect a mistaken understanding, inquiries that reflect a mistaken understanding (as opposed to mere musing), and other types of misdirected communications, inquiries, and impressions whose import is less clear.

¶ 171         This factor has two components. First, the court must determine where the persons at issue fall within the spectrum of those whose views are relevant and how credible the persons are. Based on the discussion in Section III, categories of confused persons are relevant in roughly the following descending order: actual purchasers; potential purchasers with proximity to a purchasing decision; investors, suppliers, service providers, or others whose decisions may affect the good will of the trademark owner where such persons are in proximity to a relevant decision; advertisers or members of the media who are likely to impact the purchasing public; potential purchasers who are more remote from actual purchasers; potential investors, suppliers, service providers, or others whose decisions may affect the good will of the trademark owner where such persons are more remote from a relevant decision; persons charged with using a high degree of care to identify the parties correctly; and others whose reactions may be indicative of the reactions of persons in some relevant category.

¶ 172         Second, courts must consider bias and credibility. Factors indicating that the person might be biased include relationships with, or hostility towards, one of the parties or other financial or economic incentive to favor one of the parties. Factors that may bear on credibility include sheer contrast between the account of confusion and the other likelihood of confusion factors, the inability of the witness to recall specifics, whether the account is consistent with the witness's degree of sophistication, whether the plaintiff was involved in instigating the incidents, whether there are circumstances that bolster the credibility of an account, and whether the relevant persons are subject to cross examination (or, at least, to being deposed by the defendant).

¶ 173         Courts must consider whether the persons at issue exercised the appropriate level of care and attentiveness for the types of products or services at issue and for the type of interaction at issue in each case. The level of apparent care used by persons allegedly confused should be assessed in light of the market conditions relevant for determining likelihood of confusion. Consumers should not be held to a higher level of care than they would exercise in the marketplace. While confusion may erode a trademark owner's good will in contexts other than a mistaken purchase, where carelessness or inattentiveness occurs in circumstances that are not indicative of potential impact on the trademark owner's good will, then such confusion should be discounted. Thus, misdirected communications regarding warranties, service, or product returns are likely to be more relevant than communications that merely report impressions about a product.

¶ 174         Much like the level of care, the sophistication of persons purportedly confused should be assessed in light of marketplace conditions and in light of what such evidence foretells regarding persons in the marketplace. Persons purportedly confused should be held to a reasonable level of ordinary intelligence—but not more than would be expected of the average member of the consuming public for the relevant product or service. To the extent persons reporting actual confusion are more sophisticated, such confusion (if credible and if occurring in a context indicative of the reactions of others) may foretell broader confusion among persons more representative of the actual marketplace.

¶ 175         To some extent this factor overlaps with the type of evidence at issue. For example, a mistaken purchasing decision will exhibit a higher degree of confusion than a casual inquiry about whether two parties are related. However, courts should also consider the degree to which reports of confusion continue to surface over time—as an indicator of the reliability of incidents at any particular time to predict of likely confusion and as an indicator of whether incidents are likely attributable to some factor other than the parties' marks (like the temporary existence of only one party in a telephone directory), which may disappear over time.

¶ 176         This factor measures the degree to which evidence of actual confusion would be expected to surface by looking to whether there has been sufficient coexistence in the marketplace to produce confusion. For that to happen, the parties must have presented their marks to a common audience—either through direct competition, through sales as complimentary or related products or services, or through advertising to a common audience in a common territory or in overlapping channels of trade. In addition, the court must take intou account factors regarding the context of the coexisting uses that may make incidents of confusion more or less likely to surface—factors such as price of the goods or services, ease of communicating with the parties, the degree to which the parties communicate directly with consumers, the degree to which communications with intermediaries such as retailers are passed to the parties, the degree to which consumers are likely to realize their mistakes and report them to the parties, the degree to which consumers are likely to be deterred from coming forward by concerns such as embarrassment or the fear of becoming a witness in litigation, and the parties' recordkeeping practices. In light of the extent and nature of the particular coexisting uses, the court must then determine whether the coexistence has endured for a sufficient period that some amount of confusion should have surfaced if confusion were likely to occur.

¶ 177         The courts should also consider the practical ability of a party to obtain evidence of confusion that has surfaced. Circumstances affecting the ability of a party to obtain such evidence include whether consumers typically communicate with third parties regarding the products or services at issue instead of the parties themselves, the degree to which the parties' have had the opportunity to conduct discovery, and the opposing party's recordkeeping practices.

¶ 178         With the various other factors in mind, for each type of person and each type of incident, the quantity of incidents will be important but the degree of importance will vary with the strength of the evidence and the degree to which evidence of actual confusion should be expected. While courts should be wary of imposing any statistical significance threshold that could only be achieved properly in the context of a survey, courts should also be mindful of benchmarks that provide context, such as extent of sales, advertising, and publicity.

¶ 179         In light of the many textual and policy reasons underpinning the rule that evidence of actual confusion is not required, no inference should automatically be drawn against a plaintiff where instances of actual confusion are minimal or nonexistent. Rather, if the defendant seeks to use this factor to overcome other evidence of likely confusion offered by the plaintiff, it should be incumbent on the defendant to put the lack of evidence of actual confusion in context and to demonstrate its significance. In some cases, this task will be easy, such as where there has been extensive side-by-side competition over many years. In other cases, the task will be more difficult. By contrast, where the plaintiff seeks to establish the actual confusion factor as pointing to likely confusion, it should be incumbent on the plaintiff to put the evidence of actual confusion in context and to demonstrate its significance. Given the many problems with evidence of actual confusion and the many contextual nuances that affect the value of such evidence, presumptions in either direction from the presence or lack of such evidence are not appropriate. However, the strength of the inference drawn in either direction may vary depending upon the showing made by the party seeking to establish such an inference.

VI.    Conclusion

¶ 180         In sum, far from serving as a stable, objective determinant of likely confusion, the actual confusion factor has been a source of inconsistency, incoherence, unpredictability, and, indeed, confusion in trademark law. Evaluating evidence of actual confusion requires highly subjective judgments. To date, the courts have not yet articulated an approach to such evidence that promotes a consistent, coherent, and predictable approach to the evidence. A multi-factor test will best achieve that end and, by providing uniform guidance, should ultimately help courts discern between evidence of actual confusion having more objective value and evidence of actual confusion that is more prone to subjective interpretation.

ENDNOTES