Towards a Trademark Rule of Reason
By Daniel M. Lifton*
Download a PDF version of this article here.
Some perceive trademark protection as a reward for a mark owner’s labor in cultivating his business goodwill. However, among legal scholars and academics, the prevailing theoretical explanation for trademark protection is utilitarian, focusing on increasing consumer welfare. Based on the “search costs” theory of trademark law, legal protection is justifiable because trademarks produce two welfare-increasing effects. First, trademarks reduce consumer search costs. Second, trademarks incentivize producers to invest in product quality and consistency.
Importantly, not all means of reducing search costs maximize consumer welfare. For example, consumer search costs would be reduced if competition was eliminated and products were offered by single providers. But nobody supports the monopolization of markets as a desirable method of reducing search costs. It is generally believed that consumers are better served by competition, even though competitive markets require more searching than do markets with single providers. This conflict suggests that trademark law must strive to achieve the goal of reducing consumer search costs only insofar as doing so facilitates the functioning of a competitive market.
It follows that enforcement of trademark rights rests on the assumption that mark owners, acting as quasi-economic regulators, will prevail when their infringement claim runs parallel to the consumer welfare goal of promoting effective competition. All other conduct should be left unregulated. Trademark law built on enhancing competition should limit liability to conduct that has a net effect of harming competition and should avoid liability for conduct that has a net effect of benefiting competition.
Unfortunately, in the name of reducing search costs, courts have lost sight of trademark law’s underlying competition policy. Rather than develop a system in which consumer confusion is actionable only insofar as it relates to the competitive goals of trademark law, over time courts have created one in which consumer confusion is the harm itself. Because of this confusion-centric analysis, trademark liability has expanded over the past half-century to encompass many different forms of confusion, such as initial-interest confusion and post-sale confusion. Trademark’s expansion of actionable confusion, coupled with its distribution of proof burdens, has contributed to its departure from its goal of promoting competitive markets.
In this Note, I argue that the unitary per se rule is ill-suited for assessing the vast amount of confusion that trademark law now governs. Trademark liability should, instead, reflect the model set out in the field of antitrust, a body of law similarly tasked with condemning conduct that distorts the competitive markets. Antitrust teaches that liability should oscillate between rules and standards and that, in designing a binary liability scheme, a preference for reducing false positives is most appropriate. As applied to trademark law, infringement liability would reflect a similar binary regime.
This Article proceeds as follows. Part I provides an overview of the competition policy justification that grounds trademark law. Part II will discuss trademark infringement liability and the various confusion-based liability doctrines. Part III will describe the weaknesses of trademark law’s current liability scheme. Drawing inspiration from antitrust law in Part IV, I will argue for a reformation in trademark liability that reflects antitrust by featuring rules and standards and reducing false positives.
While it is well recognized that trademark law aims to promote competition, most accounts of trademark law begin with the two economic functions that trademarks serve. A trademark is a word, symbol, or other signifier used to distinguish a good or service produced by one firm from the goods or services of the other. The range of what constitutes a trademark is broad; it includes words, colors, building shapes, and even scents. However, regardless of what form they take, at their most basic level, trademarks communicate information to consumers about the source and quality of the products on which they are displayed.
Trademarks serve two critical functions in the marketplace. First, trademarks reduce consumer search costs. Consumers can rely on trademarks as a method of obtaining accurate information about a product, such as its source and quality, and of ensuring, based on that information, that the item has the desired characteristics. By providing the means for consumers to obtain purchase-relevant information without having to expend endless time and effort in search, trademarks reduce transaction costs and thereby enhance competition. The ability of consumers to rely on trademarks in order to distinguish between producers gives rise to the second function of trademarks in the marketplace. Specifically, trademarks allow producers to profit from the goodwill they cultivate among consumers. By restricting the ability of other producers to use identical or confusingly similar marks on their competing products, trademark law ensures that producers themselves, and not their competitors, reap the benefit of their investments in quality and consistency. This profit-motive incentivizes producers to continue to make such investments, which in turn enhances competition.
Again, while the reduction of consumer search costs and the encouragement of goodwill investment represent critical intermediate objectives of the trademark system, it is important not to lose sight of the fact that neither of these goals is an end in itself. The ultimate policy goal of trademark law is to facilitate the functioning of a competitive marketplace. Informed consumers make well-informed purchase decisions, which increases their overall utility and spurs producers to offer higher quality products. Thus, the underlying aim of this body of law is to encourage more competitive markets by improving the quality of information in those markets. Conceptualized this way, trademark law is a species of competition law.
The goal of promoting competition justifies not only the affirmative rights trademark law confers on markholders, but also the limitations that the law should place on those rights. A trademark law that is built on enhancing competition should limit liability to information-distorting conduct that has a net effect of harming competition and should avoid imposing liability for conduct that has a net effect of benefiting competition. As will be discussed in Part III, the current trademark liability regime offends these basic principles.
In the name of reducing search costs, courts focus on the narrow issue of consumer confusion when determining liability for trademark infringement.
The bedrock of a trademark infringement action is the likelihood of confusion test. The “likelihood of confusion” is the probability that an alleged infringer’s mark is the cause in fact of confusion in the minds of potential consumers. Liability depends on whether the defendant’s use of a mark is likely to cause consumers to be confused or deceived as to the source or nature of the defendant’s product or service. Courts use a multi-factor test to determine whether such a likelihood is present. Although the factors differ among the circuit courts, the traditional set of factors developed by the Second Circuit in Polaroid Corp. v. Polarad Electronics Corp. is illustrative of the typical factors considered. These factors include: (1) the strength of the plaintiff’s mark; (2) the degree of similarity between plaintiff’s and defendant’s marks; (3) the proximity of the products or services; (4) the likelihood that plaintiff will bridge the gap; (5) evidence of actual confusion; (6) defendant’s good faith in adopting the mark; (7) the quality of defendant’s product or service; and (8) the sophistication of the buyers.
Importantly, the plaintiff’s prima facie case starts and ends with a confusion analysis. If the plaintiff succeeds in establishing a likelihood of confusion, the court will hold the defendant liable for infringement, implicitly presuming that the confusion causes harm to the consumer, the mark owner, and the market.
B. Expansion of Trademark Liability Over Time
Traditionally, the scope of consumer confusion targeted by trademark law was limited to purchaser confusion as to the source of goods or services at the time of sale. This is known as point-of-sale confusion. However, over the last fifty years the number of forms of actionable confusion has expanded dramatically. For example, the point of purchase is no longer the only relevant period in which to assess confusion. Under the modern initial-interest confusion and post-sale confusion theories of liability, actionable confusion now extends to the periods before and after the transaction has taken place.
Additionally, the population of confused persons is no longer limited to purchasers or likely purchasers. In a claim of post-sale confusion, the confusion does not arise among the class of purchasing consumers; rather, actionable confusion arises among third-party observers who view the product on the street and mistake the source of that product. For example, Levi Strauss launched a post-sale confusion theory to enjoin Lois Sportswear from selling designer denim jeans with stitching that resembled Levi’s trademarked stitching pattern. Even though the packaging and labeling of Lois’s jeans eliminated any possibility of consumer confusion at the time of purchase, Levi claimed that third parties who view the jeans when worn in public would mistakenly infer from the stitching pattern that Lois’s jeans were Levi’s jeans.
Under a theory of initial-interest confusion, the plaintiff claims that consumers are attracted to the defendant’s product due to the resemblance between the defendant’s mark and that of the plaintiff, but then realizes the true source of the goods before the sale is consummated. The doctrine originated from the prohibition of bait-and-switch advertising, where the concern was that consumers, once drawn into the decision-making process, may not back out even upon discovering that the offered product or service is not what they expected.
In recent years, however, courts have gone well beyond this traditional instance of initial-interest confusion. Some have found liability even when there is no confusion beyond a moment of uncertainty; others have gone even further, holding defendants liable for using another’s mark merely to gain attention and hence not to confuse. For example, courts have used an initial-interest confusion theory to enjoin the use of a competitor’s mark to attract customers to websites. In Brookfield Communications v. West Coast Entertainment, the plaintiff, which sold software that allowed customers to look up movie information, was able to prevent the defendant, an internet video rental and movie information supplier, from using the plaintiff’s mark, MOVIEBUFF, as part of a website metatag. Use of the plaintiff’s mark in a metatag means that an internet user who enters the term “MOVIEBUFF” into a search engine would pull up a list of websites that includes both the plaintiff’s and the defendant’s domain names. Although the court acknowledged that the user was not likely to be confused about any connection with the plaintiff after visiting the defendant’s site, it was enough for the court that she might click on the defendant’s site believing it to be related to the plaintiff, simply because it appeared in the same search results.
Despite the expansion of actionable confusion across both the temporal axis—to include confusion occurring before and after purchase, as well as at the time of purchase—and the consumer population axis—to include non-purchasers, as well as purchasers—liability for all three types of actionable confusion is governed by the same likelihood of confusion test outlined above. Trademark doctrine has expanded to impose liability for new types of confusion, but no mechanism has developed to distinguish between confusion that always or almost always results in trademark-related harms from confusion that is less likely to produce those harms. Upon demonstrating that the defendant’s conduct is likely to create confusion pre, post or at the point of sale, such conduct is deemed illegal per se. No further inquiry is made into whether that particular form of confusion is likely to harm or benefit consumers, producers, or competition at large. In this way, the confusion analysis operates as a per se rule.
By implementing the same likelihood of confusion test as a threshold for liability under all three theories, courts implicitly make two key assumptions: (1) that each type of confusion is equally likely to result in trademark-related harm, and (2) that the defendant’s proscribed conduct provides insignificant, if any, procompetitive benefits. However, these assumptions are unsupported in cases of initial-interest and post-sale confusion cases, which have lower likelihoods of competitive harm and may, in fact, increase competition.
As stated above, upon finding a likelihood of confusion, courts simply presume harm to the plaintiff and rule in her favor. This presumption is justified when the confusion arises at the point of sale. According to the current doctrinal reasoning, causing confusion at the point of sale undermines both the intermediate goals of trademark law, namely reducing search costs and incentivizing producer investments in product quality, and trademark law’s ultimate goal of promoting competitive markets. Consumers who are confused as to the source or nature of a good at the time of purchase are harmed in two ways—first, they end up with unwanted or misidentified products, and second, they can no longer rely on the trademark in the future to relay accurate information, thereby increasing their search costs. Similarly, producers are harmed twofold: first, they will suffer lost sales, and second, they may suffer injury to their reputations when consumers mistakenly identify them as the source of shoddy products. Therefore, when consumers are confused about source at the point-of-sale, the defendant’s conduct is harmful to competition generally. Because confusion at the point-of-sale results in the archetypal harms that trademark law seeks to avoid, presuming harm from a showing of a likelihood of such confusion is justified and consistent with overarching trademark principles.
The relationship between consumer confusion and harm to both consumers and producers is far more attenuated in the context of initial-interest and post-sale confusion. In the context of initial-interest confusion, consumer confusion, though ultimately dispelled, may increase search costs. This harm can be better understood by way of a helpful analogy. A job-seeker embellishes his skills and background on his resume, is invited for an interview by an interested employer and then, at the interview, admits to the employer that his inflated resume is not completely accurate. In this scenario, the misrepresentation enables the job-seeker to obtain a coveted interview, giving him a clear advantage over other interested parties with the same skill set and background who honestly stated their achievements on their resumes. Similarly, a producer who misrepresents the source of its products or services by likening itself to another desirable producer may put others at a competitive disadvantage even though consumers are not confused at the time of purchase.
However, this potential harm may not always be realized. The scenario described above is one in which the misrepresentation effectively supplants a competitor from being considered, impairing their ability to compete effectively in the market. However, if the misrepresentation merely offers consumers an alternative, without preventing the competitor’s products from being considered at all, no market harm will manifest. In other words, if the employer in the previous example offered an infinite number of interview slots and bore no opportunity costs in conducting them, no competitive injury would result from the job-seeker’s misrepresentations.
Importantly, whether or not the producer is able to effectively compete depends on the costs of the search and the ease with which the purchaser’s confusion is dispelled. For example, a consumer in search of a pain reliever may initially be looking for the Advil brand. While searching through the shelves, she comes across a pain reliever with packaging resembling that of Advil, but upon closer examination realizes that it is produced by a generic brand and offered at a cheaper price. The generic contains ibuprofen—the active ingredient of Advil—making it chemically identical to Advil. If the consumer sincerely favors Advil, possibly out of concerns for quality assurance, her search costs of finding it have not been raised in any meaningful respect—she puts the generic down, sees the Advil on the same shelf, and moves on. Because the effect on search costs is trivial, the mark owner does not lose the capacity to compete on the basis of price, quality, or reputation. This inference is supported by the fact that the confusion did not displace the competing product, rather the purchaser considered the products of both producers and made an informed, unconfused decision.
Furthermore, unlike in the point-of-sale context, where consumer confusion is directly related to the purchasing decision and is therefore the proximate cause of the producer’s harm, in the post-sale context, there is no indication that confused parties are likely to be potential consumers of the plaintiff’s product. In fact, there is no indication that the observations made by those who are confused are likely to be material to a purchasing decision. Therefore, the link between the confusion caused and harm felt by producers is severely weakened; the latter cannot be presumed from the former.
Trademark-relevant harm arises post-sale only when a third-party observes a product bearing the defendant’s allegedly confusing mark, inaccurately identifies its source as the plaintiff, makes an adverse judgment as to the quality of the product, attributes that judgment to the plaintiff, and subsequently refrains from purchasing products correctly attributed to the plaintiff based on that prior false association. This is a tenuous chain of events. Only in these narrow circumstances, where the confusion of a bystander is linked to a negative purchasing decision, would a plaintiff alleging post-sale confusion suffer harm that undermines the goals of trademark law. Because harm to the producer, in the form of lost sales or harm to reputation for quality, is significantly less likely to result from confusion in this context, such harm cannot and should not be presumed from mere confusion.
A review of judicial decisions reveals that in post-sale confusion cases, courts have credited theories of harm that lie outside the confines of trademark interests. This is particularly apparent in cases brought by luxury brands seeking to enjoin the production of knock-off, or look-alike, goods—the same class of cases in which post-sale confusion was originally invoked. In Hermès International v. Lederer de Paris Fifth Avenue, Inc., the Second Circuit identified the harms flowing from post-sale confusion as the following:
“[T]he purchaser of an original is harmed by the widespread existence of knockoffs because the high value of originals, which derives in part from their scarcity, is lessened. . . . A loss [to the public] occurs when a sophisticated buyer purchases a knockoff and passes it off to the public as the genuine article, thereby confusing the viewing public and achieving the status of owning the genuine article at a knockoff price.”
The first harm affects the purchasers of the original, authentic luxury goods in that the dilution of the market with similar goods diminishes the exclusivity of the original. The second harm affects the general public who will be less able to attribute the “appropriate” status to other members who use products bearing luxury marks. These harms laid out by the Second Circuit and many other courts are not the product of anticompetitive conduct––rather, they are simply the result of competition over something that has consumptive value, in this case the mark itself. Because trademark is a species of competition law, protecting against these harms is not only unfounded, but is counterproductive to the trademark cause in that it undermines the overarching goal of promoting competition.
In addition to the faulty presumption of harm across the varying liability theories, another shortcoming of the current trademark liability regime is its failure to consider the potential procompetitive benefits of the defendant’s conduct. Focusing on the narrow question of whether the defendant’s conduct is likely to confuse consumers, without assessing the possible procompetitive effects of that conduct, may be appropriate in the context of point-of-sale confusion because the potential anticompetitive effects are so substantial. On balance, the likelihood that the conduct is actually a net benefit to competition is extremely low.
In contrast, confusion that arises prior to the sale is more likely to result in procompetitive benefits. In the Advil example described above, if the purchaser is satisfied that the generic pain reliever is an adequate substitute for Advil, it follows that her “initial confusion” is what enabled her to identify that substitute and thus to cut costs. Because she was not previously familiar with the alternative option, the temporary, pre-sale confusion she experienced worked to her benefit by broadening her awareness of Advil’s cheaper competitors. This consumer’s preference was not for Advil, but rather for the most economical ibuprofen product she could find. By resembling Advil, the trade dress of the generic company signals to purchasers that it is in the same product category. In this scenario, a trademark plaintiff may complain that the conduct injures it personally, due to the lost sale, even though the conduct actually benefits competition overall. Some courts recognize this benefit in trade dress cases, and the same dynamic is likely to be at play in initial-interest confusion cases involving other types of marks.
These potential benefits also extend to the Internet. In the online context, for instance, the use of metatags in search engines can give rise to a claim of initial-interest confusion. However, this same conduct can reduce consumer search costs and provide easy access to comparative quality and price information, thereby enhancing competition. Competition on the merits improves consumer welfare by providing consumers with competing goods. As long as any confusion is dispelled by the time consumers buy goods or services, which a theory of initial-interest confusion assumes, consumers may have actually found alternative goods at least as desirable as the mark owners’ goods.
Similarly, in the context of post-sale confusion, the defendant’s conduct does in fact have potential procompetitive effects. The factual presumption in a claim of post-sale confusion is that the actual purchasers of the product bearing the defendant’s confusing mark are fully aware of its source at the time of purchase. These consumers entered into the transaction with the full and accurate knowledge that, although the sneakers resembled those of the plaintiff, they were in fact made by the defendant. This is important because the purchasers in a post-sale confusion world benefit from a competitive market for the merchandised goods and therefore benefit from the defendant’s conduct.
The heated and long-fought battle between luxury and knockoff brands is demonstrative of this point. Luxury brands are often highly litigious and aggressive in their efforts to protect one of the most important aspects of their businesses, namely their trademarks. However, the countervailing consumer interest is the availability of products with similar aesthetic appearance to prestige goods but at cheaper prices. The recent rise in companies producing knockoff goods, such as H&M, Forever21, and Zara, provide further evidence of the high consumer demand for cheaper alternatives to designer goods. The availability of designer look-alikes can be characterized as a procompetitive benefit of the allegedly infringing conduct at issue in post-sale confusion cases, and it is this benefit that is lost when producers are enjoined from producing similar goods.
Resolving post-sale confusion claims without consideration for the benefited consumers effectively subordinates their interests to the interests of producers who seek to protect the prestige and exclusivity associated with their luxury goods and of consumers of luxury goods who wish to project their prestige to society. Whether or not it is economically or morally desirable to provide legal protection for prestige and exclusivity is a discussion for another day. However, it is plain that such interests are outside the scope of interests that trademark law is designed to protect. As such, claims of post-sale confusion that result in enjoining defendant’s conduct risk stifling competition and innovation.
Trademark’s unitary liability regime fails to take into account the different market effects, both harmful and beneficial, produced by the various forms of confusion. This failure likely gives rise to an influx of “false positives,” that is, cases that wrongly find violations where no trademark-related harm exists. These false positives produce associated social costs by preventing or chilling procompetitive conduct. Other areas of competition policy, such as antitrust, deal with similar risks of error by applying a combination of rules and standards that attempt to minimize the total cost of false positives. This approach would be appropriate in the world of trademark.
It is only natural to look to antitrust and its terms for inspiration in redesigning trademark liability because, like trademark law, antitrust is designed to ensure that the competitive market functions well. In this sense, both bodies of law are species of competition policy. Trademark law achieves this goal by ensuring that consumers can rely on signs and symbols in order to glean information and ultimately make informed purchasing decisions. Antitrust law accomplishes this goal by deterring collusive and monopolistic conduct that impairs valuable competition. Additionally, both trademark law and antitrust law represent affirmations of, rather than departures from, the competitive model that drives the U.S. economy. Therefore, inherent in both legal doctrines is the default assumption that under ordinary circumstances, competitive markets will ensure efficient resource allocation and bring consumers the highest quality products at the lowest prices. As we will see in the following section, this default assumption plays an important role in the way liability rules are structured in competition law. Of course, the goal of protecting competition is much more explicit in the field of antitrust and as a result, its legal rules are tailored to achieve that goal. Trademark law, on the other hand, has had a more checkered history, resulting in liability rules that are disconnected from competition policy.
For these reasons, antitrust should serve as guidance in the process of realigning trademark liability with its original underlying purpose of promoting competition.
Unlike trademark law, antitrust law operates under a binary liability structure. This two-tiered liability scheme divides conduct into two general categories: per se violations and rule of reason violations. The per se category is effectively a rule, making the treatment of the conduct quite simple. If a plaintiff shows that the defendant deliberately engaged in such conduct, courts will hold the defendant liable for engaging in anticompetitive behavior in violation of the Sherman Act. Importantly, a plaintiff need not provide economic evidence of actual harm to competition; instead, harm to competition is presumed, and the defendant is held liable per se.
The second liability category involves a standard, confusingly referred to as the rule of reason standard. In rule of reason cases, courts require a well-developed investigation into the competitive effects of the defendant’s behavior. In other words, plaintiffs must conduct a more searching investigation into the harms and benefits produced by the defendant’s allegedly anticompetitive conduct. In order to establish a prima facie case, plaintiffs must demonstrate that the conduct in question has harmed or will harm competition. If successful, the burden shifts to defendants to offer plausible economic theories that justify their conduct as procompetitive. If the defendant does present evidence of a procompetitive justification, the plaintiff will have an opportunity to prove that the same procompetitive effects could have been achieved by a less restrictive alternative. If no less restrictive alternative is available, the court will attempt to calculate the net effect of the defendant’s conduct, by balancing the procompetitive justification of the conduct against its potential for anticompetitive harm.
The scope of each of the two categories of liability differs quite drastically. The per se category is narrow in scope—it only includes conduct that courts have previously identified as harmful to competition and that, as a whole, have no redeeming procompetitive justifications. Some examples include horizontal price fixing, bid rigging, and dividing markets. These classes of conduct are archetypally harmful to competition and are therefore deemed to be per se anticompetitive, with no opportunity to put forth redeeming justifications. In comparison, the rule of reason category is much wider in scope. It includes conduct that has ambiguous effects on competition, and which therefore requires a more searching investigation into the harms and benefits produced by the conduct.
The makeup of these two categories is not random, but rather is driven by the courts’ avoidance of certain undesirable outcomes, known as false positives.
Although the social costs of erroneous outcomes play a role in the development of all legal regimes, they are particularly important for laws that police marketplace behavior, like antitrust and trademark. Due to the limitations of current economic empirical analysis, legal decisions regarding the marketplace are necessarily made under uncertainty. The chronic degree of uncertainty throughout competition law makes mistaken conclusions inevitable to some degree—but it does not render them any less costly. Therefore, competition laws display a unique fixation with error costs.
Errors come in two forms—false positives and false negatives. A false positive occurs when a result is reached that should not have been reached. By contrast, a false negative occurs when a result is not reached but should have been. The social costs of these errors are the product of two factors: (1) the probability that the error will occur, and (2) the magnitude of the social cost when it does occur. It is important to distinguish between the two types of errors because they may produce different social costs. Many laws reduce the frequency of one type of error only to increase the frequency of the other, and thus, all legal regimes must determine which of the two types of errors is most crucial to avoid. For example, the rule imposing a beyond-a-reasonable-doubt burden of persuasion in criminal cases reduces the frequency of erroneous convictions but increases the frequency of erroneous acquittals. The rule can be justified on the grounds that, in terms of lost liberty, the social cost of an erroneous conviction is much higher than the social cost of an erroneous acquittal.
Like criminal law, antitrust rules and burdens of proof reflect a preference for reducing false positives, while tolerating a somewhat increased possibility of false negatives. Though not applied as an explicit rule, this preference has significantly influenced what types of conduct fall into each of the two categories of antitrust liability. Indeed, minimizing false positives and mitigating their chilling effects on procompetitive conduct are often cited as justifications for abandoning some of the per se rules that were applied prior to the late 1970s, especially with regard to vertical agreements.
This is most clearly illustrated in Leegin Creative Leather Products, Inc. v. PSKS, Inc., where the Supreme Court addressed the question of whether resale-price maintenance (“RPMs”) should be analyzed under a per se rule. The Court found that RPMs have ambiguous welfare effects on the market. In other words, they could have procompetitive or anticompetitive effects, depending upon the circumstances in which the resale agreements are formed. The Court acknowledged the risk of false positives under a per se rule and the associated costs of “prohibiting procompetitive conduct the antitrust laws should encourage” and thus, notwithstanding the risk of allowing unlawful conduct to go unpunished, it decided to abandon the per se rule in favor of the rule of reason standard. As applied to RPMs, the Court explained that the per se rule proscribes a significant amount of procompetitive conduct, making RPMs ill-suited for per se condemnation.
By adopting the rule of reason standard for conduct that produces ambiguous welfare effects, courts have not only expressed a preference for reducing false positives, but also an acceptance of increased false negatives. This preference finds its roots in a principle of competition policy that was famously espoused by Judge Frank Easterbrook: Preventing procompetitive behavior is more harmful than allowing anticompetitive behavior. The assumption is that the social costs of false positives far exceed the social costs of false negatives. Additionally, the correction costs of false positives are much higher than those of false negatives because court decisions have lasting impact on behavior in the market. By contrast, false negatives would largely be dissipated by the self-correcting tendencies of markets. As a species of competition policy, these same principles apply with equal force to trademark law.
Trademark law’s current unitary liability scheme should be restructured to resemble something approximating the binary structure applied in antitrust law. Under this new, but not novel, liability regime, some forms of confusion would fall in the per se category, while others would fall within a rule of reason category. The per se form of liability would serve as a rule—once a likelihood of confusion is shown, trademark harm is presumed and the defendant is held liable. Under the rule of reason category, a plaintiff would face a two-pronged burden. First, she must establish a likelihood of confusion under the current multifactor likelihood of confusion test. Second, she would be required to show that the confusion is likely to produce a trademark-related harm. If the plaintiff successfully satisfies her burden, the burden would shift to the defendant to put forth procompetitive justifications for her conduct. Finally, if the defense succeeds, the court would balance the harms against the procompetitive justifications before reaching its decision.
The question remains, what is the guiding principle by which the types of confusion should be delegated to the rule category or the standard category? By treating consumer confusion as a proxy for harm for all claims of infringement, trademark liability in its current form guards most vigorously against the wrong type of error, false negatives, and as a result does more harm to competition than good. For the reasons discussed above in the context of antitrust law, it is now clear that a reformed trademark liability structure should be designed to reduce false positives and to tolerate a somewhat increased possibility of false negatives. Where the chance of a harmful false negative is remote, trademark liability should lean toward protecting potentially procompetitive behavior. Trademark law must incorporate this priority when determining the relevant liability rule for different claims of confusion.
According to this guiding principle, conduct with ambiguous effects on the market would be held under a rule of reason standard. The expansive forms of confusion-based liability, namely initial-interest and post-sale confusion, would be analyzed within this rule of reason framework. Conduct that gives rise to these forms of confusion produces ambiguous welfare effects on the market—it could have procompetitive or anticompetitive effects, depending upon the circumstances. A rule that presumes harm from these types of confusion results in an increased number of false positives and, in turn, such errors chill procompetitive conduct. Bearing in mind the aforementioned preference for reducing false positives, a rule of reason standard is the more appropriate treatment for these types of claims, as it will ensure that liability is imposed only when the underlying conduct produces greater market harms than benefits. Furthermore, for the reasons described above, the likelihood of harmful false negatives is much lower in the context of these types of confusion.
Mirroring antitrust, the per se category of trademark liability would consist of conduct that in most cases results in a net-negative effect on the market. This class of conduct would include confusion as to source or quality at the point-of-sale. Because confusion at the point of sale is harmful to both consumers and producers in most instances in which it arises, it merits a more stringent per se rule that treats confusion as a proxy for harm. It follows that a per se rule conclusively presuming harm from confusion at the point of sale will eliminate erroneous acquittals and their associated costs. To be sure, the rule also increases erroneous liability findings, but given the fact that there is likely no benefit that the underlying conduct produces, the increase would be slight and the social costs not terribly high. Applying a standard, rather than a rule, would result in a large increase in seriously harmful false negatives. Those errors are much more likely to produce anticompetitive effects and thus relatively high social costs and a decrease in overall welfare.
Trademark law should be seen, first and foremost, as a law aimed at promoting competition. With this framework in mind, the doctrine should be reformed to differentiate between confusion likely to have a net harm to competition and confusion likely to have a net benefit. To do so, trademark liability should take its cue from the world of antitrust, a doctrine similarly grounded in the protection of competition. Antitrust law teaches that liability should oscillate between rules and standards and that a preference for reducing false positives, even at the expense of causing a slight increase in false negatives, should be adopted. Accordingly, conduct with ambiguous effects on the market should be reviewed under a rule of reason standard. As such, the expanded forms of confusion-based liability, namely initial-interest and post-sale confusion, should be analyzed within a rule of reason framework.
 See Mark P. McKenna, The Normative Foundations of Trademark Law, 82 Notre Dame L. Rev. 1839, 1873–93 (2007) (setting out the Lockean account of trademark law).
 See William M. Landes & Richard A. Posner, Trademark Law: An Economic Perspective, 30 J.L. & Econ. 265 (1987) (setting out the economic account of trademark law).
 See Mark P. McKenna, A Consumer Decision-Making Theory of Trademark Law, 98 Va. L. Rev. 67, 73–74 (2012) (explaining the “search costs” theory of trademark law).
 See McKenna, supra note 3, at 87 n.45 (“Indeed, the Supreme Court has rejected the argument that horizontal agreements to eliminate credit sales can be justified under the antitrust laws on the ground that an industry-wide agreement reduces the cost of learning price and credit terms.”).
 Stacey L. Dogan & Mark A. Lemley, A Search-Costs Theory of Limiting Doctrines in Trademark Law, 97 Trademark Rep. 1223, 1227 (2007) [hereinafter Dogan & Lemley, Search-Costs]; see also Stacey L. Dogan & Mark A. Lemley, The Merchandising Right: Fragile Theory or Fait Accompli?, 54 Emory L.J. 461, 467 (2005) [hereinafter Dogan & Lemley, Merchandising Right].
 See McKenna, supra note 3, at 71 (“Anything that can be characterized in confusion-based terms seems to raise search costs, and if search costs are the harm to be avoided, then anything that causes confusion ought to be at least prima facie actionable.”).
 See Barton Beebe & C. Scott Hemphill, The Scope of Strong Marks: Should Trademark Law Protect the Strong More Than the Weak?, 92 N.Y.U. L. Rev. 1339, 1387–89 (detailing the ways in which several elements of trademark doctrine attempt to achieve this goal); John F. Coverdale, Comment, Trademarks and Generic Words: An Effect-on-Competition Test, 51 U. Chi. L. Rev. 868, 869 (“The law regulating trade and commerce frequently seeks to promote competition as a means of allocating resources efficiently and insuring reasonable prices.”).
 These economic functions have dominated both judicial and scholarly accounts. See Qualitex Co. v. Jacobsen Prods. Co., 514 U.S. 159, 163–64 (1995) (stating that trademark law “reduce[s] the customer’s costs of shopping and making purchasing decisions [and] . . . helps assure a producer that it (and not an imitating competitor) will reap the financial, reputation-related rewards associated with a desirable product”) (internal citations omitted); Union Nat’l Bank of Tex., Laredo, Tex. v. Union Nat’l Bank of Tex., Austin, Tex., 909 F.2d 839, 844 (5th Cir. 1990) (“The idea is that trademarks are ‘distinguishing’ features which lower consumer search costs and encourage higher quality production by discouraging free-riders.”); 1 J. Thomas McCarthy, McCarthy On Trademarks And Unfair Competition § 2:3 (5th ed.) (quoting William N. Landes & Richard A. Posner, The Economics of Trademark Law, 78 Trademark Rep. 267, 267 (1988), for the proposition that trademark law is best understood as “trying to promote economic efficiency”); Barton Beebe, The Semiotic Analysis of Trademark Law, 51 UCLA L. Rev. 621, 623–24 (2004) (“The Chicago School of law and economics has long offered a totalizing and, for many, quite definitive theory of American trademark law. . . . The influence of this analysis is now nearly total. . . . No alternative account of trademark doctrine currently exists.”).
 15 U.S.C. § 1127 (1994).
 See, e.g., U.S. Reg. No. 1,078,312, Nov. 29, 1977 (APPLE for computers).
 See, e.g., Qualitex, 514 U.S. at 159.
 See, e.g., Two Pesos, Inc. v. Taco Cabana, Inc., 505 U.S. 763 (1992).
 See U.S. Reg. No. 5,467,089, May 15, 2018 (for toy modeling compounds, where “[t]he mark is a scent of a sweet, slightly musky, vanilla fragrance, with slight overtones of cherry, combined with the smell of a salted, wheat-based dough”).
 Although trademarks originally indicated source explicitly, consumers today rely on these marks principally for information about product features and quality, which—in turn—depend upon consistency of source. See Nicholas S. Economides, The Economics of Trademarks, 78 Trademark Rep. 523, 527 (1988) (“Presently the trademark typically identifies the product (the full combination of features that constitute the product), and its role of identifying the source is secondary in the minds of consumers.”).
 See Robert G. Bone, Enforcement Costs and Trademark Puzzles, 90 Va. L. Rev. 2099, 2105 (2004).
 This ultimate goal is illustrated in several areas of trademark law. For example, the non-protectability of generic and functional marks, and the defense of nominative fair use. See Beebe & Hemphill, supra note 10, at 1387 (“Limitations on the scope of a mark reduce consumer search costs, freeing up rivals to use similar marks and thereby increasing industry supply and consumer welfare.”).
 See Maureen A. O’Rourke, Shaping Competition on the Internet: Who Owns Product and Pricing Information? 53 Vand. L. Rev. 1965, 1968 (2000) (describing conditions for perfectly competitive market).
 See Dogan & Lemley, Merchandising Right, supra note 7, at 467; see also Dogan & Lemley, Search-Costs, supra note 7, at 1224 (“The evolution of trademark law reflects a continual balancing act that seeks to maximize the informational value of marks while avoiding their use to suppress competitive information.”).
 See Landscape Forms, Inc. v. Columbia Cascade Co., 113 F.3d 373, 379 (2d Cir. 1997) (“[T]he Lanham Act must be construed in light of a strong federal policy in favor of vigorously competitive markets, which is exemplified by the Sherman Act and other antitrust laws.”); Coverdale, supra note 10, at 870 (“Because the policy of the trademark law is to promote competition, a trademark, unlike a patent or copyright, affords no monopoly over the product to which it is affixed . . . indeed, the Supreme Court has noted that there is a strong federal policy that goods unprotected by patents or copyrights should be copyable by anyone.”).
 See Two Pesos, Inc. v. Taco Cabana, Inc., 505 U.S. 763, 769 (1992) (“It is, of course, also undisputed that liability under § 43(a) requires proof of the likelihood of confusion.”); Restatement Third, Unfair Competition § 20, comment d (1995) (“The term ‘likelihood of confusion’ has long been used to describe the standard of liability for trademark infringement in actions at common law and under federal and state trademark and unfair competition statutes.”).
 Lanham Act § 32, 15 U.S.C.A. § 1114(1) (2018).
 See 4 McCarthy on Trademarks and Unfair Competition § 24:28 (5th ed.) [hereinafter McCarthy] (“Through decades of case law precedent and the influence of the Restatement, the federal courts have developed a multi-factor test to assist in the difficult determination of whether there is or is not a likelihood (probability) of confusion.”).
 The following cases set forth the factors considered by the Circuits: First Circuit, see Keds Corp. v. Renee Int’l Trading Corp., 888 F.2d 215, 222 (1st Cir. 1989); Third Circuit, see Merchant & Evans, Inc. v. Roosevelt Bldg. Prods. Co., 963 F.2d 628, 637 (3d Cir. 1992); Fourth Circuit, see Anheuser-Busch, Inc. v. L & L Wings, Inc., 962 F.2d 316, 320 (4th Cir. 1992); Fifth Circuit, see Sno-Wizard Mfg., Inc. v. Eisemann Prods. Co., 791 F.2d 423, 428 (5th Cir. 1986); Sixth Circuit, see Homeowners Group, Inc. v. Home Mktg. Specialists, Inc., 931 F.2d 1100, 1106 (6th Cir. 1991); Seventh Circuit, see Smith Fiberglass Prods., Inc. v. Ameron, Inc., 7 F.3d 1327, 1329 (7th Cir. 1993); Eighth Circuit, see Squirtco v. Seven-Up Co., 628 F.2d 1086, 1091 (8th Cir. 1980); Ninth Circuit, see E. & J. Gallo Winery v. Gallo Cattle Co., 967 F.2d 1280, 1290 (9th Cir. 1992); Tenth Circuit, see Coherent, Inc. v. Coherent Techs., Inc., 935 F.2d 1122, 1125 (10th Cir. 1991); Eleventh Circuit, see Dieter v. B&H Indus. of Sw. Fla., Inc., 880 F.2d 322, 326 (11th Cir. 1989); Federal Circuit, see In re E.I. DuPont DeNemours & Co., 476 F.2d 1357, 1361 (C.C.P.A. 1973).
 287 F.2d 492 (2d Cir. 1961).
 See id. at 495.
 It is important to note that modern trademark doctrine has expanded beyond its traditional core in many ways. Trademark law now recognizes more types of symbols as protectible than ever before and extends protection beyond the trademark owner’s primary market. However, while these issues are important and require further discussion, they lie outside the scope of my analysis. In this article, I focus on the doctrinal expansion of the circumstances that create a basis for liability and specifically on the expansion of actionable confusion to now include confusion at every stage of the transaction.
 Under the Lanham Act of 1946, to establish trademark infringement, the plaintiff had to prove that the infringing mark was “likely to cause confusion or mistake or to deceive purchasers as to the source of origin of such goods.” Lanham Act, Pub. L. No. 79-489, § 32(1), 60 Stat. 427, 437 (1946) (codified as amended at 15 U.S.C. § 1114(1) (1994)). Because of the express reference to purchasers, courts accordingly focused their likelihood of confusion examination on whether actual purchasers were likely to buy a product bearing an infringing mark while mistakenly believing it to be the plaintiff’s product.
 See Jennifer E. Rothman, Initial Interest Confusion: Standing at the Crossroads of Trademark Law, 27 Cardozo L. Rev. 105, 160–61 (2005) (describing the history of initial-interest confusion).
 See Jeremy N. Sheff, Veblen Brands, 96 Minn. L. Rev. 769, 776–77 (2012) (explaining how the idea of point-of-sale confusion has dramatically expanded with the creation of doctrines like post-sale confusion); Lois Sportswear, U.S.A., Inc. v. Levi Strauss & Co., 799 F.2d 867 (2d Cir. 1986); Bone, supra note 19, at 609–10 (discussing the Lois Sportswear case).
 Ferrari S.P.A. v. Roberts, 944 F.2d 1235, 1245 (6th Cir. 1991) (where the plaintiff’s proposition was that “members of the public, but not necessarily purchasers, were actually confused by the similarity of the products”).
 Lois Sportswear, 799 F.2d at 867.
 The alleged harm to the trademark owner is not that the third-party observers will go on to wrongly buy the defendant’s product—that would be point-of-sale confusion. Rather, the harm is that the third-party observers will not buy the plaintiff’s product due to the misinformation they received upon viewing the defendant’s product, such as the notion that the real item is low quality. See Sheff, supra note 35, at 802.
 See Promatek Indus., Ltd. v. Equitrac Corp., 300 F.3d 808, 812 (7th Cir. 2002), as amended (Oct. 18, 2002) (“Initial interest confusion, which is actionable under the Lanham Act, occurs when a customer is lured to a product by the similarity of the mark, even if the customer realizes the true source of the goods before the sale is consummated.”); see also Grotrian, Helfferich, Schulz, Th. Steinweg Nachf. v. Steinway & Sons, 523 F.2d 1331, 1342 (2d Cir. 1975) (finding “initial confusion” when the declaratory plaintiff used the mark GROTRIAN-STEINWEG for pianos even though no consumers ultimately purchased the plaintiff’s pianos believing them to be STEINWAY pianos).
 See, e.g., Brookfield Commc’ns, Inc. v. W. Coast Entm’t Corp., 174 F.3d 1036, 1062–63 (9th Cir. 1999) (finding trademark liability, even though consumer confusion was not likely, because consumers might be diverted to defendant’s website).
 174 F.3d 1036 (9th Cir. 1999).
 Id. at 1062–63 (holding the defendant liable for creating “initial interest confusion” by using the plaintiff’s website in its metatag terms and by diverting people to its website through confusion about the domain name). But see Network Automation, Inc. v. Advanced Sys. Concepts, Inc., 638 F.3d 1137, 1148–49 (9th Cir. 2011) (limiting Brookfield to domain name disputes).
 See Qualitex Co. v. Jacobsen Prods. Co., 514 U.S. 159, 160–66 (1995); see also Restatement (Third) Of Unfair Competition § 2 cmt. a (1995) (“As confidence in the truth of advertising diminishes, prospective purchasers may be forced to expend additional resources.”).
 See Smith v. Chanel, Inc., 402 F.2d 562, 566 (9th Cir. 1968) (“Preservation of the trademark as a means of identifying the trademark owner’s products . . . makes effective competition possible in a complex, impersonal marketplace by providing a means through which the consumer can identify products which please him and reward the producer with continued patronage. Without some such method of product identification, informed consumer choice, and hence meaningful competition in quality, could not exist.”).
 See Bone, supra note 19, at 2108 (“[I]f consumers lacked the ability to distinguish one brand from another, firms would have no reason to create brands with more costly but higher quality characteristics.”).
 See id. (“In such a situation, it is not possible to say that the misrepresentation caused no competitive damage.”).
 See id.; Michael Grynberg, The Road Not Taken: Initial Interest Confusion, Consumer Search Costs, and the Challenge of the Internet, 28 Seattle U. L. Rev. 97, 112 (2004).
 See Landes & Posner, supra note 2, at 275 (“The fact that two goods have the same chemical formula does not make them of equal quality to even the most coolly rational consumer. That consumer will be interested not in the formula but in the manufactured product and may therefore be willing to pay a premium for greater assurance that the good will actually be manufactured to the specifications of the formula.”).
 See generally Kal Raustiala & Christopher Sprigman, Rethinking Post-Sale Confusion, 108 Trademark Rep. 881, 881–891 (2018) (identifying the chain of events needed to create post-sale observer confusion).
 See Mastercrafters Clock and Radio Co. v. Vacherin-Constantin Le Coultre Watches, Inc., 221 F.2d 464 (2d Cir. 1955).
 219 F.3d 104 (2d Cir. 2000).
 Id. at 108–109.
 See, e.g., Rolex Watch U.S.A., Inc. v. Canner, 645 F. Supp. 484, 495 (S.D. Fla. 1986) (“Others who see the watches bearing the Rolex trademarks on so many wrists might find themselves discouraged from acquiring a genuine because the items have become too common place and no longer possess the prestige once associated with them.”); Ferrari S.P.A. v. Roberts, 944 F.2d 1235, 1244 (6th Cir. 1991); I.P. Lund Trading ApS v. Kohler Co., 163 F.3d 27, 44 (1st Cir. 1998); Gucci Am., Inc. v. Dart, Inc., 715 F. Supp. 566, 567 (S.D.N.Y. 1989) (“Others will be discouraged from acquiring a genuine Gucci because the items have become too commonplace and no longer possess the prestige and status associated with them.”); Coach, Inc. v. Treasure Box, Inc., 2013 WL 2402922, at *8 (N.D. Ind. May 31, 2013).
 See e.g., Glynn S. Lunney, Trademark Monopolies, 48 Emory L.J. 367, 408 (claiming that the rationale for trademark protection against knockoffs, “while it may couch itself in terms of confusion and reputation, seems to rest on the sense that the ordinary rule of competition should not apply to prestige goods. . . . Thus, where competition resulting in lower priced goods is generally thought desirable, courts often complain about the lower prices that imitations of prestige goods generate.”).
 There are, of course, cases that do produce procompetitive benefits. Take, for example, a twist on the classic case of point-of-sale confusion. A plaintiff sells a simple product, such as bars of soap, and the defendant sells the identical product under the same mark at a lower price. Here, one could argue that only the plaintiff suffers harm in the form of lost sales, but that consumers, despite being confused as to the source of the soap, actually benefit from the ability to buy the same product at a cheaper price. However, such cases are outweighed by the great social costs of false negatives. See infra Part IV.B.
 As the district court explained in Am. Home Prods. Corp. v. Barr Labs., Inc., 656 F. Supp. 1058, 1068 (D.N.J. 1987), aff’d, 834 F.2d 368 (3d Cir. 1987):
“The resemblance between two products can alert consumers to the functional or utilitarian equivalence between them, to the fact that one product may be substituted for the other in the ultimate uses for which the products are intended. The free flow of information regarding the substitutability of products is valuable to individual consumers and to society collectively, and by providing it a supplier engages in fair competition based on those aspects—for example, price—in which the products differ.”
 See, e.g., Gibson Guitar Corp. v. Paul Reed Smith Guitars, LP, 423 F.3d 539 (6th Cir. 2005) (rejecting the plaintiff’s claim that the defendant’s guitar impermissibly created initial-interest confusion, explaining that “many legitimately competing product shapes are likely to create some initial interest in the competing product due to the competing product’s resemblance to the better-known product when viewed from afar”) (emphasis added).
 See, e.g., Horphag Research, Ltd. v. Pelligrini, 337 F.3d 1036, 1039–42 (9th Cir. 2003) (holding that the metatags used by the defendant were infringing because they caused initial-interest confusion).
 The potential procompetitive benefits associated with a competitor advertising its products and services extend further. In recent years, plaintiffs have successfully utilized the doctrine in “knock-off cases.” For example, the district court in Cartier, Inc. v. Four Star Jewelry Creations, Inc., No. 01 Civ. 11295, 2003 WL 21056809, at *1 (S.D.N.Y. May 8, 2003), held that watches designed to look similar to Cartier watches could be found to infringe Cartier’s trade dress if consumers were initially “attracted” to the watches, even if consumers knew that the knock-offs were not Cartier watches at the time of purchase.
 See, e.g., Karl Storz Endoscopy-Am., Inc. v. Surgical Techs., Inc., 285 F.3d 848, 854 (9th Cir. 2002) (“The law in the Ninth Circuit is clear that ‘post-purchase confusion,’ i.e., confusion on the part of someone other than the purchaser who, for example, simply sees the item after it has been purchased can establish the required likelihood of confusion under the Lanham Act.”); see also Adidas Am., Inc. v. Payless Shoesource, Inc., 546 F. Supp. 2d 1029, 1058 (D. Or. 2008) (finding that the plaintiff’s failure to allege point-of-sale confusion “is of no consequence” to the viability of its initial-interest and post-sale claims).
 Michael Grynberg, Trademark Litigation as Consumer Conflict, 83 N.Y.U. L. Rev. 60, 102 (2008) (criticizing post-sale confusion for “subordinat[ing] the interests of these consumers to those wishing to cultivate the status that comes with the purchase of artificially scarce goods”).
 See id. at 107.
 See Thomas F. Cotter, The Procompetitive Interest in Intellectual Property Law, 48 Wm. & Mary L. Rev. 483, 490–91 & n.17 (2006) (discussing the use of error-costs analysis in antitrust and intellectual property law).
 See Steven C. Salop, Exclusionary Conduct, Effect on Consumers, and the Flawed Profit-Sacrifice Test, 73 Antitrust L. J. 311, 311–12 (2006); Brunswick Corp. v. Pueblo Bowl-O-Mat, Inc., 429 U.S. 477, 488 (1977); (stating that “the antitrust laws . . . were enacted for ‘the protection of competition not competitors’”) (quoting Brown Shoe Co. v. United States, 370 U.S. 294, 320 (1962)); 15 U.S.C. § 14 (1914) (stating that practices are unlawful when they “may . . . substantially lessen competition or tend to create a monopoly. . . .”).
 See Dogan & Lemley, Search-Costs, supra note 7, at 1224–27. The primacy of competition makes trademark law distinct from the rest of intellectual property law, which protects authors and inventors from competition in order to incentivize investments in invention and creation. See Mark A. Lemley, The Economics of Improvement in Intellectual Property Law, 75 Tex. L. Rev. 989 (1997); Dogan & Lemley, Merchandising Right, supra note 7, at 467–68; 17 U.S.C. § 106 (2000) (enumerating the exclusive rights of copyright holders); 35 U.S.C. § 271 (2000) (setting forth the exclusive rights of patent holders). Copyright and patent law grant authors and inventors exclusive economic rights to remedy the perceived market failure that would arise if copiers were able to replicate expressive works and inventions without incurring the costs of creating them. Coverdale, supra note 10, at 869 (“Unlike the patent and copyright laws, however, trademark protection is intended to promote, not hinder, competition.”); see generally Wendy J. Gordon, Asymmetric Market Failure and Prisoner’s Dilemma in Intellectual Property, 17 U. Dayton L. Rev. 853 (1992).
 See generally Mark A. Lemley, Property, Intellectual Property, and Free Riding, 83 Tex. L. REV. 1031 (2005).
 See Brown Shoe, 370 U.S. at 320 (stating that the antitrust laws were enacted for “the protection of competition, not competitors”).
 Robert G. Bone, Taking the Confusion Out of “Likelihood of Confusion”: Toward a More Sensible Approach to Trademark Infringement, 106 Nw. U. L. Rev. 1307, 1316–20 (2012).
 See Thomas A. Piraino, Jr., Reconciling the Per Se and Rule of Reason Approaches to Antitrust Analysis, 64 S. Cal. L. Rev. 685, 685–93 (1991); Richard M. Steuer, Indiana Federation of Dentists: The Per Se-Rule of Reason Continuum (and a Comment on State Action), 8 Cardozo L. Rev. 1101, 1120 (1987) (“[T]he Supreme Court labored at defining two categories of antitrust offenses—those that were illegal per se and those that violated the ‘rule of reason.’”).
 See Thomas Krattenmaker, Per Se Violations in Antitrust Law: Confusing Offenses with Defenses, 77 Geo. L.J. (“[T]he configuration of ‘per se violation’ seems to mean . . . that once certain conduct by a defendant is proved—e.g., horizontal price fixing, a group boycott, a tie-in sale—the plaintiff has established without doubt a violation of the antitrust laws. Nothing remains to be said, or can be said, by either side on the question of whether the defendant violated the antitrust laws.”).
 See, e.g., Newman v. Universal Pictures, 813 F.2d 1519, 1522–23 (9th Cir. 1987) (explaining that the per se rule “relieves plaintiff of the burden of demonstrating an anticompetitive effect, which is assumed”); see also United States v. Topco Assocs., Inc., 405 U.S. 596, 607 (1972) (“[T]here are certain agreements or practices which because of their pernicious effect on competition and lack of any redeeming virtue are conclusively presumed to be unreasonable and therefore illegal without elaborate inquiry as to the precise harm they have caused or the business excuse for their use.”) (quoting N. Pac. Ry. Co. v. United States, 356 U.S. 1, 5 (1958)).
 The rule of reason is not a rule at all; rather, it is a standard. Generally speaking, standards set forth a general decree, leaving interpretation for later adjudication. By contrast, rules specify ex-ante which types of conduct are forbidden, leaving only factual determinations remaining ex-post. For instance, when confronting the issue of speeding, a legislature could impose a rule making it illegal to drive above “fifty-five miles per hour,” or alternatively could set a standard making it illegal to drive at “unreasonable speeds.” The debate about which regime is better is a topic that has been heavily theorized before, both in the realm of intellectual property and elsewhere. See Matthew D. Adler & Eric A. Posner, Rethinking Cost-Benefit Analysis, 109 Yale L.J. 165 (1999).
 See Cont’l Television v. GTE Sylvania, 433 U.S. 36, 49 (1977) (noting that in rule of reason cases “the factfinder weighs all of the circumstances of a case in deciding whether a restrictive practice should be prohibited as imposing an unreasonable restraint on competition”). Generally, harm to competition is demonstrated using economic models and data.
 See, e.g., Mark A. Lemley & Christopher R. Leslie, Categorical Analysis in Antitrust Jurisprudence, 93 Iowa L. Rev. 1207, 1214–15 (2008).
 In California Dental, for example, because the majority concluded that the FTC failed to make out a prima facie case, the burden never shifted. See Cal. Dental Ass’n v. FTC, 526 U.S. 756, 775 n.12 (1999). The dissenters in that case disagreed with this conclusion. Id. at 783.
 See Nat’l Collegiate Athletic Ass’n v. Bd. of Regents of Univ. of Okla., 468 U.S. 85, 113 (1984); California Dental, 526 U.S. at 788 (Breyer, J., concurring in part and dissenting in part) (“In the usual Sherman Act § 1 case, the defendant bears the burden of establishing a procompetitive justification.”).
 See, e.g., Capital Imaging Assocs., P.C. v. Mohawk Valley Med. Assocs., Inc., 996 F.2d 537, 543 (2d Cir. 1993) (“Assuming [the] defendant comes forward with such proof, the burden shifts back to [the] plaintiff . . . to demonstrate that any legitimate collaborative objectives proffered by [the] defendant could have been achieved by less restrictive alternatives, that is, those that would be less prejudicial to competition as a whole.”); see also C. Scott Hemphill, Less Restrictive Alternatives in Antitrust Law, 116 Colum. L. Rev. 927, 941 (2016). Of course, if the defendant is unable to offer a procompetitive justification, then the plaintiff should prevail.
 See United States v. Microsoft Corp., 253 F.3d 34, 59 (D.C. Cir. 2001) (“[C]ourts routinely apply a . . . balancing approach” requiring plaintiff to “demonstrate that the anticompetitive harm . . . outweighs the procompetitive benefit.”); Am. Ad Mgmt., Inc. v. GTE Corp., 92 F.3d 781, 789 (9th Cir. 1996) (suggesting that the rule of reason requires a showing that “the restraint is unreasonable as determined by balancing the restraint and any justifications or pro-competitive effects of the restraint”).
 See Leegin Creative Leather Prods., Inc. v. PSKS, Inc., 127 S. Ct. 2705, 2713 (2007) (“To justify a per se prohibition a restraint must have ‘manifestly anticompetitive’ effects and ‘lack . . . any redeeming virtue.’”) (citations omitted); Herbert J. Hovenkamp, The Rule of Reason, 70 Fla. L. Rev. 81, 96 (2018) (“Per se illegality is appropriate if judicial experience indicates that a particular class of restraints rarely has any effect but to reduce output and increase price.”).
 See Robert H. Bork, The Rule of Reason and the Per Se Concept: Price Fixing and Market Division, 75 Yale L.J. 373, 391 (1966) (stating that “all horizontal price-fixing and market division is illegal per se”); Verizon Commc’ns Inc. v. Law Offices of Curtis V. Trinko, L.L.P., 540 U.S. 398, 408 (2004) (identifying collusion for purposes such as price fixing as the “supreme evil of antitrust”).
 N. Pac. Ry. Co. v. United States, 356 U.S. 1, 5 (1958) (identifying price fixing, division of markets, group boycotts, and tying arrangements as unlawful activities “in and of themselves”).
 Notably, most conduct analyzed as antitrust violations are considered under a rule of reason standard rather than a per se rule. See Cont’l T.V., Inc. v. GTE Sylvania Inc., 433 U.S. 36, 49 (1977) (noting that the rule of reason is “applied for the majority of anticompetitive practices challenged under § 1 of the Act”); see also State Oil Co. v. Khan, 522 U.S. 3, 10 (1997).
 Sylvania, 433 U.S. at 49 (noting that in rule of reason cases “the factfinder weighs all of the circumstances of a case in deciding whether a restrictive practice should be prohibited as imposing an unreasonable restraint on competition”).
 See John M. Newman, Procompetitive Justifications in Antitrust Law, 94 Ind. L.J. 501, 530(2019) (“The modern antitrust enterprise is concerned with the social costs of erroneous decisions.”).
 See Andrew I. Gavil, Antitrust Bookends: The 2006 Supreme Court Term in Historical Context, 22 Antitrust 21, 21 (2007); C. Scott Hemphill, An Aggregate Approach to Antitrust: Using New Data and Rulemaking To Preserve Drug Competition, 109 Colum. L. Rev. 629, 669 (2009).
 See Richard A. Posner, An Economic Approach to Legal Procedure and Judicial Administration, 2 J. Legal Stud. 399 (1973); Richard A. Posner, Economic Analysis Of Law 16–17 (5th ed. 1998). The expected value of each error cost can be illustrated as a mathematical formula, E.V. = P(x)*n. Here, P(x) is the probability of the event (either a false positive or false negative); n is the social cost produced by that type of error; by multiplying the P(x) with n, we get E.V. or the expected value of the type of error.
 See Scott E. Sundby, The Reasonable Doubt Rule and the Meaning of Innocence, 40 Hastings L.J. 457, 460 (1989).
 See Alan Devlin & Michael Jacobs, Antitrust Error, 52 Wm. & Mary L. Rev. 75, 83–84 (2010) (comparing standards used in antitrust to criminal law). In the antitrust context, the condemnation of procompetitive behavior is deemed a “false positive,” while allowing anticompetitive behavior is called a “false negative.”
 See, e.g., Richard A. Posner, Antitrust Policy and the Supreme Court: An Analysis of the Restricted Distribution, Horizontal Merger and Potential Competition Decisions, 75 Colum. L. Rev. 282, 283–99 (1975) (discussion of restricted-distribution cases); Howard A. Shelanski, The Case for Rebalancing Antitrust and Regulation, 109 Mich. L. Rev. 683, 712 (2011) (“[A]ntitrust jurisprudence has evolved to reduce significantly the likelihood of false positives.”).
 See Leegin Creative Leather Prods., Inc. v. PSKS, Inc., 127 S. Ct. 2705 (2007).
 Id. at 2718.
 Id.; see also Verizon Commc’ns Inc. v. Law Offices of Curtis V. Trinko, LLP, 540 U.S. 398, 414 (2004) (“Mistaken inferences and the resulting false condemnations ‘are especially costly, because they chill the very conduct the antitrust laws are designed to protect.’”) (quoting Matsushita Elec. Indus. Co. v. Zenith Radio Corp., 475 U.S. 574, 594 (1986)).
 See Leegin, 127 S. Ct. at 2718.
 See Fred S. McChesney, Talking ’Bout My Antitrust Generation: Competition for and in the Field of Competition Law, 52 Emory L.J. 1401, 1413 (2003).
 See Frank H. Easterbrook, The Limits of Antitrust, 63 Tex. L. Rev. 1, 2–3 (1984).
 See Joshua D. Wright, Abandoning Antitrust’s Chicago Obsession: The Case for Evidence-Based Antitrust, 78 Antitrust L.J. 241, 248 (“[T]he costs of false convictions in the antitrust context are likely to be significantly larger than the costs of false acquittals.”).
 For example, the Supreme Court’s decision in Dr. Miles Med. Co. v. John D. Park & Sons Co., 220 U.S. 373 (1911), holding that a RPM is illegal per se, took nearly a century to be overturned by the Court in Leegin. The durability of erroneous judicial precedent was a principle concern for Judge Easterbrook. See Easterbrook, supra note 103, at 2. (“If the court errs by condemning a beneficial practice, the benefits may be lost for good. Any other firm that uses the condemned practice faces sanctions in the name of stare decisis, no matter the benefits.”); id. at 15 (“There is no automatic way to expunge mistaken decisions of the Supreme Court.”); Wright, supra note 104, at 248 (“[J]udicial errors that wrongly excuse an anticompetitive practice may eventually be undone by competitive forces attracted by the presence of monopoly rents.”).
 William McGeveran & Mark P. McKenna, Confusion Isn’t Everything, 89 Notre Dame L. Rev. 253, 304 (2013) (“Instead of protecting legitimate competition interests notwithstanding the accompanying risk of some consumer confusion, courts strive to prevent confusion even if doing so harms competition.”).
 See supra Part III.B. (detailing how conduct found liable under initial-interest confusion can produce procompetitive effects by reducing consumer search costs and providing easy access to comparative quality and price information); id. (detailing how, in a post-sale confusion world, the actual purchasers of the defendant’s product benefit from a competitive market for the merchandised goods and therefore benefit from the defendant’s conduct).
 See Grynberg, supra note 49, at 99 (“[P]ermitting initial interest confusion may also harm consumers. The class of initially confused consumers includes those who are specifically seeking a particular brand to the exclusion of others. They must expend extra effort to determine which product is which, and to find their preferred choice. For these consumers, initial interest confusion impedes the trademark’s function of reducing consumer search costs. This perspective suggests that a balancing is possible: Courts should police initial interest confusion only when it produces greater harms than benefits.” (emphasis in original)).