Checklist · Intellectual Property

Working Up Likelihood of Confusion: A Practical Checklist

Type
Checklist
Jurisdiction
National
Date
2025-01-03
Topic
Intellectual Property

Likelihood of confusion decides most trademark disputes, and it rewards a disciplined workflow rather than a first impression. This checklist gives you a repeatable method for working the question up—choosing the controlling circuit test, assembling the factual record, analyzing each factor on the evidence, and reaching a documented, defensible conclusion. It covers the Second Circuit's Polaroid factors, the Federal Circuit and TTAB's DuPont factors, the Ninth Circuit's Sleekcraft factors, and the regional variants, along with the confusion theories (reverse, initial-interest, post-sale, sponsorship) that quietly reorder the analysis. Each phase carries WHY notes, trap flags, and invented worked examples, and the whole thing ties back to survey design, clearance practice, and litigation posture. Treat it as a research resource and verify the controlling circuit's exact factor list and any deadlines before you rely on it.

Working Up Likelihood of Confusion: A Practical Checklist

Intellectual Property > Trademark Infringement | Published 3 January 2025 - Updated 21 February 2025 | UpLaw - uplaw.us

A factor-by-factor workflow for predicting whether an appreciable number of ordinary consumers are likely to be confused about the source, sponsorship, or affiliation of goods or services—usable for a clearance opinion, an enforcement decision, or an infringement claim or defense, across the Polaroid, DuPont, and Sleekcraft tests.

Keywords: likelihood of confusion, Polaroid factors, DuPont factors, Sleekcraft factors, trademark infringement, strength of mark, similarity of marks, relatedness of goods, actual confusion, reverse confusion, initial interest confusion, post-sale confusion, consumer sophistication, survey evidence, net confusion, Lanham Act

Nearly every U.S. trademark dispute comes down to one question, and it is not "who copied whom" or "who registered first." It is whether a defendant's use of its mark is likely to confuse consumers about where the goods come from, who stands behind them, or whether the two businesses are connected. That single question drives infringement under Lanham Act § 32, 15 U.S.C. § 1114 (registered marks) and § 43(a), 15 U.S.C. § 1125(a) (unregistered marks and unfair competition); it drives registration refusals under § 2(d), 15 U.S.C. § 1052(d); and it decides oppositions and cancellations before the Trademark Trial and Appeal Board. Get the confusion analysis right and most of the case follows. Get it wrong—or work it up sloppily—and everything downstream inherits the error.

The temptation is to treat confusion as a feeling: look at two logos, decide they seem close enough (or not), and move on. Courts have spent sixty years warning against exactly that. Confusion is a legal conclusion built from a defined set of factors, weighed against a developed factual record, and expressed as a probability rather than a possibility. The famous multifactor tests—Polaroid in the Second Circuit, DuPont at the Federal Circuit and TTAB, Sleekcraft in the Ninth—are not scorecards where the side with more checkmarks prevails. They are structured lenses for a holistic prediction, and the same marks can produce different answers depending on the forum, the goods, the buyers, and the theory of confusion in play.

This checklist turns that prediction into a workflow. Work the phases in order. Each - [ ] line is an action you can actually perform, not an abstraction. The WHY and trap notes explain the reasoning and flag the mistakes that sink real analyses. The worked examples are invented—clearly hypothetical—and exist to show doctrine in motion, not to stand in for authority. For the doctrinal background this checklist assumes, see Likelihood of Confusion in Trademark Law: A Brand Owner's Guide; for the enforcement context, see Trademark Infringement and Related Rights: The Enforcement Landscape.

How to use this checklist

You can enter at any phase, but the phases build on one another. Clearance work usually runs all six with a light touch, front-loading Phases 1–3. Enforcement work leans on Phases 2–4, where the marketplace facts and the confusion theory live. Litigation work uses everything, and Phase 6 becomes the deliverable. Whatever the posture, resist the urge to jump straight to "the marks look alike" or "the marks look different." Those are conclusions; the factors are how you earn them.

The doctrinal frame, before you touch a single factor

A few fixed points govern everything below. Fix them in your mind first, because the factors only make sense against them.

Confusion means source, sponsorship, or affiliation confusion. The Lanham Act reaches uses "likely to cause confusion, or to cause mistake, or to deceive." 15 U.S.C. § 1114(1)(a); § 1125(a)(1)(A). That includes confusion about whether the senior user makes the junior user's goods, but also whether it sponsors, endorses, or is affiliated with them. A consumer who correctly believes two products come from different companies but wrongly assumes one licensed or approved the other is still confused in the eyes of the law. See Grotrian, Helfferich, Schulz, Th. Steinweg Nachf. v. Steinway & Sons, 523 F.2d 1331 (2d Cir. 1975).

The standard is probability, judged by the ordinary prudent purchaser. A mere possibility of confusion is not enough; an appreciable number of reasonably prudent consumers must be likely to be confused. Courts do not ask whether the most careless shopper alive could be fooled, nor whether a side-by-side comparison would dispel doubt. They ask about ordinary marketplace conditions and imperfect recollection.

It is a holistic judgment, not arithmetic. Every circuit says no single factor is dispositive and the list is a guide rather than a rote formula. In practice, a handful of factors do most of the work. The leading empirical study of the case law found that similarity of the marks, proximity of the goods, and the defendant's intent predict outcomes far more reliably than the rest, and that courts sometimes let a strong finding on one factor "stampede" the others. Barton Beebe, An Empirical Study of the Multifactor Tests for Trademark Infringement, 94 Cal. L. Rev. 1581 (2006). Knowing which factors carry weight in your circuit and on your facts is the whole game.

The Restatement and McCarthy supply the connective tissue. The Restatement (Third) of Unfair Competition §§ 20–23 (1995) synthesizes the factors across circuits, and McCarthy on Trademarks and Unfair Competition remains the standard treatise citation for every proposition below. When you need a neutral, court-respected articulation of a factor, those are your anchors.

With that frame set, work the phases.

Phase 1 — Frame the question and choose the controlling test

  • [ ] Identify the forum and circuit that will (or would) decide the dispute; the factor list and its weighting differ by circuit.
  • [ ] Determine whether you are analyzing registrability (TTAB / Federal Circuit → apply the DuPont factors) or the right to use / infringement (district court → apply the regional circuit's test).
  • [ ] Select the controlling list: Polaroid (2d Cir.), DuPont (Fed. Cir. / TTAB), Sleekcraft (9th Cir.), or the equivalent regional list (Third Circuit Lapp, Fifth Circuit "digits of confusion," Sixth Circuit Frisch's, Eleventh Circuit Frehling, and so on).
  • [ ] Pin down the legal standard: probability, not mere possibility, of confusion among ordinarily prudent purchasers.
  • [ ] Identify which theory or theories of confusion apply: forward (direct), reverse, sponsorship/affiliation, initial-interest, or post-sale.
  • [ ] Note whether confusion is alleged among purchasers or non-purchasers (dealers, downstream observers, the general public), since both can be actionable.

WHY / traps. The single most consequential early error is running the wrong test in the wrong posture. A DuPont registrability analysis compares the marks and the goods as recited in the application and registration, and it presumes the goods travel in all normal channels to all usual classes of purchasers regardless of what the parties actually do in the market. See In re Detroit Athletic Co., 903 F.3d 1297 (Fed. Cir. 2018). A district-court infringement analysis, by contrast, examines real-world marketplace use—actual packaging, actual trade channels, actual buyers. The two can diverge on identical marks: an applicant can lose at the TTAB on a broad identification while the same parties coexist peacefully in the market, or win registration yet lose an infringement suit because its real-world presentation looks nothing like the drawing. Decide which world you are in before you compare anything.

Flag the confusion theory early, too. In reverse confusion, the strength inquiry flips: what matters is the junior user's commercial saturation—its power to swamp the market and make consumers believe the senior user's goods come from the junior—rather than the senior mark's fame. Big O Tire Dealers, Inc. v. Goodyear Tire & Rubber Co., 561 F.2d 1365 (10th Cir. 1977); A & H Sportswear, Inc. v. Victoria's Secret Stores, Inc., 237 F.3d 198 (3d Cir. 2001). Intent to pass off is not required for reverse confusion, and a court that mechanically applies the forward-confusion strength and intent factors will reach the wrong result. Miss the theory in Phase 1 and every later factor tilts the wrong way.

The circuit factor map

The tests share a common core—strength, similarity, proximity, actual confusion, channels, intent, care—and differ mainly in labels, count, and emphasis. Use this map to pick your list; then apply that circuit's precedent, not a generic composite.

Test Forum Factors, in brief
Polaroid (2d Cir.) S.D.N.Y. and Second Circuit infringement (1) strength of the senior mark; (2) similarity of the marks; (3) proximity of the products; (4) likelihood the senior user bridges the gap; (5) actual confusion; (6) defendant's good faith; (7) quality of defendant's product; (8) sophistication of buyers. Polaroid Corp. v. Polarad Elecs. Corp., 287 F.2d 492 (2d Cir. 1961).
DuPont (Fed. Cir. / TTAB) Registration, oppositions, cancellations Thirteen factors, including similarity of marks; similarity of goods/services; similarity of trade channels; conditions of purchase and buyer sophistication; fame of the prior mark; number and nature of similar marks on similar goods; actual confusion; length of concurrent use without confusion; market interface; and extent of potential confusion. In re E.I. du Pont de Nemours & Co., 476 F.2d 1357 (C.C.P.A. 1973).
Sleekcraft (9th Cir.) Ninth Circuit infringement (1) strength of the mark; (2) proximity of the goods; (3) similarity of the marks; (4) evidence of actual confusion; (5) marketing channels; (6) type of goods and degree of care; (7) defendant's intent; (8) likelihood of expansion. AMF Inc. v. Sleekcraft Boats, 599 F.2d 341 (9th Cir. 1979), abrogated on other grounds by Mattel, Inc. v. Walking Mountain Prods., 353 F.3d 792 (9th Cir. 2003).
Lapp (3d Cir.) Third Circuit infringement Ten factors adapting Polaroid, applied to competing and non-competing goods. Interpace Corp. v. Lapp, Inc., 721 F.2d 460 (3d Cir. 1983).
Frisch's (6th Cir.) Sixth Circuit infringement Eight factors closely tracking Sleekcraft. Frisch's Rests., Inc. v. Elby's Big Boy of Steubenville, Inc., 670 F.2d 642 (6th Cir. 1982).
"Digits of confusion" (5th Cir.) Fifth Circuit infringement Type of mark; similarity of marks; similarity of products; outlets and purchasers; advertising media; intent; actual confusion; degree of care. See Bd. of Supervisors for La. State Univ. Agric. & Mech. Coll. v. Smack Apparel Co., 550 F.3d 465 (5th Cir. 2008).
SquirtCo (8th Cir.) Eighth Circuit infringement Six factors—strength, similarity, competitive proximity, intent, actual confusion, degree of care. SquirtCo v. Seven-Up Co., 628 F.2d 1086 (8th Cir. 1980).
Frehling (11th Cir.) Eleventh Circuit infringement Seven factors; strength and similarity weighted most heavily. Frehling Enters., Inc. v. Int'l Select Grp., Inc., 192 F.3d 1330 (11th Cir. 1999).

The First Circuit uses the Pignons factors, Pignons S.A. de Mecanique de Precision v. Polaroid Corp., 657 F.2d 482 (1st Cir. 1981); the Fourth uses Pizzeria Uno Corp. v. Temple, 747 F.2d 1522 (4th Cir. 1984); the Seventh uses a seven-factor test emphasizing similarity, intent, and actual confusion; the Tenth uses King of the Mountain Sports, Inc. v. Chrysler Corp., 185 F.3d 1084 (10th Cir. 1999). For the mechanics of how the Second Circuit specifically deploys these factors on a dispositive motion, see Applying the Polaroid Factors at Summary Judgment in the Second Circuit, and for how appellate courts review the whole exercise, Standards of Review in Second Circuit Trademark Appeals.

Worked example (hypothetical). Suppose Harbor & Vale, a Rhode Island coffee roaster, opposes an application by Harborvale Tea Co. at the TTAB and, on the same facts, threatens an infringement suit in the District of Connecticut. At the Board, you apply DuPont and compare the marks against the goods as identified—"roasted coffee" versus "tea-based beverages"—presuming both move through all normal grocery and online channels to ordinary consumers. In the district court, you apply Polaroid and look at what the parties actually do: Harbor & Vale sells only through its two cafés and a regional grocery chain; Harborvale Tea ships direct-to-consumer nationwide with completely different packaging. Same marks, same parties, two forums, potentially two different answers. Phase 1 is where you notice that before you have invested in the wrong record.

Which theory of confusion?

These variants get their own treatment in Phase 4, because each one quietly rewrites how you read strength, intent, and even who counts as a "consumer."

Phase 2 — Build the factual record

  • [ ] Collect the senior mark exactly as used and as registered: the drawing, specimens, registration certificate, first-use dates, and current status (live, incontestable, subject to a pending cancellation).
  • [ ] Collect the junior mark as actually used in the marketplace: packaging, trade dress, house marks, advertising, point-of-sale presentation, and dated web and app screenshots.
  • [ ] Document the goods or services of each party, their price points, and their conditions of sale.
  • [ ] Map the trade and marketing channels: where and how each party advertises and sells, and whether those channels genuinely overlap or merely share the internet.
  • [ ] Identify the relevant consumers and their sophistication, purchase frequency, and degree of care.
  • [ ] Gather actual-confusion evidence: misdirected emails, calls, orders, invoices, warranty claims, complaints, and reviews—each dated and sourced.
  • [ ] Investigate third-party use of similar marks on related goods (actual, promoted use, not just registrations) to test the senior mark's real-world strength.
  • [ ] Pull the adoption history of the junior mark: clearance searches, counsel opinions, internal naming and branding files.
  • [ ] Capture and authenticate web evidence properly before it changes or disappears.

WHY / traps. Factors talk to one another, and you cannot weigh them without the underlying facts. Two recurring traps deserve special attention.

First, do not confuse third-party registrations with third-party use. In district court, a crowded-field argument that a senior mark is weak generally requires evidence of actual, marketplace-visible use by others; a stack of dormant registrations proves little about what consumers have been conditioned to see. See Star Indus., Inc. v. Bacardi & Co., 412 F.3d 373 (2d Cir. 2005). At the TTAB and Federal Circuit, the calculus differs: extensive third-party registrations and uses can show that a term is conceptually weak and entitled to a narrow scope, even without full marketplace proof. Juice Generation, Inc. v. GS Enters. LLC, 794 F.3d 1334 (Fed. Cir. 2015); Jack Wolfskin Ausrustung fur Draussen GmbH & Co. KGaA v. New Millennium Sports, S.L.U., 797 F.3d 1363 (Fed. Cir. 2015). Match your evidence to your forum.

Second, anecdote is not admissible confusion. "A customer told me they thought we were the same company" is hearsay if offered for its truth, and it will not survive a motion. There are non-hearsay uses—an out-of-court statement offered to show the declarant's confused state of mind can come in under the state-of-mind exception—but you must plan the evidentiary theory, not stumble into it. Surveys exist precisely to capture confusion in an admissible, quantified form, which is why Phase 5 gets its own treatment.

On web evidence: screenshots of the parties' sites, archived pages, and social media are usually the backbone of the marketplace record, and they are worthless if you cannot get them admitted. Authenticate them under Federal Rule of Evidence 901 with a witness who can describe the capture, or with Wayback Machine and hashing foundations, before the pages mutate. The mechanics are covered in Authenticating Website Screenshots and Web Evidence in Federal Court and, in checklist form, Getting Website Evidence Admitted.

Phase 3 — Work each factor on the evidence

Now analyze each factor against the record you built. Do not score them equally. Most circuits treat mark similarity, mark strength, and relatedness of the goods as the heavy hitters, and a strong showing there can carry a thin record elsewhere; conversely, marks that are genuinely dissimilar can defeat confusion regardless of everything else. Work each factor, assign it a weight, and record your reasoning.

Strength of the senior mark

The two halves of strength do different jobs. Conceptual strength asks how inherently distinctive the mark is; a coined term like KODAK sits at the fanciful end and gets a wide berth, while a descriptive term like "Creamy" for yogurt earns protection only if buyers have come to treat it as a source identifier. Commercial strength asks how much recognition the mark has actually earned. A conceptually weak mark can become commercially powerful through massive use, and a conceptually strong mark can languish in obscurity. Famous, heavily promoted marks receive the broadest scope of protection, on the theory that consumers are more likely to assume any similar mark traces back to the famous source. Kenner Parker Toys Inc. v. Rose Art Indus., Inc., 963 F.2d 350 (Fed. Cir. 1992); Recot, Inc. v. Becton, 214 F.3d 1322 (Fed. Cir. 2000).

Incontestability matters here in a specific way. An incontestable registration cannot be challenged on the ground that the mark is merely descriptive, so an infringer cannot defend by arguing the senior mark never acquired distinctiveness. Park 'N Fly, Inc. v. Dollar Park & Fly, Inc., 469 U.S. 189 (1985). It does not, however, make a weak mark strong for scope-of-protection purposes; a crowded field can still narrow the mark's reach. For what registration and incontestability actually buy the owner, see What Federal Trademark Registration Actually Buys You, and for the distinctiveness spectrum in depth, What the Law Will Protect: Substantive Standards for Trademarks.

Worked example (hypothetical). Lumen for solar-panel installation services is arbitrary as applied—"lumen" describes light output, not installation—so it is conceptually strong. But if a clearance search turns up eleven actively promoted regional installers using LUMEN-, LUMENA-, or LUMEN-adjacent names, the mark's effective scope shrinks; consumers in that field have learned to distinguish small differences. The senior user still owns a valid mark, but its power to stop a newcomer named Lumina Solar is weaker than the raw Abercrombie placement suggests. Strength is a finding about the market, not a label on the mark.

Similarity of the marks

  • [ ] Compare the marks in sight, sound, and meaning, as encountered in the marketplace, not side by side.
  • [ ] Apply the anti-dissection rule: compare overall commercial impression, not isolated shared syllables. Estate of P.D. Beckwith, Inc. v. Comm'r of Patents, 252 U.S. 538 (1920).
  • [ ] Identify the dominant element of each mark and give it appropriate weight, while still assessing the whole. In re Nat'l Data Corp., 753 F.2d 1056 (Fed. Cir. 1985).
  • [ ] Account for house marks and added matter that may (or may not) distinguish otherwise similar marks.
  • [ ] Judge similarity on the basis of imperfect recollection—the consumer who saw the senior mark last week, not the litigator holding both exhibits.

Similarity is the factor courts weight most heavily, and the one where sloppy analysis does the most damage. Two disciplines govern it. The anti-dissection rule forbids chopping marks into fragments to manufacture similarity or difference; the comparison is of the marks as wholes, as consumers actually perceive them. At the same time, courts may identify a dominant portion—often the first word, or the distinctive rather than descriptive element—and give it more weight, because that is what buyers remember. National Data holds both truths at once: you consider the whole, but the whole has a center of gravity. Descriptive or generic components carry little source-identifying weight, so "SUPREME" in two marks contributes far less to similarity than an unusual coined element.

The comparison happens across three axes. Sight covers spelling, length, and visual layout. Sound covers how the marks are pronounced and misheard—critical for goods bought by phone or by ear. Meaning covers connotation; two marks that look and sound different can converge in meaning (the classic being marks that translate to the same idea, engaging the doctrine of foreign equivalents). A powerful showing on any one axis can support similarity even where the others diverge.

Two hard rules trap the careless. First, no side-by-side comparison. Consumers rarely see the marks together; they carry an imperfect memory of one and encounter the other later. Analyzing the marks as if displayed on adjacent shelf-talkers understates confusion. Second, marketplace context matters. A mark printed in a distinctive script, always paired with a house brand, and sold in a controlled channel presents differently than the bare word on a drawing.

Worked example (hypothetical). Cedar & Ash and Cedarash for home fragrances share their entire dominant term. A defendant argues they differ because one adds an ampersand and a space. Under the anti-dissection rule and imperfect recollection, that argument fails: a shopper who remembers a "cedar ash" candle from last month is not parsing punctuation. Now flip it—Cedar & Ash versus Birch & Bone. Both use the "[Wood] & [X]" formula, but the dominant terms differ entirely in sight, sound, and meaning; the shared syntactic frame is not source-identifying. Same structural observation, opposite result, because the analysis turns on commercial impression rather than surface pattern.

Relatedness / proximity of the goods or services

  • [ ] Ask whether consumers would expect the goods to come from, or be affiliated with, the same source—not whether the parties directly compete.
  • [ ] Consider whether the goods are complementary, sold to the same buyers, or used together.
  • [ ] Weigh the strength of the senior mark here: stronger marks reach further across product lines.

Direct competition is not required. The question is whether the goods are related enough that a consumer, seeing similar marks, would assume a common or affiliated source. AMF Inc. v. Sleekcraft Boats, 599 F.2d 341 (relating high-speed and family recreational boats). Complementary products, products sold side by side, and products a single company would plausibly offer under one brand all count as proximate. The stronger and more famous the senior mark, the more distant a product can be and still trigger confusion, because buyers assume the famous house has expanded. Recot, 214 F.3d 1322.

Worked example (hypothetical). Meridian for premium fountain pens sues Meridian for leather notebook covers. The goods do not compete—one writes, one holds paper—but they are classic complements, sold in the same stationery boutiques to the same gift buyers, and it is entirely natural to assume a pen house would extend into leather goods. High relatedness. Change the junior's goods to industrial lubricants and the proximity collapses; no ordinary buyer expects a pen maker to supply factory oil, and only a truly famous senior mark could bridge that distance.

Bridging the gap

  • [ ] Assess the senior user's likely expansion into the junior user's market.
  • [ ] Rely on publicly known or objectively expected expansion, not secret subjective intent.

This Polaroid-specific factor asks whether the senior user is likely to enter the junior's market, such that present coexistence risks future confusion. It cuts both ways: a senior user poised to expand strengthens its claim, while a senior user with no realistic path into the junior's field weakens the proximity story. The intent that counts is objective—announced plans, natural product-line extensions, industry norms—not an unstated hope. In circuits without a distinct bridging factor, the same concern folds into relatedness and the likelihood-of-expansion inquiry (Sleekcraft factor eight).

Actual confusion

Because the ultimate question is likelihood, evidence that confusion has already happened is powerful—courts logically credit real-world confusion over predictions about it. But actual confusion is notoriously hard to gather and easy to overstate. Judicial treatment is inconsistent across and within circuits, both on admissibility and on weight. A few misdirected emails prove little if the parties have done millions of dollars of business alongside each other; the confusion must be more than trivial relative to the opportunity for it. Conversely, the absence of actual confusion after a long period of significant, overlapping use cuts hard against the plaintiff—if these marks were going to confuse people, they would have by now.

Quality matters as much as quantity. Confusion by actual or potential purchasers about source is worth far more than confusion by unrelated third parties, or momentary mix-ups that would never affect a buying decision. Internet-era evidence complicates this: social media surfaces more apparent confusion than ever, but much of it is fleeting, anonymous, and unreliable, and courts increasingly discount casual online mix-ups that do not reflect a genuine purchasing consumer. For the deeper treatment, the actual-confusion factor is where survey evidence (Phase 5) most often substitutes for anecdote.

Worked example (hypothetical). Tidewater Bank and Tidewater Wealth both operate in coastal Virginia. Over four years, the bank logs 38 dated instances: wire transfers sent to the wrong entity, a subpoena served on the wrong company, and clients calling the wealth firm to dispute checking fees. That is high-quality actual confusion—purchasers, about source, with financial consequences. Contrast a single tweet reading "wait, are these the same?"—one anonymous, non-purchasing observer, dispelled in the next reply. The first is close to dispositive; the second is noise.

Intent / good faith

  • [ ] Distinguish deliberate copying or intent to trade on goodwill (supports confusion, and may raise an inference of likely confusion) from mere knowledge of the senior mark (not, by itself, bad faith).
  • [ ] Look for adoption of the junior mark after a clearance search flagged the senior mark and counsel advised against it, or for copying of distinctive, non-functional features.
  • [ ] In reverse confusion, recognize that intent to pass off is not required; carelessness in adopting a mark the senior user already owns can still matter.

Intent is a heavily weighted factor because a defendant who set out to trade on the plaintiff's goodwill has, in effect, conducted its own market test and bet that consumers will be confused. Courts will infer likely confusion from a proven intent to deceive. But the doctrine draws a sharp line: knowing about the senior mark is not the same as intending to exploit it. Businesses run clearance searches precisely so they will know about senior marks; punishing that knowledge would discourage diligence. Bad faith requires something more—an intent to capitalize on the senior user's reputation, evidenced by copying trade dress, mimicking distinctive elements, or pressing forward after counsel warned of the conflict. See Star Indus., 412 F.3d 373; Streetwise Maps, Inc. v. VanDam, Inc., 159 F.3d 739 (2d Cir. 1998).

This is also where the good-faith record you preserve during clearance pays off defensively. A documented clearance search plus a reasoned opinion of counsel is strong evidence of good faith and a primary tool for defeating willfulness and enhanced remedies after Romag Fasteners, Inc. v. Fossil, Inc., 590 U.S. 212 (2020). Building and keeping that record is its own discipline; see Rebutting Willful Trademark Infringement: Clearance Searches, Opinions of Counsel, and USPTO Approval. And note the geographic wrinkle: in some circuits, a junior user's knowledge of a senior mark destroys the good-faith needed for the Tea Rose–Rectanus remote-user defense, as the Ninth Circuit held in analysis worth reading in Stone Creek v. Omnia.

Marketing and trade channels

  • [ ] Assess specific overlap in how and where the parties advertise and sell.
  • [ ] Discount channels that are shared but ubiquitous—"both sell online," "both advertise on social media"—unless the overlap is targeted and real.

Convergent trade channels raise the odds that the same consumer encounters both marks in the same buying context. But the analysis has to be specific. Nearly everyone sells online and advertises on the same handful of platforms, so "both have websites" proves almost nothing. What moves the needle is genuine overlap: the same retailers, the same trade shows, the same targeted ad placements, the same aisle. A senior user selling exclusively through independent bike shops and a junior selling exclusively through big-box sporting-goods chains may reach different buyers even though both technically sell "athletic gear."

Consumer sophistication and degree of care

  • [ ] Calibrate care to price, importance, frequency, and buyer expertise.
  • [ ] Remember that even careful, sophisticated buyers can be confused by very similar marks on very similar goods. Kenner Parker, 963 F.2d 350.
  • [ ] Watch for reverse-confusion situations where sophistication cuts differently.

The more expensive, important, or infrequent the purchase, the more care buyers exercise, and the less likely they are to be confused—a corporate CIO selecting an enterprise software platform will not mistake vendors the way a shopper grabbing a $3 impulse item might. But sophistication is not a magic defense. Expert buyers can still be confused where the marks are nearly identical and the goods closely related, and sophistication in the subject matter does not equal sophistication about trademarks. In the enterprise-software hypothetical, the CIO is sophisticated about software but may reasonably assume that "AcmeFlow" and "Acme Flow" are two products from one company—an assumption about corporate structure, not a lapse in diligence.

Quality of the junior user's goods, and circuit-specific factors

  • [ ] Note the quality factor, which cuts both ways and is usually minor: markedly inferior goods can heighten harm to reputation, while comparable quality can suggest an intent to pass off.
  • [ ] Add any circuit-specific factors: the DuPont length-of-concurrent-use and market-interface factors, Sleekcraft's likelihood-of-expansion factor, and the like.

Quality rarely decides a case, but it colors the equities and the damages story. The circuit-specific factors can matter more than their obscurity suggests. A long period of concurrent use without confusion (DuPont factor eight; and see Nabisco) is strong defensive evidence. The "market interface" factor—consent agreements, prior dealings, laches—can reshape the analysis entirely; for the equitable defenses that ride alongside confusion, see Equitable Defenses: Laches, Acquiescence, Waiver, and Equitable Estoppel.

Phase 4 — The confusion variants that rewrite the analysis

The default is forward, point-of-sale confusion. The variants below change which facts matter, and each one has quietly reversed cases that looked clear under the standard factors.

  • [ ] Reverse confusion. Analyze the junior user's commercial strength and advertising saturation, not the senior's fame. Do not require intent to pass off. Ask whether the junior's market presence is so large that consumers assume the senior's goods originate with, or are sponsored by, the junior. Big O Tire, 561 F.2d 1365; A & H Sportswear, 237 F.3d 198.
  • [ ] Initial-interest confusion. Ask whether the junior mark captures consumer attention through momentary confusion that dissipates before purchase—drawing buyers to the junior's door by borrowing the senior's draw. Recognized in classic form in Mobil Oil, 818 F.2d 254, and Brookfield, 174 F.3d 1036, but treated cautiously in the digital context, where the Ninth Circuit's "Internet troika" and later cases demand real, not theoretical, diversion. Network Automation, Inc. v. Advanced Sys. Concepts, Inc., 638 F.3d 1137 (9th Cir. 2011); Multi Time Machine, Inc. v. Amazon.com, Inc., 804 F.3d 930 (9th Cir. 2015).
  • [ ] Post-sale confusion. Look beyond the buyer to downstream observers. A consumer who knowingly buys a cheap look-alike is not confused, but members of the public who later see the item may be, harming the senior mark. Lois Sportswear, 799 F.2d 867; Ferrari, 944 F.2d 1235.
  • [ ] Sponsorship / affiliation confusion. Separate "who made this" from "who stands behind this." Endorsement, licensing, and affiliation confusion are independently actionable even where source is clear.
  • [ ] Non-purchaser confusion. Confusion among dealers, brokers, and the observing public can be actionable, not just confusion among buyers at the register.

WHY / traps. Reverse confusion is the variant most often missed, and missing it is fatal. A small senior user (say, a regional coffee roaster) confronts a national chain that adopts the same name and blankets the market with advertising. Under a naive forward analysis, the senior mark looks "weak" (little fame) and the junior looks innocent (no intent to trade on an unknown roaster's goodwill), so a court might find no confusion. That is backwards. In reverse confusion, the junior's saturation is the engine of harm: consumers come to believe the roaster is the copycat, or an unauthorized affiliate of the chain, and the senior loses control of its own identity. Analyze the junior's strength and the senior's swamped position, and the case looks entirely different.

Initial-interest confusion demands equal care. Keyword advertising, metatags, and search-result placement generated a wave of initial-interest theories, but the modern courts—especially the Ninth Circuit—have narrowed the doctrine, requiring evidence that consumers were actually diverted and deceived rather than merely exposed to a competitor's labeled result. The keyword-advertising and online-enforcement mechanics live in Defending a Brand on the Internet: A Strategic Enforcement Playbook, and the special problems of virtual goods and expressive uses appear in Trademark Doctrine for Virtual Goods and NFTs: The MetaBirkins Landscape.

One more caution, post-Jack Daniel's. Where a defendant argues its use is an expressive parody, do not assume the Rogers v. Grimaldi threshold sweeps the multifactor test aside. In Jack Daniel's Props., Inc. v. VIP Prods. LLC, 599 U.S. 140 (2023), the Supreme Court held that when a defendant uses the borrowed mark as a source identifier for its own goods, the ordinary likelihood-of-confusion analysis governs in full—parody or not. The expressive-use argument narrows; the factors return.

Phase 5 — Decide whether to commission a survey

  • [ ] Decide whether the case turns on consumer perception enough to justify a survey's cost and lead time.
  • [ ] Choose the format: Eveready (senior mark strong or famous; show only the junior mark and ask open-ended source questions) or Squirt (weaker or genuinely co-encountered marks; show both, with rigorous controls). Union Carbide Corp. v. Ever-Ready Inc., 531 F.2d 366 (7th Cir. 1976); SquirtCo v. Seven-Up Co., 628 F.2d 1086 (8th Cir. 1980).
  • [ ] Define the universe correctly: prospective purchasers of the junior user's goods for forward confusion; the senior user's purchasers for reverse confusion.
  • [ ] Require a control cell and report net confusion (test minus control), not the raw gross figure.
  • [ ] Plan for double-blind administration, neutral non-leading questions, marketplace-realistic stimuli, and objective coding.

WHY / traps. A survey is the only way to put an admissible number on confusion, and a good one can win a close case. A bad one is worse than none: it hands the opponent a Daubert target and a headline percentage to attack, and an excluded survey can poison the expert's credibility on everything else. The design choices are where cases are won and lost.

The universe is the most common fatal flaw. Survey the wrong population—the senior user's customers in a forward-confusion case, or the general public instead of likely buyers—and the results are irrelevant no matter how clean the execution. The control cell is the second: without a control that isolates confusion caused by the marks' similarity from background noise (guessing, pre-existing beliefs, survey artifacts), the gross confusion figure overstates the real effect. Report net confusion. As a rough and heavily caveated benchmark drawn from the case law and the treatises, net confusion around 15% or higher tends to support a likelihood of confusion, figures below roughly 10% tend to negate it, and the 10–15% band is a genuine gray zone that depends on the rest of the record. These are not statutory thresholds; they are patterns, and courts have found confusion on lower numbers and rejected it on higher ones.

Format follows strength. The Eveready format—show only the junior mark and ask consumers who they think puts it out and whether that company makes anything else—works when the senior mark is strong enough that it should spring to mind unprompted; it is the gold standard precisely because it does not lead the respondent. The Squirt format—show both marks in a realistic array and ask about source—suits weaker marks or genuine side-by-side marketplace conditions, but it carries a higher risk of leading respondents and demands tighter controls.

For the full methodology and the litigation stakes, work from the Reference Manual on Scientific Evidence (Fed. Jud. Ctr., 3d ed. 2011), Shari Seidman Diamond's Reference Guide on Survey Research, alongside these UpLaw resources: Reading and Building Trademark Confusion Surveys, Daubert and the Consumer Survey, and the phase-by-phase Designing a Defensible Confusion Survey. Admissibility runs through Federal Rule of Evidence 702 and the Daubert–Kumho gatekeeping framework; a survey with a defensible universe, a proper control, and neutral questions usually clears the bar, with methodological weaknesses going to weight rather than admissibility.

Phase 6 — Weigh holistically and document the conclusion

  • [ ] Weigh the factors as an integrated judgment, identifying which ones actually move the needle in this market—do not count checkmarks.
  • [ ] State whether the result is clear or close, and, if litigation is contemplated, whether the objective factors line up for a possible summary-judgment disposition or whether triable issues remain.
  • [ ] Apply any circuit rule that resolves doubt in favor of the senior / prior user, particularly in registrability. In re Hyper Shoppes (Ohio), Inc., 837 F.2d 463 (Fed. Cir. 1988).
  • [ ] Where the analysis clears a junior mark for adoption, preserve the good-faith record—clearance search plus reasoned opinion—to defeat willfulness later.
  • [ ] Write a dated, evidence-cited memorandum that walks each factor, states the weight assigned, and explains the bottom line.
  • [ ] Flag any dilution angle separately (famous marks under 15 U.S.C. § 1125(c)); dilution does not require confusion and has its own elements.

WHY / traps. The conclusion is a prediction, not a sum. Two analysts can find the same factors present and reach opposite results because they weighed similarity and proximity differently against a weak-mark, crowded-field record. That is not a flaw in the method; it is the method. What separates a defensible opinion from a guess is the reasoning: say which factors you treated as dispositive, why, and what evidence supports each weight. A memo that simply lists eight factors and declares a winner will not survive a skeptical adversary, an examining attorney, or a judge ruling on summary judgment.

On posture: likelihood of confusion is often treated as a fact question (or a mixed question) that resists summary judgment, but where the objective factors overwhelmingly favor one side and the subjective factors cannot rescue the other, courts do grant it. The Second Circuit practice on that maneuver is mapped in Applying the Polaroid Factors at Summary Judgment in the Second Circuit. If the matter is headed to trial, note which issues go to the jury versus the judge, because damages and disgorgement, injunctions, and the confusion question itself fall on different sides of the legal-equitable line—worked through in Judge or Jury in Trademark Litigation.

Keep dilution in its own lane. A famous mark can be diluted by blurring or tarnishment without any likelihood of confusion at all—the harm is to the mark's distinctiveness or reputation, not to consumers' ability to identify source. 15 U.S.C. § 1125(c); Moseley v. V Secret Catalogue, Inc., 537 U.S. 418 (2003) (pre-TDRA). Analyze it separately, with its own fame threshold and its own elements, or you will conflate two doctrines that answer different questions.

Finally, remember what the conclusion feeds. If you are clearing a mark, it flows into the go/no-go decision and the clearance opinion; see Clearing a New Trademark: A Practical Checklist and Trademark Clearance Searching: A Comprehensive Guide. If you are enforcing, it drives the cease-and-desist calculus and the choice among TTAB, UDRP, and federal court—How to Draft a Trademark Cease-and-Desist Letter That Works and the full Trademark Enforcement and Litigation Lifecycle toolkit. If you are responding to a demand, it anchors the defenses and the response options in Answering a Trademark Cease-and-Desist Letter. And if the matter is a registration fight, the same analysis becomes the § 2(d) argument in Overcoming a USPTO Trademark Office Action or the pleaded ground in Opposing a Trademark Before the TTAB.

Common mistakes, in one place

  • Tallying factors instead of weighing them; letting one factor "stampede" the analysis without saying so on the record.
  • Running the wrong test for the posture—a DuPont registrability analysis where an infringement analysis is needed, or vice versa.
  • Missing a reverse-confusion theory, then applying strength and intent backward.
  • Comparing marks side by side, or dissecting them into shared fragments, instead of assessing overall commercial impression on imperfect recollection.
  • Treating mere knowledge of the senior mark as bad faith.
  • Relying on third-party registrations (without use) to prove a crowded field in district court, or ignoring that the TTAB weighs registrations differently.
  • Offering anecdotal "someone told me" confusion with no admissible theory.
  • Commissioning a survey on the wrong universe, with no control, or in the wrong format—and reporting gross instead of net confusion.
  • Assuming a parody or expressive-use argument displaces the factors after Jack Daniel's when the mark is used as a source identifier.
  • Conflating dilution with confusion.

A worked example, end to end (hypothetical)

Pull the phases together on invented facts.

Northwind Outfitters has sold technical rain shells under the mark NORTHWIND since 2009, through independent outdoor retailers and its own website, with steady but modest advertising in the Pacific Northwest. In 2024, a venture-backed apparel startup launches NORTH WYND, a line of urban "weatherproof" jackets, with a nine-figure marketing budget, national department-store placement, and heavy social advertising. Northwind Outfitters comes to you.

Phase 1. Forum: likely the Western District of Washington, so Sleekcraft governs an infringement suit; a parallel TTAB opposition would apply DuPont. Theory: this is textbook reverse confusion—a small senior user swamped by a large junior. Flag it now, because it reorders strength and intent.

Phase 2. Build the record: NORTHWIND's registration and first-use proof; NORTH WYND's packaging, ad spend, and channel data; overlapping retailers; any misdirected customer emails; a crowded-field search for other WIND-formative outdoor marks; and the startup's internal naming files (did a clearance search flag NORTHWIND?).

Phase 3, read through the reverse-confusion lens. Strength—do not dwell on NORTHWIND's modest fame; the engine is NORTH WYND's saturation, which threatens to make consumers think Northwind Outfitters is the copycat. Similarity—near-identical in sound (the "Wynd" respelling changes nothing a listener hears) and meaning, high on sight; very strong. Proximity—both sell weatherproof outerwear to overlapping buyers; high. Actual confusion—if retailers or customers are already asking whether Northwind is "the new national brand's cheaper line," that is high-quality reverse-confusion evidence. Intent—the startup's failure to clear, or its decision to proceed after a hit, matters, though passing-off intent is not required. Channels—overlapping department and outdoor retail; real, not merely "both online." Care—mid-price apparel bought with moderate care; sophistication will not save the junior given the closeness of the marks and goods.

Phase 4. Confirm the reverse-confusion framing controls, and check for post-sale confusion (someone seeing a NORTH WYND jacket on the street assuming it is Northwind).

Phase 5. Commission an Eveready-style survey if NORTHWIND is strong enough in its regional market, or a Squirt design if not, defining the universe as the senior user's prospective purchasers (the reverse-confusion universe), with a proper control, reporting net confusion.

Phase 6. The objective factors—similarity, proximity, channels—line up strongly for Northwind under a correctly framed reverse-confusion analysis, with the junior's saturation supplying the mechanism of harm. Document each weight, note the survey result, flag that the modest-fame senior mark still prevails because the theory is reverse rather than forward, and preserve the reasoning for summary judgment. The mistake to avoid is the one a naive forward analysis invites: concluding "the senior mark isn't famous and the junior didn't mean to trade on it, so no confusion." Framed correctly, this is a strong case.

Primary and secondary authority

Statutes. Lanham Act § 32, 15 U.S.C. § 1114 (infringement of registered marks); § 43(a), 15 U.S.C. § 1125(a) (unregistered marks, unfair competition); § 43(c), 15 U.S.C. § 1125(c) (dilution of famous marks); § 2(d), 15 U.S.C. § 1052(d) (registration bar); §§ 15, 33(b), 15 U.S.C. §§ 1065, 1115(b) (incontestability); § 45, 15 U.S.C. § 1127 (definitions).

The multifactor tests. Polaroid Corp. v. Polarad Elecs. Corp., 287 F.2d 492 (2d Cir. 1961); In re E.I. du Pont de Nemours & Co., 476 F.2d 1357 (C.C.P.A. 1973); AMF Inc. v. Sleekcraft Boats, 599 F.2d 341 (9th Cir. 1979); Interpace Corp. v. Lapp, Inc., 721 F.2d 460 (3d Cir. 1983); Frisch's Rests., Inc. v. Elby's Big Boy, 670 F.2d 642 (6th Cir. 1982); SquirtCo v. Seven-Up Co., 628 F.2d 1086 (8th Cir. 1980); Frehling Enters., Inc. v. Int'l Select Grp., Inc., 192 F.3d 1330 (11th Cir. 1999); Pignons S.A. de Mecanique v. Polaroid Corp., 657 F.2d 482 (1st Cir. 1981); Pizzeria Uno Corp. v. Temple, 747 F.2d 1522 (4th Cir. 1984).

Distinctiveness and similarity. Abercrombie & Fitch Co. v. Hunting World, Inc., 537 F.2d 4 (2d Cir. 1976); Two Pesos, Inc. v. Taco Cabana, Inc., 505 U.S. 763 (1992); Estate of P.D. Beckwith, Inc. v. Comm'r of Patents, 252 U.S. 538 (1920); In re Nat'l Data Corp., 753 F.2d 1056 (Fed. Cir. 1985); Park 'N Fly, Inc. v. Dollar Park & Fly, Inc., 469 U.S. 189 (1985).

Strength, proximity, and third-party use. Star Indus., Inc. v. Bacardi & Co., 412 F.3d 373 (2d Cir. 2005); Recot, Inc. v. Becton, 214 F.3d 1322 (Fed. Cir. 2000); Kenner Parker Toys Inc. v. Rose Art Indus., Inc., 963 F.2d 350 (Fed. Cir. 1992); Juice Generation, Inc. v. GS Enters. LLC, 794 F.3d 1334 (Fed. Cir. 2015); Jack Wolfskin v. New Millennium Sports, 797 F.3d 1363 (Fed. Cir. 2015); Virgin Enters. Ltd. v. Nawab, 335 F.3d 141 (2d Cir. 2003).

Confusion variants and actual confusion. Big O Tire Dealers, Inc. v. Goodyear Tire & Rubber Co., 561 F.2d 1365 (10th Cir. 1977); A & H Sportswear, Inc. v. Victoria's Secret Stores, Inc., 237 F.3d 198 (3d Cir. 2001); Mobil Oil Corp. v. Pegasus Petroleum Corp., 818 F.2d 254 (2d Cir. 1987); Brookfield Commc'ns, Inc. v. West Coast Entm't Corp., 174 F.3d 1036 (9th Cir. 1999); Network Automation, Inc. v. Advanced Sys. Concepts, Inc., 638 F.3d 1137 (9th Cir. 2011); Multi Time Machine, Inc. v. Amazon.com, Inc., 804 F.3d 930 (9th Cir. 2015); Lois Sportswear, U.S.A., Inc. v. Levi Strauss & Co., 799 F.2d 867 (2d Cir. 1986); Ferrari S.p.A. v. Roberts, 944 F.2d 1235 (6th Cir. 1991); Nabisco, Inc. v. Warner-Lambert Co., 220 F.3d 43 (2d Cir. 2000); Mut. of Omaha Ins. Co. v. Novak, 836 F.2d 397 (8th Cir. 1987); Grotrian, Helfferich, Schulz v. Steinway & Sons, 523 F.2d 1331 (2d Cir. 1975).

Intent, remedies, and expressive use. Streetwise Maps, Inc. v. VanDam, Inc., 159 F.3d 739 (2d Cir. 1998); Romag Fasteners, Inc. v. Fossil, Inc., 590 U.S. 212 (2020); Jack Daniel's Props., Inc. v. VIP Prods. LLC, 599 U.S. 140 (2023); Moseley v. V Secret Catalogue, Inc., 537 U.S. 418 (2003).

Surveys and experts. Union Carbide Corp. v. Ever-Ready Inc., 531 F.2d 366 (7th Cir. 1976); SquirtCo v. Seven-Up Co., 628 F.2d 1086 (8th Cir. 1980); Fed. R. Evid. 702; Daubert v. Merrell Dow Pharm., Inc., 509 U.S. 579 (1993); Kumho Tire Co. v. Carmichael, 526 U.S. 137 (1999); Fed. Jud. Ctr., Reference Manual on Scientific Evidence (3d ed. 2011) (Reference Guide on Survey Research).

Secondary sources. McCarthy on Trademarks and Unfair Competition (multi-vol.); Restatement (Third) of Unfair Competition §§ 20–23 (1995); Barton Beebe, An Empirical Study of the Multifactor Tests for Trademark Infringement, 94 Cal. L. Rev. 1581 (2006); USPTO, Trademark Manual of Examining Procedure § 1207 (likelihood-of-confusion refusals).

Verify the controlling circuit's exact factor list and current case law, and confirm any procedural deadlines, before relying on this analysis.

Related UpLaw Resources

Related Guides and Articles

Related Checklists

Related Toolkits

This checklist is general legal information published by UpLaw (uplaw.us), not legal advice, and it creates no attorney-client relationship. Likelihood of confusion is intensely fact-specific and varies by circuit and forum; consult qualified trademark counsel about any particular matter.