Stand in a drugstore aisle and you are surrounded by a quiet act of faith. You reach for a familiar blue box because its color, its logo, and its name promise that the thing inside will be what it was last time. You do not run a chemical assay; you trust the brand. Trademark law exists to protect that small, repeated act of trust at industrial scale. A trademark is a word, name, symbol, design, or combination of these that identifies the source of goods or services and distinguishes them from those of others, and the whole system depends on one premise: the mark reliably tells you who stands behind the product. When a second comer's mark makes consumers believe they are buying from, or dealing with, the first comer, the law calls it a likelihood of confusion -- and confusion is the engine that drives almost every trademark infringement case in the United States.

This article is a deep, practical guide to that engine. It is written so that a judge weighing a preliminary injunction motion, a trademark lawyer drafting an opposition, and a founder choosing a name for a new product can all follow it. We define every term of art the first time it appears, we work through concrete (and deliberately invented) examples labeled with names like "Acme" and "Brightline," and we cite the primary authorities -- statutes, Supreme Court and circuit decisions, and Trademark Trial and Appeal Board (TTAB) precedent -- so that the assertions here can be checked rather than merely trusted.

We will cover the statutory framework; the three great multifactor tests (the Second Circuit's Polaroid factors, the Federal Circuit and TTAB's DuPont factors, and the Ninth Circuit's Sleekcraft factors); each individual factor in depth; the recognized variants of confusion -- direct, reverse, sponsorship, initial-interest, and post-sale; the slippery role of consumer sophistication; how confusion surveys are built and torn apart; who else can be on the hook through secondary liability; the separate doctrine of dilution under the Trademark Dilution Revision Act; the special pressures the internet and keyword advertising place on the analysis; the Supreme Court's 2023 reshaping of expressive-use defenses in Jack Daniel's Properties, Inc. v. VIP Products LLC; the remedies and defenses that determine what a confusion finding is actually worth; and a concrete brand-protection program. The throughline is that confusion analysis is not a checklist to be tallied mechanically. It is a holistic, fact-intensive judgment about the probable reactions of real consumers in a real marketplace.

For readers who want the foundational concepts first -- what a trademark is, how rights arise, and how registration works -- start with our trademark basics primer and the companion trademark overview on substantive standards for protection. This piece assumes those fundamentals and goes deep on the confusion question itself.

The Statutory Foundation: What a Plaintiff Must Prove

Federal trademark rights live primarily in the Lanham Act of 1946. The Act creates two principal causes of action for infringement. Section 32(1), codified at 15 U.S.C. § 1114(1), protects marks registered on the Principal Register of the U.S. Patent and Trademark Office (USPTO). Section 43(a), codified at 15 U.S.C. § 1125(a), protects unregistered common-law trademarks and trade dress, and also reaches false designations of origin and false or misleading representations of fact. (For the unfair-competition and false-advertising side of Section 43(a), see false advertising and Lanham Act Section 43(a).) Because owners of registered marks may sue under Section 43(a) as well, most infringement complaints plead both sections, and the two use overlapping likelihood-of-confusion standards -- so the same multifactor analysis governs registered and unregistered marks alike. Section 32 reaches the unauthorized use of any "reproduction, counterfeit, copy, or colorable imitation" of a registered mark, and a "colorable imitation" is statutorily defined as any mark so resembling a registered mark "as to be likely to cause confusion or mistake, or to deceive." 15 U.S.C. § 1127.

To prevail under either section, a plaintiff must prove three things. First, that it owns a valid, legally protectable mark -- a mark that is distinctive (either inherently or through acquired secondary meaning) and not generic or functional. (Common-law rights and their territorial limits are explored in trademark rights under common law.) Second, that the defendant used the mark in commerce without authorization after the plaintiff's rights arose. And third -- the issue at the heart of this article -- that the defendant's use is likely to cause consumer confusion. The plaintiff carries the burden of proving each element by a preponderance of the evidence. See KP Permanent Make-Up, Inc. v. Lasting Impression I, Inc., 543 U.S. 111, 118-20 (2004) (confirming that the plaintiff bears the burden of proving likelihood of confusion, even when the defendant asserts fair use).

A foundational and frequently misunderstood point: the standard is likelihood, not mere possibility. Courts uniformly hold that a plaintiff must show a probability of confusion, not just a theoretical or speculative one. See, e.g., Streamline Production Systems, Inc. v. Streamline Manufacturing, Inc., 851 F.3d 440, 453 (5th Cir. 2017); Nora Beverages, Inc. v. Perrier Group of America, Inc., 269 F.3d 114, 121 (2d Cir. 2001); Rodeo Collection, Ltd. v. West Seventh, 812 F.2d 1215, 1217 (9th Cir. 1987). The classic formulation asks whether an appreciable number of ordinarily prudent purchasers are likely to be confused as to the source, sponsorship, or affiliation of the goods or services. See Savin Corp. v. Savin Group, 391 F.3d 439, 456 (2d Cir. 2004). The reasonable, ordinarily prudent consumer -- not the most careless shopper and not the most fastidious expert -- is the touchstone.

Confusion is not limited to purchasers, however. Although courts often phrase the inquiry in terms of "actual or potential purchasers," the Lanham Act sweeps more broadly. Courts have held that confusion among non-purchasers can be actionable, including confusion among the general public in the post-sale context (General Motors Corp. v. Keystone Automotive Industries, Inc., 453 F.3d 351, 356 (6th Cir. 2006)); among brokers, dealers, and suppliers (Arrowpoint Capital Corp. v. Arrowpoint Asset Management, LLC, 793 F.3d 313, 320-21 (3d Cir. 2015)); and among any person whose confusion can harm the trademark owner's reputation, such as when misdirected claim forms and correspondence reach the wrong company (Beacon Mutual Insurance Co. v. OneBeacon Insurance Group, 376 F.3d 8 (1st Cir. 2004)). What is not actionable is confusion among people with no knowledge of or interest in the parties' goods or services. See Estee Lauder Inc. v. The Gap, Inc., 108 F.3d 1503, 1511 (2d Cir. 1997).

The Three Faces of Confusion: Source, Sponsorship, and the Direction of the Mistake

Before turning to the factors, it helps to fix the kinds of mistaken belief the law guards against. Courts evaluate confusion from several angles at once.

Direct (forward) confusion is the paradigm. It occurs when consumers encounter the junior user's goods and mistakenly believe they come from, or are made by, the senior user. Imagine a newcomer markets a cola under a script logo nearly identical to a famous incumbent's; shoppers reach for the newcomer thinking it is the incumbent's product. The newcomer free-rides on goodwill the incumbent built.

Sponsorship, affiliation, or connection confusion is broader. Here consumers may correctly understand that the goods come from a different company, yet wrongly believe the senior user authorized, licensed, endorsed, or is otherwise associated with the junior user. This is the theory behind many merchandising, co-branding, and celebrity-endorsement disputes. The statute expressly reaches confusion "as to the affiliation, connection, or association" of the parties or "as to the origin, sponsorship, or approval" of the goods. 15 U.S.C. § 1125(a)(1)(A).

Reverse confusion flips the direction of the mistake and is discussed in detail below. Instead of the junior user passing itself off as the senior user, a large junior user saturates the market so heavily that consumers come to believe the smaller, senior user is the infringer, or is affiliated with the junior user. The harm is the obliteration of the senior user's identity rather than misappropriation of its goodwill.

These categories are not mutually exclusive; a single set of facts can present forward, reverse, and sponsorship theories simultaneously. The multifactor tests below are the analytical machinery courts use to evaluate all of them.

The Multifactor Tests: One Question, Twelve-Plus Answers

Every regional circuit and the Federal Circuit has developed its own enumerated list of likelihood-of-confusion factors. The lists differ in number and labeling, but they probe the same underlying realities: how strong is the senior mark, how similar are the marks and the goods, and what does the real-world evidence (actual confusion, intent, surveys) show? No circuit treats its list as a scorecard to be totaled. The factors are, in the courts' recurring phrase, a guide and not a "rote checklist," and they must be weighed holistically with attention to the particular market. Three lists dominate the national conversation and deserve to be set out in full.

The Second Circuit: The Polaroid Factors

The Second Circuit's test comes from Judge Henry Friendly's opinion in Polaroid Corp. v. Polarad Electronics Corp., 287 F.2d 492, 495 (2d Cir. 1961). The Polaroid factors are:

  1. The strength of the plaintiff's mark.
  2. The similarity of the two marks.
  3. The proximity (competitive relatedness) of the products or services.
  4. The likelihood that the senior user will "bridge the gap" -- that is, expand into the junior user's market.
  5. Actual confusion.
  6. The junior user's good faith (or bad faith) in adopting its mark.
  7. The quality of the junior user's product.
  8. The sophistication of the relevant consumers.

Polaroid remains the most influential list in the country because of Friendly's stature and because the Second Circuit decides a large share of fashion, media, and financial-services trademark cases.

The Federal Circuit and the TTAB: The DuPont Factors

For appeals from the TTAB -- which adjudicates oppositions to registration and petitions to cancel registrations -- and in registrability disputes generally, the governing list is the DuPont factors, from In re E.I. du Pont de Nemours & Co., 476 F.2d 1357, 1361 (C.C.P.A. 1973). The thirteen DuPont factors, considered to the extent the record supports them, are:

  1. The similarity of the marks in appearance, sound, connotation, and commercial impression.
  2. The similarity and nature of the goods or services as described in the application or registration.
  3. The similarity of established, likely-to-continue trade channels.
  4. The conditions of purchase -- impulse versus careful, sophisticated purchasing.
  5. The fame of the prior mark (a factor that, before the TTAB, can be decisive).
  6. The number and nature of similar marks in use on similar goods (third-party use and registrations).
  7. The nature and extent of any actual confusion.
  8. The length of time of concurrent use without evidence of actual confusion.
  9. The variety of goods on which a mark is or is not used.
  10. The market interface between the parties (consent agreements, assignments, laches).
  11. The extent to which the applicant has a right to exclude others.
  12. The extent of potential confusion (de minimis or substantial).
  13. Any other established fact probative of the effect of use.

A crucial structural point distinguishes the DuPont analysis from infringement litigation. Because the TTAB decides only registrability -- not the right to use a mark or any claim for damages -- it compares the marks and goods as they appear in the application and registration, not as they are actually used in the marketplace. The TTAB must presume the goods travel in all normal channels to all usual classes of purchasers for goods of that type. See, e.g., In re i.am.symbolic, llc, 866 F.3d 1315 (Fed. Cir. 2017). A registration that broadly describes "clothing" is treated as covering all clothing in all channels, even if the applicant in fact sells only one item in one store. This makes the DuPont analysis more abstract -- and often more favorable to the party opposing registration -- than the marketplace-grounded inquiry a district court conducts. Practitioners moving between the two forums must keep this difference front of mind. For the mechanics of an opposition or cancellation, see our discussion of discovery practice in TTAB trademark proceedings; for the registration process the TTAB polices, see the trademark process.

The Ninth Circuit: The Sleekcraft Factors

The Ninth Circuit -- whose docket is rich in technology, entertainment, and internet cases -- applies the factors from AMF Inc. v. Sleekcraft Boats, 599 F.2d 341, 348-49 (9th Cir. 1979):

  1. The strength of the mark.
  2. The proximity or relatedness of the goods.
  3. The similarity of the marks.
  4. Evidence of actual confusion.
  5. The marketing channels used.
  6. The type of goods and the degree of care likely to be exercised by the purchaser.
  7. The defendant's intent in selecting the mark.
  8. The likelihood of expansion of the product lines.

The Ninth Circuit has repeatedly stressed that the Sleekcraft factors are "not a rote checklist" and that their relative weight shifts with the facts, particularly in internet cases. See Network Automation, Inc. v. Advanced Systems Concepts, Inc., 638 F.3d 1137, 1145-46 (9th Cir. 2011).

The remaining circuits use parallel lists -- the Third Circuit's Lapp factors, the Sixth Circuit's Frisch's factors, the Fifth Circuit's "digits of confusion," the Seventh Circuit's seven-factor test, and so on. Because the substance overlaps heavily, the rest of this article organizes the discussion around the factors themselves rather than the circuits, noting circuit-specific wrinkles where they matter. But the overlap is not perfect, and the differences can be outcome-determinative: a brand owner choosing where to file, and a lawyer writing a clearance opinion, must always identify the controlling circuit's exact formulation -- including which factors that circuit treats as most important and how it resolves close calls. For the mechanics of clearance, see how to conduct a comprehensive trademark clearance search.

The Factors in Depth

Strength of the Senior Mark

The strength of the plaintiff's mark is, in several circuits, the single most important factor, because stronger marks stand out more and are therefore both more deserving of protection and more likely to be the source consumers assume. See Variety Stores, Inc. v. Wal-Mart Stores, Inc., 888 F.3d 651, 661 (4th Cir. 2018); Frehling Enterprises, Inc. v. International Select Group, Inc., 192 F.3d 1330, 1335 (11th Cir. 1999). Strength has two independent dimensions, and a mark can be strong on one and weak on the other.

Inherent (conceptual) strength measures how distinctive the mark is in the abstract, along the classic Abercrombie spectrum drawn in Abercrombie & Fitch Co. v. Hunting World, Inc., 537 F.2d 4 (2d Cir. 1976). From strongest to weakest:

  • Coined or fanciful marks are invented words with no prior meaning -- KODAK, EXXON, PEPSI. They are the strongest because they exist only to identify source.
  • Arbitrary marks are ordinary words applied to unrelated goods -- APPLE for computers, GAP for clothing, AMAZON for an online store. Strong, because the word tells the consumer nothing about the product itself.
  • Suggestive marks require imagination to connect the mark to a feature of the goods -- COPPERTONE for suntan lotion, NETFLIX for streaming. Inherently distinctive and protectable, but generally weaker than arbitrary marks.
  • Descriptive marks directly convey a quality or characteristic -- HOLIDAY INN, BED & BATH, AMERICAN AIRLINES. They are protectable only if they have acquired secondary meaning, the consumer association of the term with a single source. See Star Industries, Inc. v. Bacardi & Co., 412 F.3d 373, 385 (2d Cir. 2005).
  • Generic terms name the product category itself -- "soap," "rice," "escalator" (once a brand, now generic). They can never be trademarks, no matter how much money is spent. See Xtreme Lashes, LLC v. Xtended Beauty, Inc., 576 F.3d 221, 227 (5th Cir. 2009).

A vital lesson of Abercrombie is that the same word can fall in different boxes depending on the goods: "safari" was generic for certain khaki apparel yet potentially protectable for other items. Distinctiveness is always assessed in relation to the specific goods, never in the abstract. (One modern wrinkle: a "generic.com" term may still function as a mark where consumers do not perceive it as generic. See USPTO v. Booking.com B.V., 591 U.S. 549 (2020), holding that BOOKING.COM was registrable because the relevant public did not understand it to signify online hotel-reservation services generally.)

Acquired (commercial) strength measures actual marketplace recognition. Even a conceptually weak descriptive mark can become commercially powerful through long use and heavy promotion, and even a coined mark can be commercially weak if it is brand-new and unknown. Courts assess commercial strength through the duration of continuous use, advertising expenditures, sales volume (ideally set against competitor sales for context), unsolicited media coverage, and direct consumer-recognition evidence such as surveys. See Virgin Enterprises Ltd. v. Nawab, 335 F.3d 141, 147-48 (2d Cir. 2003).

Two recurring battlegrounds deserve attention. First, third-party use of similar marks on related goods weakens a mark, because consumers accustomed to distinguishing many similar marks in a crowded field are less likely to be confused by one more. See Kibler v. Hall, 843 F.3d 1068, 1074 (6th Cir. 2016); Lang v. Retirement Living Publishing Co., 949 F.2d 576, 581 (2d Cir. 1991). District courts generally require proof that the third-party marks are actually in use and promoted, not merely registered. See Scarves by Vera, Inc. v. Todo Imports Ltd., 544 F.2d 1167, 1173-74 (2d Cir. 1976). The Federal Circuit and TTAB, however, treat even third-party registrations as probative of a mark's conceptual weakness. See Jack Wolfskin Ausrustung Fur Draussen GmbH v. New Millennium Sports, S.L.U., 797 F.3d 1363, 1373-74 (Fed. Cir. 2015). Second, incontestability: a registration that has reached incontestable status under 15 U.S.C. § 1115(b) is conclusive evidence of validity, but courts disagree whether it makes the mark strong. The better-reasoned view is that incontestability establishes distinctiveness without establishing the degree of distinctiveness; an incontestable mark can still be commercially weak. See Therma-Scan, Inc. v. Thermoscan, Inc., 295 F.3d 623, 632 (6th Cir. 2002).

Similarity of the Marks

Most courts rank similarity of the marks as one of the two most important factors. See A&H Sportswear, Inc. v. Victoria's Secret Stores, Inc., 237 F.3d 198, 216 (3d Cir. 2000); GoTo.com, Inc. v. Walt Disney Co., 202 F.3d 1199, 1205 (9th Cir. 2000). The comparison runs along three axes -- sight, sound, and meaning (commercial impression) -- and several rules of engagement govern how it is done:

  • Marks are compared as a whole, not dissected into components (the anti-dissection rule). See AutoZone, Inc. v. Tandy Corp., 373 F.3d 786, 795 (6th Cir. 2004).
  • Courts do not place the marks side by side, because consumers rarely encounter them simultaneously; the test is the recollection of an ordinary consumer who saw one mark at one time and the other later. See Louis Vuitton Malletier v. Dooney & Bourke, Inc., 454 F.3d 108, 117 (2d Cir. 2006).
  • The marks are compared in the marketplace context in which consumers confront them. Identical word marks may not confuse if packaging, trade dress, and conditions of sale differ markedly. See Arcona, Inc. v. Farmacy Beauty, LLC, 976 F.3d 1074, 1080-81 (9th Cir. 2020).
  • Courts may give greater weight to the dominant portion of a composite mark. See Xtreme Lashes, 576 F.3d at 228.
  • Adding generic or descriptive matter to another's mark usually does not avoid similarity; "ACME PREMIUM" is still confusingly similar to "ACME." See New York City Triathlon, LLC v. NYC Triathlon Club, Inc., 704 F. Supp. 2d 305, 317 (S.D.N.Y. 2010).

A worked example: a small mapmaker uses STREETWISE, and a competitor launches STREETSMART for the same kind of folding city maps. The words are not identical, but they share the "STREET" prefix, evoke the same idea of urban navigation savvy, and would be encountered by the same shoppers in the same racks. The overall commercial impression is close enough to support confusion -- the comparison turns on the impression left in memory, not a letter-by-letter audit. Now change one fact: suppose dozens of other publishers already sell STREET-prefixed map lines. The shared prefix loses force, the "STREET" element is shown to be conceptually weak, and the similarity factor shrinks. The factors, as we will keep seeing, talk to one another.

Relatedness (Proximity) of the Goods or Services

The closer the goods or services, the less similarity between the marks is required to confuse. See Kos Pharmaceuticals, Inc. v. Andrx Corp., 369 F.3d 700, 712-13 (3d Cir. 2004). Critically, the parties need not be direct competitors. Goods are "proximate" if consumers would reasonably expect them to come from the same source -- if they are complementary, sold to the same buyers, or used together.

The principle traces to Aunt Jemima Mills Co. v. Rigney & Co., 247 F. 407 (2d Cir. 1917), where the senior user sold pancake flour and the defendant used the same mark on pancake syrup. Even though the products were not identical, consumers would naturally assume the syrup came from the flour maker; this "related goods" doctrine (sometimes called the Aunt Jemima doctrine) extended trademark protection beyond strictly competing goods and remains good law. The modern question is always the consumer's expectation about source, not a wooden inquiry into whether the goods compete.

Likelihood of Bridging the Gap

A subset of the relatedness inquiry, this factor (prominent in the Second and Ninth Circuits) asks whether the senior user is likely to expand into the junior user's market. If a maker of premium ceramic insulators would naturally extend its line into ceramic conduit and pipe, a consumer encountering the junior user's pipe under a similar mark may assume the insulator maker has bridged the gap. Cf. Interpace Corp. v. Lapp, Inc., 721 F.2d 460 (3d Cir. 1983). The factor protects the senior user's natural zone of expansion and the consumer's reasonable assumption that established brands grow into adjacent fields.

Evidence of Actual Confusion

Actual confusion is not required to win -- likelihood is the standard -- but it is often the most persuasive evidence available, because it shows the predicted harm actually occurring. See World Carpets, Inc. v. Dick Littrell's New World Carpets, 438 F.2d 482, 489 (5th Cir. 1971) ("There can be no more positive or substantial proof of the likelihood of confusion than proof of actual confusion."). Evidence can take the form of misdirected phone calls, emails, orders, invoices, or complaints; testimony from confused consumers; and, most systematically, survey evidence (discussed below).

Two cautions temper this factor. First, de minimis instances of confusion -- a handful of stray misdirected emails against a backdrop of millions of transactions -- carry little weight. The volume of confusion must be appreciable relative to the opportunities for it. Second, the absence of actual confusion despite a long period of significant concurrent use can cut against the plaintiff, supporting an inference that confusion is unlikely. See Nora Beverages, 269 F.3d at 124. Courts give the most weight to confusion among actual purchasers in the ordinary course of trade and discount confusion among uninterested bystanders. A practical caution for litigants: anecdotal "actual confusion" evidence is also hearsay-prone. Testimony that "a customer told me he thought you two were the same company" can draw a sustainable objection unless it is offered for a non-hearsay purpose (such as the declarant's then-existing state of mind under Federal Rule of Evidence 803(3)) -- a reason surveys, which capture confusion methodically and admissibly, often do the heavy lifting.

Defendant's Intent

If the junior user adopted its mark intending to trade on the senior user's goodwill or to confuse consumers, courts often infer that confusion is likely -- reasoning that a deliberate copyist is in the best position to judge what will confuse, and presumably achieved its aim. See Mobil Oil Corp. v. Pegasus Petroleum Corp., 818 F.2d 254, 258-59 (2d Cir. 1987) (defendant's adoption of a flying-horse association to reach the same oil-industry customers supported an inference of intent). But the bar is real: mere knowledge of the senior mark, without more, is not bad faith. A junior user who knew of the senior mark but reasonably believed its own use was non-infringing -- for instance, after a clearance search and an attorney opinion -- has not acted in bad faith, and that good-faith record can be a powerful shield against enhanced remedies. We explore that dynamic in the shield of good faith. Conversely, where a party adopts a mark with knowledge of a senior user's territorial rights, that knowledge can destroy the good faith that some doctrines require -- as the Ninth Circuit explained applying the Tea Rose-Rectanus doctrine in Stone Creek v. Omnia.

Marketing Channels and Trade Channels

This factor compares how and where the parties advertise and sell. Confusion is more likely when the parties share retail outlets, advertising media, and target customers. But courts increasingly discount the factor when the shared channel is extremely broad. The fact that both parties sell on the internet or advertise nationally proves little, because nearly everyone does. See Network Automation, 638 F.3d at 1151 ("Today, it would be the rare commercial retailer that did not advertise online, and the shared use of a ubiquitous marketing channel does not shed much light on the likelihood of consumer confusion."). The question is whether the specific channels overlap in a way that brings the marks before the same consumers under conditions conducive to confusion.

Type of Goods and Degree of Consumer Care

The level of care the relevant consumer exercises can be outcome-determinative, and it interacts with every other factor. We treat it at length in its own section below.

Quality of the Junior User's Goods

This factor cuts in two directions. Markedly inferior goods support confusion under a reputational-harm and post-sale theory: consumers who encounter the shoddy junior product may blame the senior user, damaging its reputation. But comparable or high quality can also support confusion, because similar quality makes the assumption of common source more plausible. Because it cuts both ways, courts often treat quality as a minor factor.

Weighing the Factors: A Balancing Act, Not Arithmetic

The cardinal rule is that the factors are interrelated guideposts, not elements, and no single factor is dispositive. See Sleekcraft, 599 F.2d at 348-49. Courts may decline to march through every factor and instead focus on the few that matter on the facts; a powerful showing on mark similarity and goods relatedness can outweigh a thin record on the rest. In some circuits, doubt is resolved in favor of the senior user -- the newcomer, having the entire language to choose from, should stay well clear of an established mark. The art of confusion analysis lies in identifying which factors actually move the needle in the specific market and weighting them accordingly.

One more structural point matters for litigation strategy: is likelihood of confusion a question of law or of fact? The circuits split. The Second and several others treat it as a question of fact reviewed for clear error (with the application of the legal standard to the facts sometimes reviewed de novo), while the Federal Circuit treats it as a legal conclusion based on underlying factual findings, see Recot, Inc. v. Becton, 214 F.3d 1322, 1326 (Fed. Cir. 2000). That characterization drives two practical questions: how hard it is to win summary judgment (fact-intensive balancing often defeats it -- see, e.g., Levi Strauss & Co. v. Blue Bell, Inc., 778 F.2d 1352, 1356 n.5 (9th Cir. 1985)), and how much deference a verdict gets on appeal. Plan the case around the standard of review from day one.

Beyond the Point of Sale: The Variants of Confusion

Traditional analysis imagines confusion at the cash register. But courts have long recognized that the relevant confusion can occur before, after, or in a different direction from the sale.

Reverse Confusion

In ordinary (forward) confusion, the junior user feeds off the senior user's fame. Reverse confusion is the mirror image: a large, heavily advertising junior user floods the market so thoroughly that consumers come to associate the senior user's mark with the junior user -- and may even think the smaller senior user is the copycat. The doctrine protects small businesses from being swallowed by deep-pocketed latecomers. The seminal modern case is Big O Tire Dealers, Inc. v. Goodyear Tire & Rubber Co., 561 F.2d 1365 (10th Cir. 1977), where Goodyear's massive "Bigfoot" advertising campaign overwhelmed a small tire dealer's senior "Big Foot" mark.

Reverse confusion changes how several factors are applied. The strength inquiry focuses on the junior user's commercial strength (its advertising saturation), because that strength is what overwhelms the senior mark; and intent looks not at an intent to pass off but at whether the junior user carelessly or deliberately ignored the senior user's rights. See Dreamwerks Production Group, Inc. v. SKG Studio, 142 F.3d 1127 (9th Cir. 1998) (recognizing reverse confusion where a small "Dreamwerks" science-fiction convention organizer faced the new and far larger DreamWorks studio). Brand owners launching a major campaign must clear their marks against small senior users precisely because reverse confusion liability does not require any intent to deceive -- a counterintuitive but expensive lesson for big companies that assume their size insulates them.

Sponsorship and Affiliation Confusion

As noted, the statute reaches the mistaken belief that the senior user sponsors, endorses, licenses, or is affiliated with the junior user, even when consumers know the companies are different. This theory dominates merchandising disputes (think unlicensed team logos on apparel), endorsement and influencer cases, and disputes over whether a use falsely implies an official connection. The factors apply as usual, but the inquiry zeroes in on whether consumers would assume an authorized relationship.

Initial-Interest Confusion

Initial-interest confusion occurs when a junior user uses a senior mark to capture the consumer's initial attention, even though the confusion is dispelled before any purchase. The theory is that the junior user has unfairly used the senior user's goodwill as bait to divert customers. The doctrine rose to prominence in early internet cases. In Brookfield Communications, Inc. v. West Coast Entertainment Corp., 174 F.3d 1036 (9th Cir. 1999), the court analogized a competitor's use of a rival's mark in website metatags to a billboard that misdirects highway drivers to a competitor's exit; even motorists who realize the mistake have been diverted.

Initial-interest confusion is among the most criticized doctrines in trademark law. Critics argue it untethers liability from the consumer-protection rationale, especially online where users can click "back" in an instant and suffer no real deception. The Ninth Circuit itself has since narrowed Brookfield's reach in the keyword-advertising context, emphasizing that clearly labeled search results rarely confuse sophisticated, fast-clicking internet users. See Network Automation, 638 F.3d at 1148-54. Brand owners should treat initial-interest confusion as a viable but increasingly fact-sensitive theory, strongest where the use is genuinely deceptive (e.g., counterfeiters) and weakest where labeling and context make the source obvious.

Post-Sale Confusion

Post-sale confusion focuses on confusion among the observing public after the sale, not the buyer. The classic context is luxury knockoffs. A purchaser of a $40 "Rolex" from a street vendor knows it is fake -- there is no point-of-sale confusion. But bystanders who later see the watch on the buyer's wrist may believe it is genuine, which harms Rolex in two ways: it cheapens the prestige and exclusivity that drive genuine sales, and it can cause reputational damage if the knockoff malfunctions. See Rolex Watch U.S.A., Inc. v. Canner, 645 F. Supp. 484 (S.D. Fla. 1986); see also Hermès International v. Lederer de Paris Fifth Avenue, Inc., 219 F.3d 104, 108-09 (2d Cir. 2000); General Motors v. Keystone Automotive, 453 F.3d at 356. Post-sale confusion is also invoked in trade-dress and design cases where the look of a product, observed in use, signals source to the public -- a theme we develop in the intricate world of trade dress protection. Critics contend the doctrine protects status and exclusivity more than it protects against deception, but it is firmly established, particularly against counterfeiters.

Consumer Sophistication and the Degree of Care

The degree of care the relevant consumer exercises is a uniquely powerful and contested factor. The general rule is intuitive: the more expensive, important, or infrequent the purchase, the more care consumers take, and the less likely they are to be confused; cheap, impulse purchases invite less scrutiny and more confusion. Compare Versa Products Co. v. Bifold Co. (Manufacturing) Ltd., 50 F.3d 189 (3d Cir. 1995) (industrial valve actuators costing thousands of dollars are bought with great care), with Beer Nuts, Inc. v. Clover Club Foods Co., 805 F.2d 920, 926 (10th Cir. 1986) ("Buyers typically exercise little care in the selection of inexpensive items that may be purchased on impulse.").

Several variables shape the analysis: the price and importance of the purchase; its frequency (a once-a-decade appliance versus daily groceries); the sophistication of the target market (specialized industrial or professional buyers are presumed careful -- see Checkpoint Systems, Inc. v. Check Point Software Technologies, Inc., 269 F.3d 270 (3d Cir. 2001)); and the sales environment (a guided showroom purchase versus a hurried e-commerce add-to-cart). Importantly, even sophisticated buyers can be confused when the marks are very similar; sophistication reduces, but does not eliminate, the risk.

The doctrine attracts criticism for resting on class-based and inconsistent assumptions about consumers. Some opinions announce sweeping generalizations -- that "purchasers of relatively inexpensive goods are held to a lesser standard of purchasing care and do not give much thought to the purchase of such inexpensive goods," Munsingwear, Inc. v. Jockey International, Inc., 31 U.S.P.Q.2d 1146, aff'd, 39 F.3d 1184 (Fed. Cir. 1994) -- that may not track how real consumers behave. The most vivid illustration is wine, where courts have reached opposite conclusions about the very same product category. In Banfi Products Corp. v. Kendall-Jackson Winery, Ltd., 74 F. Supp. 2d 188, 199 (E.D.N.Y. 1999), the court credited evidence that "wine drinkers tend to be older, wealthier, and better educated than the average population" and therefore reasonably careful, while in E. & J. Gallo Winery v. Consorzio del Gallo Nero, 782 F. Supp. 457, 466 (N.D. Cal. 1991), the court treated the wine-buying public as generally unsophisticated impulse buyers -- a characterization that significantly enhanced the likelihood of confusion. These contradictory portraits of the same shoppers expose how malleable the factor is and how much it turns on which side builds the better record. The lesson for litigants is not to leave the question to judicial intuition. Build the record: commission a survey of how the actual consumers behave, offer expert testimony on purchasing patterns, and supply concrete evidence about the buyers, the price points, and the conditions of sale. A brand owner who shows that its market involves quick, low-cost, impulse decisions can substantially strengthen its confusion case; a defendant who shows careful, expert, high-dollar purchasing can substantially weaken it.

Survey Evidence: Building and Attacking Proof of Confusion

Because likelihood of confusion is a prediction about consumer perception, a well-designed consumer survey is often the most direct evidence available. Surveys are not required -- and counsel should always research the presiding judge's attitude toward them -- but some courts have drawn an adverse inference from a well-resourced party's failure to run one. Two survey designs dominate likelihood-of-confusion litigation.

The Eveready Format

The Eveready survey, named for Union Carbide Corp. v. Ever-Ready Inc., 531 F.2d 366 (7th Cir. 1976), is the gold standard when the senior mark is strong and well known. Respondents are shown only the junior user's product or mark -- the senior mark is not displayed -- and are asked open-ended questions: "Who do you think puts out this product?" "Do you think the company that makes this product makes any other products? If so, what?" "Do you think the company that makes this product needs permission from anyone to use this name?" Because the senior mark is never shown, an Eveready survey relies on the strength of the senior mark to elicit it from memory; a respondent who spontaneously names the senior brand, or assumes an affiliation, demonstrates confusion drawn from real-world recollection. Its great virtue is realism and resistance to leading the witness.

The Squirt Format

The Squirt survey, named for SquirtCo v. Seven-Up Co., 628 F.2d 1086 (8th Cir. 1980), is used when the senior mark is not strong enough for respondents to summon it unaided. Here both marks are shown -- sometimes sequentially, sometimes in an array -- and respondents are asked whether they believe the products come from the same source or affiliated sources. Squirt surveys are more useful for newer or niche marks, but they carry a higher risk of suggestiveness because placing the marks together can cue an association that would not arise in the marketplace. Courts scrutinize Squirt designs closely for that reason.

Surveys Beyond Confusion

The Eveready and Squirt formats answer the confusion question, but surveys do other jobs in trademark litigation, and counsel should match the instrument to the issue. Secondary-meaning surveys measure whether a descriptive mark has come to identify a single source in consumers' minds -- the prerequisite to protectability for descriptive marks. Dilution surveys test whether a junior use is blurring or tarnishing a famous mark's distinctiveness, supplying evidence for the "actual association" factor of the TDRA blurring analysis. Genericness surveys -- typically in the Teflon format, which asks respondents to sort terms as "brand names" or "common names" -- gauge whether the relevant public understands a term as a source identifier or merely as the name of the product category, the dispositive question in a genericide attack. Each shares the methodological discipline described below; only the question being asked changes.

Anatomy of a Defensible Survey

Whichever format is used, courts (and the Daubert and Federal Rule of Evidence 702 gatekeeping inquiry) evaluate the same methodological pillars:

  • Proper universe. The respondents must represent the relevant population -- the actual or prospective purchasers of the junior user's goods, not the general public. An over-inclusive universe is one of the most common grounds for excluding or discounting a survey.
  • Representative sample. The sample must be drawn so as to fairly represent that universe.
  • Marketplace conditions. The stimulus should replicate how consumers actually encounter the marks, including realistic context and packaging.
  • Neutral, non-leading questions. Questions must not suggest the desired answer or prompt guessing.
  • A proper control. A control cell (e.g., a non-infringing variant) isolates "noise" -- guessing and pre-existing beliefs -- from confusion actually caused by the challenged mark. The net confusion figure is the test-cell result minus the control-cell result.
  • A qualified, independent expert who designed and supervised the survey using accepted methods and who did not work backward from a desired result.

The migration of surveys online has added a further fault line. Modern panels and click-through formats are cheaper and faster, but they invite fresh attacks: were respondents actually screened into the proper universe, were "speeders" and inattentive panelists filtered out, and did the on-screen presentation realistically mimic the shopping experience? Methodological flaws of this kind usually go to weight rather than admissibility, but a sloppy online survey can do a party more harm than no survey at all. Net confusion levels around 15% or higher are frequently treated as supporting likelihood of confusion, though there is no magic number, and figures below that range can still matter in context. Survey evidence is expensive, which can disadvantage smaller parties, and it almost always provokes a battle of experts over methodology. But a clean, well-controlled survey remains one of the most persuasive tools a brand owner can deploy.

Who Else Is on the Hook: Secondary Liability

Direct infringers are not always the most attractive -- or most solvent -- defendants. A trademark owner facing a swarm of anonymous counterfeiters, or a marketplace teeming with infringing listings, often wants to reach the platform, supplier, or service provider that made the infringement possible. Two theories supply that reach, and both require proof of an underlying act of direct infringement.

Contributory infringement applies where a defendant either induces another to infringe or continues to supply its product or service to one it knows or has reason to know is infringing. The doctrine descends from Inwood Laboratories, Inc. v. Ives Laboratories, Inc., 456 U.S. 844, 854 (1982), a case about a drug manufacturer whose generic capsules were mimicking a branded competitor's. The modern digital battleground is the online marketplace. In Tiffany (NJ) Inc. v. eBay Inc., 600 F.3d 93 (2d Cir. 2010), the Second Circuit held that generalized knowledge that some counterfeits are sold on a platform is not enough; contributory liability requires that the service provider have specific knowledge of particular infringing listings and fail to act. That holding shapes how brand owners structure takedown programs and how platforms design their notice-and-removal systems.

Vicarious infringement is narrower and turns not on the defendant's involvement but on its relationship with the infringer -- an agency relationship, or joint control over the infringing goods or services. Separately, corporate officers and directors can incur personal liability where they actively participated in, directed, or authorized the infringing conduct; the corporate form is not an automatic shield. These theories matter enormously in counterfeiting and online-enforcement campaigns; we address the operational side in brand protection online.

Dilution: Protecting Famous Marks Without Confusion

Some marks are so famous that the law protects them even when there is no likelihood of confusion at all. This is the doctrine of dilution, available under Section 43(c) of the Lanham Act, 15 U.S.C. § 1125(c), as enacted by the Trademark Dilution Revision Act of 2006 (TDRA). Where infringement law protects consumers from confusion, dilution law protects the famous mark itself from erosion of its uniqueness and selling power, regardless of competition or confusion. See Rosetta Stone Ltd. v. Google, Inc., 676 F.3d 144, 167 (4th Cir. 2012).

The Fame Threshold

Dilution protection is reserved for a narrow elite. A mark qualifies only if it is widely recognized by the general consuming public of the United States as a designation of source. 15 U.S.C. § 1125(c)(2)(A). This is a demanding "household name" standard -- think COCA-COLA, NIKE, APPLE, MCDONALD'S. Crucially, the TDRA eliminated "niche fame": it is not enough that a mark is famous within a specialized industry or region. See Coach Services, Inc. v. Triumph Learning LLC, 668 F.3d 1356, 1373 (Fed. Cir. 2012). Many marks strong enough to win an infringement case fall far short of dilution fame. See Avery Dennison Corp. v. Sumpton, 189 F.3d 868, 877 (9th Cir. 1999) (AVERY and DENNISON were distinctive but not famous). Proving fame is its own evidentiary project -- sales, advertising reach, market share, third-party recognition, and survey data -- and the bar is high enough that many plaintiffs plead dilution defensively and never seriously pursue it.

The TDRA also fixed a key proof problem from the prior statute. In Moseley v. V Secret Catalogue, Inc., 537 U.S. 418 (2003), the Supreme Court read the earlier Federal Trademark Dilution Act to require proof of actual dilution -- an exceedingly hard burden. The TDRA legislatively overruled that result: a plaintiff now need only prove a likelihood of dilution. 15 U.S.C. § 1125(c)(1).

Blurring and Tarnishment

The TDRA recognizes two harms. Dilution by blurring is an "association arising from the similarity between a mark or trade name and a famous mark that impairs the distinctiveness of the famous mark." 15 U.S.C. § 1125(c)(2)(B). Blurring is the gradual whittling-away of a mark's singular identity through its use on unrelated goods -- the hypothetical worlds of "Kodak pianos," "Buick aspirin," or "Tiffany restaurants." To assess blurring, courts weigh six statutory factors: (i) the degree of similarity between the marks; (ii) the degree of inherent or acquired distinctiveness of the famous mark; (iii) the extent to which the owner is engaging in substantially exclusive use; (iv) the degree of recognition of the famous mark; (v) whether the junior user intended to create an association with the famous mark; and (vi) any actual association between the marks. 15 U.S.C. § 1125(c)(2)(B)(i)-(vi).

Dilution by tarnishment is an "association arising from the similarity between a mark or trade name and a famous mark that harms the reputation of the famous mark." 15 U.S.C. § 1125(c)(2)(C). Tarnishment typically arises when a junior use links the famous mark to unsavory, low-quality, or offensive contexts.

Statutory Exclusions

The TDRA contains important carve-outs that protect speech. There is no dilution liability for: (a) any fair use of a famous mark other than as a designation of source, including use in connection with parody, criticism, or commentary on the famous mark or its owner; (b) all forms of news reporting and news commentary; and (c) any noncommercial use of a mark. 15 U.S.C. § 1125(c)(3). These exclusions, especially the parody and noncommercial-use provisions, do a great deal of work -- but, as the next section shows, the Supreme Court has held they do not shield a humorous or parodic use when the defendant is using the famous mark as its own source identifier.

Expressive Use and the Narrowing of Rogers: Jack Daniel's v. VIP Products

For decades, defendants who used another's mark in an expressive work -- a movie, a song, a book, an artwork -- invoked the test from Rogers v. Grimaldi, 875 F.2d 994 (2d Cir. 1989). Under Rogers, the Lanham Act does not bar an expressive use of a mark unless the use either (1) has no artistic relevance to the underlying work, or (2) explicitly misleads as to source. Where Rogers applied, it functioned as a threshold filter that often ended a trademark case before any likelihood-of-confusion analysis, in service of the First Amendment.

The Supreme Court sharply cabined that doctrine in Jack Daniel's Properties, Inc. v. VIP Products LLC, 599 U.S. 140 (2023). VIP sold a squeaky dog toy called "Bad Spaniels," shaped like a Jack Daniel's whiskey bottle and parodying its label (replacing "Old No. 7 Tennessee Sour Mash Whiskey" with scatological humor). VIP argued the toy was an expressive parody entitled to the Rogers threshold protection. A unanimous Court disagreed, holding that Rogers does not apply when the alleged infringer uses the challenged mark as a designation of source for its own goods -- that is, as a trademark to identify and distinguish its own product. Because VIP used "Bad Spaniels" (and the bottle trade dress) as a source identifier for its own dog toys, the ordinary likelihood-of-confusion analysis governed, and the case returned for that analysis. The Court was careful to say it was not deciding whether Rogers is ever correct; it held only that Rogers has no role when the mark is used as a mark.

The Court also addressed dilution. It held that the TDRA's noncommercial-use exclusion does not categorically shield parody when the parody is used as a designation of source. The fair-use parody exclusion in 15 U.S.C. § 1125(c)(3)(A) expressly does not apply to use "as a designation of source for the person's own goods or services," and the Court declined to let the broader noncommercial-use exclusion swallow that limitation. Jack Daniel's, 599 U.S. at 160-63.

The practical upshot for brand owners and would-be parodists is significant. Humor and parody remain relevant -- they can make confusion less likely under the multifactor test, as the influential Haute Diggity Dog line of cases ("Chewy Vuiton" dog toys) illustrates (Louis Vuitton Malletier S.A. v. Haute Diggity Dog, LLC, 507 F.3d 252 (4th Cir. 2007)). But a defendant can no longer use parody or expression to avoid the confusion inquiry altogether when it has adopted the borrowed mark as its own brand. After Jack Daniel's, the central question in expressive-use cases is whether the defendant is using the plaintiff's mark to identify the source of the defendant's own goods; if so, the full Polaroid/Sleekcraft/DuPont analysis applies.

The Digital Marketplace: Keyword Advertising, Domains, and Social Media

The internet has not changed the legal standard, but it has stress-tested how the factors apply. Several digital contexts recur.

Keyword Advertising

When a competitor buys a brand owner's trademark as a search-advertising keyword, so that searching "Brightline" returns a sponsored ad for "Acme," is that infringement? The trend is that keyword purchasing is generally not actionable by itself. Two threshold questions loom: whether buying a keyword is even a "use in commerce" of the mark, and whether the resulting ads are clearly labeled and non-deceptive. The practical reality is that infringement suits against legitimate search engines are hard to win when sponsored links sit in a clearly labeled area separate from organic results, and where the advertiser's use is nominative -- using the mark to refer to a product, not as the advertiser's own brand. In Rosetta Stone Ltd. v. Google, Inc., 676 F.3d 144 (4th Cir. 2012), the Fourth Circuit reversed summary judgment for Google, holding that genuine fact issues existed on confusion and contributory liability where the keyword program was alleged to facilitate counterfeit sales -- but the opinion did not declare keyword advertising inherently infringing, and it pointedly did not decide the "use in commerce" or nominative-fair-use questions. An earlier and more plaintiff-friendly data point is Playboy Enterprises, Inc. v. Netscape Communications Corp., 354 F.3d 1020 (9th Cir. 2004), where the search engine "keyed" unlabeled adult-content banner ads to searches for "playboy" and "playmate." Relying on an initial-interest-confusion theory and survey evidence that a statistically significant share of users believed Playboy sponsored the ads, the Ninth Circuit reversed summary judgment for the defendant -- illustrating that keyword and ad-targeting practices become actionable precisely when the resulting ads are unlabeled and genuinely deceptive. The Ninth Circuit's later Network Automation decision reframed the Sleekcraft factors for this context, emphasizing the clarity of the ad's labeling, the appearance and content of the ad, and the sophistication of internet users, and discounting the shared "internet" marketing channel. The practical rule: keyword suits succeed most reliably when aimed at a competitor or counterfeiter whose ads actually confuse or who sells infringing goods -- not at a search engine displaying clearly labeled sponsored links. For a broader treatment of online enforcement, see brand protection online.

Domain Names and Cybersquatting

Bad-faith registration of domain names containing others' marks is governed by the Anticybersquatting Consumer Protection Act (ACPA), 15 U.S.C. § 1125(d), enacted in 1999, which supplies a list of bad-faith factors and remedies including transfer of the domain and statutory damages of up to $100,000 per domain. See Sporty's Farm L.L.C. v. Sportsman's Market, Inc., 202 F.3d 489 (2d Cir. 2000) (one of the first ACPA decisions). Brand owners also have a faster, cheaper administrative path through the Uniform Domain-Name Dispute-Resolution Policy (UDRP); we cover the mechanics in how to file a UDRP complaint for domain name disputes and the structure of the namespace in top-level domain names.

Social Media, Hashtags, and Influencers

Trademarks now appear as usernames, hashtags, and in influencer content, generating new sponsorship-confusion questions. Whether a hashtag use creates confusion is fact-bound; many hashtags are merely descriptive or functional and do not designate source. Cf. Eksouzian v. Albanese, 2015 WL 4640809 (C.D. Cal. Aug. 7, 2015) (treating certain hashtags as non-trademark uses in a settlement-interpretation dispute). Influencer campaigns also implicate FTC endorsement-disclosure rules, which intersect with -- but are distinct from -- trademark confusion. For the broader regulatory backdrop, see social media law basics.

Virtual Goods and the Metaverse

Branded virtual goods and NFTs raise the question whether a trademark for physical products extends to digital replicas. The high-profile Hermès International v. Rothschild "MetaBirkins" litigation produced a jury verdict for Hermès, applying (pre-Jack Daniel's) the Rogers framework and finding the digital handbags were not protected artistic expression. Hermès International v. Rothschild, 654 F. Supp. 3d 268 (S.D.N.Y. 2023). After Jack Daniel's, source-identifying uses of marks on virtual goods will be analyzed under ordinary confusion principles. We address these frontier issues in trademark challenges in the metaverse and virtual goods.

Artificial Intelligence in Brand Selection and Clearance

Artificial intelligence is reshaping trademark practice from two directions. On the input side, AI tools increasingly generate candidate brand names and screen them, and machine-learning systems are augmenting (not yet replacing) the traditional clearance search by surfacing phonetic, visual, and semantic near-matches across registers, domains, and social handles faster than manual review. On the analytical side, predictive models trained on past decisions promise to estimate likelihood-of-confusion outcomes, though such tools remain advisory: the multifactor analysis is a holistic, fact-bound judgment that resists mechanical scoring, and courts decide cases, not algorithms. A mark cleared or coined with AI assistance should still receive a human-reviewed clearance opinion -- both to catch what the model missed and to preserve the good-faith record that defeats willfulness.

A Comparative Glance: Confusion Beyond U.S. Borders

Because brands operate globally, U.S. brand owners increasingly confront confusion analyses that differ in structure and emphasis from the Lanham Act's. Two jurisdictions illustrate the range.

In the European Union, trademark law is harmonized across member states and administered for EU-wide rights through the European Union Intellectual Property Office (EUIPO). The foundational confusion standard comes from the European Court of Justice in SABEL BV v. Puma AG, Case C-251/95, [1997] E.C.R. I-6191, which directs courts to make a "global appreciation" of the visual, aural, and conceptual similarity of the marks based on the overall impression they create, bearing in mind their distinctive and dominant components. Crucially, SABEL held that a mere likelihood of association -- the later mark calling the earlier one to mind -- is not by itself enough to establish a likelihood of confusion. EU law also reaches further than U.S. infringement doctrine in one respect: marks with a reputation are protected against uses that take "unfair advantage" of, or are detrimental to, their distinctive character or repute, without any showing of confusion. EU practice likewise gives more weight to negotiated coexistence agreements between parties than U.S. courts typically do.

China, now one of the world's largest consumer markets, departs from the U.S. model in two structural ways that repeatedly trap foreign brand owners. First, China is a first-to-file jurisdiction: rights generally flow from registration, not use, which fuels trademark squatting and makes early filing essential. Second, China overlays the international classification with a unique sub-class system, so that goods nominally in the same class may be treated as dissimilar (or vice versa) depending on sub-class placement -- a trap that can leave a brand unprotected against confusingly similar marks in adjacent sub-classes. The difficulty of protecting personal names is illustrated by Michael Jordan v. Qiaodan Sports, in which China's Supreme People's Court in 2016 ((2016) ZGFXZ No. 27) held that the basketball star held name rights in the Chinese characters "乔丹" (Qiaodan) because the public strongly associated them with him, while declining protection for the Latin-alphabet pinyin "Qiaodan" -- a partial victory that became a touchstone for protecting personal names against Chinese trademark pirates. Foreign brand owners entering these markets should clear, file, and enforce according to local doctrine rather than assuming the U.S. framework travels.

What a Confusion Finding Is Worth: Remedies and Defenses

Proving likelihood of confusion answers the liability question. It does not, by itself, answer two further questions that often matter more to the client: what can the plaintiff get, and what can the defendant say back? A confusion analysis untethered from remedies and defenses is only half a strategy.

Injunctive Relief

The primary remedy in trademark cases is an injunction stopping the infringing use. 15 U.S.C. § 1116(a). Injunctions -- preliminary and permanent -- are the workhorses of trademark enforcement because trademark harm (lost control of reputation and goodwill) is notoriously hard to quantify in dollars. To obtain one, a plaintiff must satisfy the traditional four-factor equitable test of eBay Inc. v. MercExchange, L.L.C., 547 U.S. 388 (2006): likely success on the merits (or actual success, for a permanent injunction), irreparable harm, the balance of hardships, and the public interest. For years many courts presumed irreparable harm once likelihood of confusion was shown; eBay unsettled that presumption, and Congress responded in the Trademark Modernization Act of 2020 by codifying a rebuttable presumption of irreparable harm upon a finding (or, at the preliminary stage, a likelihood) of infringement. 15 U.S.C. § 1116(a). The presumption can be rebutted, but it restores much of the plaintiff's traditional advantage.

Monetary Relief

Money is harder to come by than an injunction, and clients should be told so early. The Lanham Act authorizes recovery of (1) the defendant's profits, (2) the plaintiff's actual damages, and (3) the costs of the action, all subject to the principles of equity. 15 U.S.C. § 1117(a). Crucially, monetary relief is not automatic -- "subject to the principles of equity" gives courts broad discretion, and in many cases courts grant only an injunction. See Lindy Pen Co. v. Bic Pen Corp., 982 F.2d 1400, 1404-05 (9th Cir. 1993). Any award must be compensatory or restitutionary, not punitive. 15 U.S.C. § 1117(a). And a registrant who failed to give statutory notice of registration (the ® symbol or equivalent) generally cannot recover profits or damages absent proof the defendant had actual notice. 15 U.S.C. § 1111.

On the much-litigated question of whether willfulness is required to disgorge a defendant's profits, the Supreme Court resolved a circuit split in Romag Fasteners, Inc. v. Fossil, Inc., 590 U.S. 212 (2020). The Court held that the Lanham Act does not impose a categorical precondition of willfulness for a profits award in a Section 43(a) case; the defendant's mental state remains "a highly important consideration," but it is not an inflexible gateway. After Romag, an innocent or merely negligent infringer can, in the right equitable circumstances, be ordered to disgorge profits -- a meaningful expansion of exposure.

In counterfeiting cases (a heightened species of Section 32 infringement), the plaintiff gets sharper tools: treble damages or profits and mandatory attorney's fees for use of a counterfeit mark absent extenuating circumstances, 15 U.S.C. § 1117(b), or, at the plaintiff's election, statutory damages of up to $2,000,000 per counterfeit mark per type of goods for willful counterfeiting, 15 U.S.C. § 1117(c). Cybersquatting under the ACPA similarly offers elective statutory damages. These regimes exist precisely because counterfeiters rarely keep honest books, making actual damages impossible to prove.

Attorney's fees are available to the prevailing party in "exceptional cases." 15 U.S.C. § 1117(a). Although that standard predates the patent analogue, courts widely apply the flexible, totality-of-the-circumstances definition the Supreme Court announced for the parallel Patent Act fee provision in Octane Fitness, LLC v. ICON Health & Fitness, Inc., 572 U.S. 545 (2014) -- an "exceptional" case being one that stands out from others in the substantive strength of a party's position or the unreasonable manner in which it was litigated. See, e.g., Fair Wind Sailing, Inc. v. Dempster, 764 F.3d 303 (3d Cir. 2014) (importing Octane Fitness into the Lanham Act).

Principal Defenses

A defendant's best defense is often to defeat the confusion case on its own terms -- attacking mark validity (genericness, mere descriptiveness without secondary meaning, functionality of claimed trade dress) or showing the factors do not add up. But the Lanham Act and the case law also supply affirmative defenses worth knowing:

Classic (descriptive) fair use protects a defendant who uses a term that happens to be (or resemble) another's mark for its ordinary descriptive meaning -- to describe its own goods, in good faith, and not as a mark. 15 U.S.C. § 1115(b)(4). The canonical examples are a cranberry juice described as "sweet-tart" despite the SWEETARTS mark, Sunmark, Inc. v. Ocean Spray Cranberries, Inc., 64 F.3d 1055 (7th Cir. 1995), and "King Size" used to describe large men's clothing. Critically, KP Permanent Make-Up holds that a defendant invoking classic fair use need not disprove all likelihood of confusion -- some consumer confusion can coexist with a valid fair-use defense. 543 U.S. at 121-23.

Nominative fair use protects a defendant who uses another's mark to refer to the trademark owner itself or its goods -- comparative advertising, repair and resale, commentary, compatibility statements. The Ninth Circuit's three-part test (New Kids on the Block v. News America Publishing, Inc., 971 F.2d 302 (9th Cir. 1992)) asks whether the product is readily identifiable without the mark, whether the defendant used only as much of the mark as necessary, and whether the defendant did anything to suggest sponsorship or endorsement. The closely related first-sale doctrine lets a reseller of genuine goods use the brand to identify what it is selling.

Laches and acquiescence can bar relief where the owner unreasonably delayed enforcement to the defendant's prejudice (laches) or affirmatively led the defendant to believe its use was tolerated (acquiescence). These equitable defenses, and their cousins waiver and estoppel, are explored in understanding equitable defenses -- a reminder that sleeping on your rights can be as costly as never having registered them. Finally, abandonment (including loss of rights through a "naked license" with no quality control) and genericide (the mark becoming the generic name of the product) can extinguish the mark entirely.

The takeaway: design the enforcement strategy around the remedy you actually want and the defenses you will actually face. A meritorious confusion claim aimed at the wrong defendant, with stale facts, no registration notice, and no plausible damages, can win on liability and still deliver little.

A Practical Brand-Protection Program

Doctrine matters most when it informs action before a dispute. Brand owners who treat trademark protection as a continuous program, not a one-time filing, dramatically reduce both the risk of being an infringer and the cost of stopping one. The core elements:

Clear before you commit. Before adopting any new mark, conduct a comprehensive clearance search -- USPTO records, state registrations, common-law uses, domains, and internet/social-media uses -- and obtain a written clearance opinion. This both avoids picking an infringing mark and, if a dispute later arises, supplies the good-faith record that defeats willfulness and enhanced damages. See how to conduct a comprehensive trademark clearance search and the shield of good faith. On timing, see also when to trademark your brand.

Register early and broadly. File federal applications for your core marks in every relevant class of goods and services, and consider defensive registrations of close variants and key product lines. Registration provides nationwide constructive priority, a presumption of validity, and access to enhanced remedies. For the filing mechanics and class selection, see how to file a trademark application with the USPTO, federal trademark application checklists, USPTO trademark application checklists, and USPTO trademark classes.

Build and preserve a strength record. Keep contemporaneous records of advertising spend, sales, market share, unsolicited media, and consumer recognition. These are the raw materials of the "strength" and "fame" factors, and they are far easier to assemble in real time than to reconstruct in litigation.

Monitor relentlessly. Implement watch services for new applications, marketplace listings, domains, app stores, and social handles. Early detection lets you oppose at the TTAB or send a measured cease-and-desist before the junior user builds equity -- and before reverse-confusion exposure (if you are the large party) or laches (if you are the senior party) sets in. When you do write that letter, calibrate it; see drafting a trademark cease and desist letter and, for the other side of the exchange, responding to a trademark cease and desist letter.

Enforce proportionately and consistently. Police your mark, but avoid overreaching "trademark bullying," which courts and the public penalize and which can backfire in fee awards and PR. Calibrate the response to the actual risk of confusion, document your enforcement to support the strength factor, and use coexistence or consent agreements where the markets genuinely do not overlap.

Govern licensing with quality control. When you license your mark, impose and actually exercise quality-control standards; a "naked license" without supervision can forfeit rights through abandonment. Set clear co-branding and usage guidelines and train employees and partners on proper trademark use.

Plan litigation evidence in advance. If a dispute is foreseeable, budget for a properly controlled Eveready or Squirt survey and identify the consumer-care and channel evidence you will need. Surveys built late and cheaply rarely survive Daubert.

Frequently Asked Questions

Do I have to prove people were actually confused to win a trademark case? No. The legal standard is likelihood of confusion -- a probability, not a certainty. Actual confusion is powerful evidence and can be decisive, but a plaintiff can prevail without it by showing that the multifactor test points to a probability of confusion. See KP Permanent Make-Up, 543 U.S. 111.

Are the Polaroid, DuPont, and Sleekcraft tests really different? They use different labels and number their factors differently, but they probe the same realities -- mark strength, mark similarity, goods relatedness, actual confusion, intent, channels, and consumer care. The most important practical difference is forum: the DuPont factors govern registrability before the TTAB and Federal Circuit and compare the marks and goods as described in the application/registration, while Polaroid and Sleekcraft govern infringement in district court and look to actual marketplace use.

My mark is strong but not famous. Can I sue for dilution? Probably not. Dilution under the TDRA requires that the mark be famous to the general U.S. consuming public -- a household-name standard that eliminates niche or regional fame. A mark can easily be strong enough to win an infringement case yet fall short of dilution fame. Your remedy in that situation is an ordinary likelihood-of-confusion claim.

Is buying a competitor's name as a Google keyword illegal? Generally not by itself. Courts have largely declined to find infringement merely from purchasing a keyword, focusing instead on whether the resulting ads are clearly labeled and non-deceptive and whether the buyer is selling infringing or counterfeit goods. The strongest keyword cases involve competitors or counterfeiters whose ads actually confuse consumers.

Does parody protect me if I put a famous brand on my product? Less than it used to. After Jack Daniel's v. VIP Products (2023), you cannot use parody to escape the likelihood-of-confusion analysis when you are using the famous mark as a source identifier for your own goods. Parody can still make confusion less likely within that analysis, and noncommercial expressive uses remain protected -- but using someone's mark as your brand triggers the full confusion inquiry.

If I win on confusion, will I get money damages? Not necessarily. The most reliable remedy is an injunction. Monetary relief -- the defendant's profits, your actual damages, costs, and (in exceptional cases) fees -- is subject to the court's equitable discretion, and many plaintiffs recover only an injunction. After Romag v. Fossil (2020) you no longer must prove willfulness to disgorge profits, but the defendant's intent remains highly important, and you may forfeit monetary recovery entirely if you failed to give statutory notice of your registration.

Can I go after the online marketplace instead of the anonymous seller? Sometimes, through contributory infringement -- but under Tiffany v. eBay, a platform is liable only if it had specific knowledge of particular infringing listings and failed to act, not merely general awareness that counterfeits exist on its site. Effective platform enforcement therefore depends on precise, well-documented takedown notices.

How much confusion does a survey need to show? There is no magic threshold, but net confusion of roughly 15% or higher is frequently treated as supporting likelihood of confusion, and figures below that can still matter depending on the market and the rest of the evidence. What matters most is methodology: a proper universe, a control cell, marketplace-realistic stimuli, and a qualified, independent expert.

Key Takeaways

Likelihood of confusion is the central question in trademark infringement, and it is answered not by a mechanical checklist but by a holistic, market-grounded judgment about how real consumers will probably react. The multifactor tests -- Polaroid in the Second Circuit, DuPont before the TTAB and Federal Circuit, Sleekcraft in the Ninth, and their cousins elsewhere -- channel that judgment through mark strength, mark similarity, goods relatedness, actual confusion, intent, channels, and consumer care, with mark similarity and goods relatedness usually carrying the most weight.

Confusion comes in more forms than the cash-register moment: direct, reverse, sponsorship, initial-interest, and post-sale. Consumer sophistication can flip outcomes and should be proven, not assumed. Surveys, when properly designed in the Eveready or Squirt format with a control cell and a credible expert, are among the most persuasive tools available. Liability can extend beyond the direct infringer through contributory and vicarious theories, though platforms answer only for infringement they specifically knew about. Dilution offers famous marks a separate, confusion-free remedy under the TDRA, but only for the rare household-name mark. The internet has not changed the standard, but it has reshaped how channels, intent, and initial-interest confusion are weighed -- and after Jack Daniel's v. VIP Products, expressive and parodic uses no longer bypass the confusion analysis when the mark is used as a source identifier. And winning on confusion is only half the battle: injunctions are reliable, money is discretionary, and a roster of defenses -- classic and nominative fair use, laches, abandonment -- can blunt even a strong claim.

For brand owners, the surest protection is a continuous program: clear before you commit, register early and broadly, build a strength record, monitor relentlessly, enforce proportionately, and govern your licenses. The brands that fare best in the maze of confusion are the ones that mapped the route before they entered it.

Related Articles

This article is provided by mclaw.io for general informational purposes only and does not constitute legal advice, nor does it create an attorney-client relationship. Trademark law is fact-specific and varies by jurisdiction and circuit. Readers should consult qualified trademark counsel about their particular circumstances before adopting, registering, enforcing, or defending a mark.