Running the Likelihood-of-Confusion Analysis: A Factor-by-Factor Checklist

By ·

Likelihood of confusion is the controlling question in nearly every trademark dispute, and answering it well is a disciplined, factor-by-factor exercise rather than a gut reaction. This checklist gives you a repeatable workflow: frame the question and pick the controlling multifactor test, assemble the evidentiary record, work each factor in turn, account for the variant flavors of confusion, decide whether to commission a survey, and weigh the whole holistically into a defensible written conclusion. It covers the Second Circuit's Polaroid factors, the Federal Circuit and TTAB's DuPont factors, and the Ninth Circuit's Sleekcraft factors, with the registrability-versus-use distinction, B&B Hardware preclusion, the Eveready and Squirt survey formats, and the post-Jack Daniel's treatment of expressive use. Two running examples, WHY notes, and trap flags accompany each phase. This is educational material, not legal advice; verify the controlling circuit's exact factor list and current law before relying on any analysis.

Intellectual Property -> Trademark | Published 28 June 2026 | rightsy.io

What this checklist does for you. Likelihood of confusion is the engine room of American trademark law. Whether you are clearing a new brand, sizing up an enforcement target, drafting a complaint, or defending one, the same question sits at the center: are ordinary consumers likely to be confused about who stands behind these goods? This checklist hands you a repeatable way to answer it — not a hunch, not a vibe, but a structured, evidence-anchored prediction you can write down and defend to a skeptical examiner, adversary, or judge. Work the phases in order. Each - [ ] is something you can actually do. The doctrinal companion to this workflow is the brand owner's field map to likelihood of confusion; this document is the operator's manual.

Confusion is a prediction, not a scoreboard

Start by unlearning the most common amateur mistake. The multifactor tests you are about to apply — Polaroid, DuPont, Sleekcraft, and their regional cousins — are not scorecards. You do not add up the factors, see who has more checkmarks, and declare a winner. Every circuit that has thought about it says the same thing in slightly different words: the factors are a non-exhaustive, non-mechanical guide to a single ultimate question, and they are to be weighed, not counted. AMF Inc. v. Sleekcraft Boats, 599 F.2d 341, 348-49 (9th Cir. 1979) ("no rigid weighing"); In re E.I. du Pont de Nemours & Co., 476 F.2d 1357, 1361 (C.C.P.A. 1973) (factors vary in weight, and "each case must be decided on its own facts").

The legal standard is probability, not possibility. Almost anything is possible; the law cares whether confusion is likely among an appreciable number of ordinary, reasonably prudent consumers exercising ordinary care. The confusion that counts is confusion as to source, sponsorship, affiliation, connection, or approval — codified in the unfair-competition provision of the Lanham Act, § 43(a)(1)(A), 15 U.S.C. § 1125(a)(1)(A), and mirrored for registered marks in § 32, 15 U.S.C. § 1114, and for registrability refusals in § 2(d), 15 U.S.C. § 1052(d). Confusion does not require that anyone think the two products are literally the same item. It is enough that consumers wrongly believe the goods come from, or are blessed by, the same commercial source.

Two consequences follow, and both will save you from rookie errors:

Trap — the tally. If your analysis ends with "five factors favor confusion, three favor the defendant, so plaintiff wins," start over. Identify which two or three factors actually drive this market and explain why. A holistic verdict that names its load-bearing factors beats an arithmetic one every time, and it is the only kind that survives appeal.

Meet the parties (two running examples)

To keep this concrete, we will follow two invented disputes through the whole workflow.

Scenario A — forward confusion. Vesperwing Distilling Co. is a mid-size craft distillery that has sold VESPERWING gin since 2013, owns an incontestable federal registration, pours heavy regional advertising into the mark, and is genuinely well known among cocktail drinkers. A newcomer, Hollow & Vine LLC, is about to launch VESPERWYNG, a line of nonalcoholic botanical "spirits," through the same liquor stores, bottle shops, and online retailers. The senior user is strong; the junior user adopted a near-twin for adjacent goods. This is the classic pattern: the public mistakes the small newcomer for the established brand.

Scenario B — reverse confusion. Hale & Hearth Soup Co. is a single-location New England soup maker that has used HALE & HEARTH since 2016, with modest sales and almost no advertising budget. A national grocery chain, Meridian Markets, rolls out a HALE & HEARTH private-label soup line backed by a multimillion-dollar campaign. Now shoppers who encounter the original assume the small, senior company is the knockoff riding the chain's coattails. Same factor list, inverted logic — and we will see exactly where the inversion bites.

Keep both in mind. The factors behave differently depending on which story you are in, and recognizing the story is half the analysis.

Phase 1 — Frame the question and choose the controlling test

Before a single factor, get the posture right. The most expensive mistakes in this whole exercise happen here, silently, before anyone has looked at a mark.

The registrability-versus-use fork (do not skip this)

The DuPont analysis the USPTO runs is not the same exercise a district court runs, even on identical marks, and confusing the two produces confident, wrong answers.

This is why a § 2(d) refusal does not automatically prove infringement, and why a clean registration does not guarantee a clean field of use. It is also why one Supreme Court case belongs at the front of your mind: B&B Hardware, Inc. v. Hargis Industries, Inc., 575 U.S. 138 (2015). When the TTAB decides likelihood of confusion and the usages adjudicated there are materially the same as those at issue in later litigation, the Board's decision can have issue-preclusive effect in district court. The lesson is practical: a TTAB record is not a sandbox. If you litigate confusion before the Board, you may be litigating it for keeps. Before you ever reach that point, it is worth pulling the opposition and cancellation history of the marks in play — Rightsy's TTAB proceedings database lets you see whether your senior mark (or your opponent's) has already been tested against similar marks, and how the Board came out.

A field guide to the circuit tests

The lists overlap heavily — strength, similarity, proximity, channels, care, intent, actual confusion, and expansion recur almost everywhere — but the labels and the count differ. Use this to translate.

| Forum / circuit | Test (leading case) | Notable features | |---|---|---| | Fed. Cir. / TTAB | DuPont, In re E.I. du Pont, 476 F.2d 1357 (C.C.P.A. 1973) | 13 factors; on-paper comparison; marks + relatedness often dominate; fame plays an outsized role | | 2d Cir. | Polaroid, 287 F.2d 492 (2d Cir. 1961) | 8 factors; includes "bridging the gap" and quality of junior's goods | | 9th Cir. | Sleekcraft, 599 F.2d 341 (9th Cir. 1979) | 8 factors; the "internet troika" (similarity, relatedness, channels) for web cases | | 3d Cir. | Lapp, Interpace v. Lapp, 721 F.2d 460 (3d Cir. 1983) | 10 factors; leading reverse-confusion jurisprudence (A&H Sportswear) | | 5th Cir. | "digits of confusion," e.g., Elvis Presley Enters. v. Capece, 141 F.3d 188 (5th Cir. 1998) | 7-8 "digits"; intent and actual confusion weighted heavily | | 6th Cir. | Frisch's, 670 F.2d 642 (6th Cir. 1982) | 8 factors; Ferrari v. Roberts (post-sale confusion) | | 7th Cir. | AutoZone v. Strick, 543 F.3d 923 (7th Cir. 2008) | 7 factors; treats actual confusion and intent as especially probative | | 8th Cir. | SquirtCo, 628 F.2d 1086 (8th Cir. 1980) | 6 factors; lends its name to the Squirt survey format | | 11th Cir. | Frehling, 192 F.3d 1330 (11th Cir. 1999) | 7 factors; mark strength and similarity weighted heavily | | 1st Cir. | Pignons, 657 F.2d 482 (1st Cir. 1981) | 8 factors |

WHY / traps. The single most consequential early error is running the wrong test in the wrong posture — a DuPont registrability analysis where a real-world infringement analysis is needed, or vice versa. The second is failing to spot a reverse-confusion theory at the outset (Scenario B). In reverse confusion, the strength inquiry flips to the junior user's commercial saturation, intent to "pass off" is not required, and a court that mechanically applies forward-confusion logic will reach the wrong result. Flag the theory before you touch the factors, because the theory rewrites how several of them are read.

Phase 2 — Assemble the evidentiary record

You cannot weigh factors you have not documented. Build the file first; analyze second.

WHY / traps. Two recurring evidentiary potholes. First, do not mistake a stack of third-party registrations for proof of a weak, crowded field in district court — most courts demand evidence of actual, marketplace use before they will discount a senior mark's strength, even though the TTAB and Federal Circuit will credit registrations as evidence of conceptual weakness and the public's habituation to a term. Juice Generation, Inc. v. GS Enters. LLC, 794 F.3d 1334 (Fed. Cir. 2015); Jack Wolfskin Ausrustung v. New Millennium Sports, 797 F.3d 1363 (Fed. Cir. 2015). Second, do not build your actual-confusion case on "a customer once told me they thought you two were the same." That is classic hearsay if offered for its truth; its admissible cousin is a properly designed survey (Phase 5) or a witness with personal knowledge of the confusion. Capture and date your web evidence carefully — undated screenshots prove little.

Phase 3 — Work the factors, one at a time

This is the heart of the exercise. Take each factor in turn, state the evidence, assign a direction and a weight, and note where the factor interacts with its neighbors. Below, each factor carries its checkbox, a plain explanation, a worked illustration from our two scenarios, and the trap that most often snares practitioners.

3.1 Strength of the senior mark

Strength is two questions wearing one label. Conceptual strength asks how inherently distinctive the word is; the ladder runs from unprotectable generic terms up through descriptive marks (protectable only with secondary meaning), suggestive, arbitrary, and fanciful coinages. We unpack the rungs in the Abercrombie spectrum, and the way a weak mark can climb the ladder through use in how marks acquire secondary meaning. Commercial strength asks how much marketplace recognition the mark has actually earned. The two can diverge: KODAK is fanciful and famous; a descriptive mark like a hypothetical "QUICK-DRY" for paint may be conceptually weak yet commercially powerful after decades of advertising.

The stronger the mark, the wider the protective moat — and the more weight similarity carries. In the Federal Circuit, fame of the prior mark "plays a dominant role" and is given great weight when present. Recot, Inc. v. Becton, 214 F.3d 1322, 1327 (Fed. Cir. 2000); Kenner Parker Toys Inc. v. Rose Art Indus., 963 F.2d 350 (Fed. Cir. 1992).

Scenario A. VESPERWING is fanciful (an invented word) — top of the conceptual ladder — and incontestable, with heavy regional advertising. Strength points hard toward confusion and magnifies the danger of a near-identical newcomer.

Scenario B. HALE & HEARTH, the senior soup maker, is conceptually suggestive but commercially modest. In a forward analysis its commercial weakness might cut against confusion. But this is reverse confusion, so the question flips: the relevant strength is Meridian Markets' advertising saturation, which is precisely what lets the giant overwhelm the original. Read the factor the wrong way and you would wrongly conclude the small company has no case.

Trap — the third-party-use shortcut. Pointing to a long list of registrations that include the shared term feels persuasive, but in most district courts dormant registrations prove little about marketplace strength. Show actual, promoted use by third parties, or the "crowded field" argument fizzles.

3.2 Similarity of the marks

Similarity is the factor courts most often call decisive, and it is governed by an old, durable rule: compare the marks in their entireties, as a consumer with an imperfect memory would encounter them, not laid side by side under a magnifying glass. Estate of P.D. Beckwith, Inc. v. Comm'r of Patents, 252 U.S. 538, 545-46 (1920) (the commercial impression of a mark is derived from it as a whole). You may identify a dominant portion — disclaimed or descriptive matter is given less weight — but you may not chop a mark into syllables, find a shared fragment, and declare victory. The "sight, sound, and meaning" trinity is the standard frame: marks are confusingly similar if they look alike, sound alike, or mean alike enough that the overall impression coincides. Palm Bay Imps., Inc. v. Veuve Clicquot Ponsardin, 396 F.3d 1369 (Fed. Cir. 2005) (comparing appearance, sound, connotation, and commercial impression; applying the doctrine of foreign equivalents).

Scenario A. VESPERWING vs. VESPERWYNG: identical sound, near-identical appearance, identical meaning. This is about as strong a similarity showing as exists. Combined with a fanciful, famous senior mark, similarity alone nearly settles the case.

Scenario B. HALE & HEARTH vs. HALE & HEARTH: literally identical. Similarity is not in dispute; the action is elsewhere (strength, read in reverse, and the equities of a giant adopting a small company's exact name).

Trap — the side-by-side and the dissection. Two errors are endemic here. The first is staging an improper side-by-side comparison — consumers do not shop with both labels in hand. The second is dissecting marks into shared letters or syllables to manufacture similarity (or hammering one trivial difference to manufacture distance). Always return to the overall commercial impression in the actual market.

3.3 Relatedness / proximity of the goods or services

Direct competition is not required. Goods are "related" for confusion purposes when the circumstances surrounding their marketing would lead consumers to believe they emanate from the same source. Recot, 214 F.3d at 1329. The relatedness factor runs on a sliding scale with similarity: identical marks on closely related goods is the danger zone; even somewhat different marks on identical goods can confuse.

Scenario A. Gin vs. nonalcoholic botanical "spirits" are not identical, but they are plainly adjacent — same shelf, same cocktail-curious buyer, same bottle aesthetics, with a well-documented market trend of alcohol brands extending into zero-proof lines. A reasonable consumer could easily assume VESPERWYNG is the distillery's nonalcoholic offshoot. Relatedness favors confusion.

Scenario B. Identical goods (soup), identical mark. Relatedness is maximal; this factor is not where the case is won or lost, but it removes any cushion the junior user might have hoped for.

Trap — demanding head-to-head competition. A defendant who says "we don't even compete" has misstated the test. Sponsorship and affiliation confusion does not require competition; brand-extension expectations do the work.

3.4 Trade and marketing channels

Channel overlap raises the odds that the same consumer meets both marks; channel separation lowers them. The honest version of this factor asks whether the same buyers encounter both products in the same context, not whether both happen to exist somewhere on the open web.

Scenario A. VESPERWING and VESPERWYNG share liquor stores, bottle shops, cocktail-focused online retailers, and overlapping advertising aimed at the same cocktail audience. Channel overlap is real and specific. Favors confusion.

Scenario B. Both soups sit on grocery shelves, but here the asymmetry matters: Meridian's national distribution dwarfs the senior user's single location, which is exactly how a junior user achieves the saturation that drives reverse confusion.

Trap — the "both online" overclaim. After the explosion of e-commerce, nearly everyone sells online, so a bare "both use the internet" shows almost nothing. Courts have said as much. Insist on specific, overlapping channels.

3.5 Purchaser sophistication and degree of care

Expensive, considered, professional purchases (industrial equipment, enterprise software) invite scrutiny that reduces confusion; cheap, impulse, everyday purchases invite the opposite. But sophistication is not a magic shield — when the marks are nearly identical and the goods nearly the same, even careful buyers can stumble. In re Research & Trading Corp., 793 F.2d 1276 (Fed. Cir. 1986).

Scenario A. A bottle of gin or nonalcoholic spirit is a modest, often impulse, purchase made by ordinary consumers scanning a shelf. Low care favors confusion.

Scenario B. Soup is a quintessential low-attention grocery buy. Low care, again, favors confusion — and undercuts any defense that shoppers would surely notice the difference.

Trap — overestimating the buyer. Defendants love to argue their customers are too savvy to be fooled. Anchor the argument to the actual relevant buyers and the actual purchase conditions, and remember that high care rarely rescues a near-identical mark on near-identical goods.

3.6 Bridging the gap / likelihood of expansion

Several circuits (notably the Second, in Polaroid) ask whether the senior user is likely to "bridge the gap" between its market and the junior user's; the Ninth's Sleekcraft phrases it as likelihood of expansion. The point is forward-looking: if the senior user would naturally grow into the junior user's space, present coexistence is a poor predictor of future confusion. Courts focus on objective evidence — industry expansion patterns, the senior user's stated plans, consumer expectations — not litigation-driven assertions of intent. Virgin Enters. Ltd. v. Nawab, 335 F.3d 141 (2d Cir. 2003).

Scenario A. Distilleries extending into nonalcoholic spirits is a visible, well-documented industry trend, so consumers might reasonably assume VESPERWING has bridged the gap itself. Favors confusion.

Scenario B. The gap is already zero — both sell soup — so bridging is irrelevant or trivially satisfied.

Trap — secret intentions. A senior user's private, unannounced wish to someday enter a market is not evidence of bridging. Build this factor on objective, public-facing facts.

3.7 Actual confusion

Evidence that real consumers were actually confused is among the most persuasive proof available, because it converts a prediction into a track record. But it is notoriously hard to gather, so its absence is weak counter-evidence unless the parties have coexisted for a long time in overlapping markets without a single credible incident — then the silence speaks. Courts discount confused professionals, fleeting misdirected mail, and confusion that does not bear on purchasing decisions.

Scenario A. Because VESPERWYNG has not launched, there is no actual-confusion evidence yet — which is exactly why this is a clearance and pre-launch enforcement problem, and why a survey (Phase 5) may be the best available proof of likely confusion.

Scenario B. If, after Meridian's rollout, the senior user's social accounts fill with shoppers congratulating it on its "new grocery deal" or complaining about a Meridian product, that is textbook reverse-confusion evidence — and the kind of thing brand monitoring catches early.

Trap — the hearsay anecdote. "Customers tell us all the time" is not admissible for its truth and will not survive an objection. Convert anecdotes into admissible form: percipient witnesses, business records of misdirected orders, or a methodologically sound survey.

3.8 Intent / good faith of the junior user

A junior user who picked its mark intending to ride the senior user's reputation invites a court to presume that it succeeded — an "intent to confuse" can tip several other factors. But knowing about a prior mark is not, by itself, bad faith; businesses adopt marks with eyes open all the time, and the law does not punish mere awareness. The line between awareness and bad faith is exactly where good-faith records earn their keep. A documented clearance search and a reasoned counsel opinion are powerful evidence that adoption was honest — the subject of the shield of good faith. When knowledge tips into deliberate copying, the analysis hardens fast, as the Ninth Circuit illustrated in Stone Creek v. Omnia, where copying a mark with knowledge destroyed any claim of good faith.

Scenario A. If Hollow & Vine's naming deck shows it chose VESPERWYNG because consumers associate the sound with premium spirits, intent swings hard toward confusion. If, instead, it ran a clearance search, got an opinion, and adopted in good faith — that record blunts willfulness even if confusion is ultimately found.

Scenario B. Meridian may have adopted HALE & HEARTH with no intent to deceive at all, and it would not matter much: reverse-confusion liability does not hinge on a passing-off motive. The factor is close to neutral, and a court will not let "we weren't trying to fool anyone" rescue a saturating junior user.

Trap — equating knowledge with bad faith. This is one of the most common analytical errors in the whole exercise. Knowledge of the senior mark is not bad faith. Reserve the bad-faith finding for evidence of intent to exploit the senior user's goodwill.

3.9 Quality of the junior user's goods

This Polaroid factor is rarely decisive. Inferior junior goods sharpen the tarnishment concern; comparable quality can actually make confusion more likely because the products feel like they belong to one family. Either way, do not let it carry weight it cannot bear.

3.10 Circuit-specific extras

WHY / traps. Resist the urge to grade every factor on the same curve. Across most circuits, the heavy hitters are mark similarity, mark strength, and relatedness of the goods; a strong showing on those can outweigh a thin record everywhere else, and dissimilarity alone can defeat confusion outright (Kellogg). Two perennial mistakes: dissecting marks into shared fragments instead of comparing overall impressions, and treating "knowledge of the senior mark" as bad faith. A third, subtler one: scoring an empty factor against the party who bears the burden, when the right move (especially at the TTAB) is to mark it neutral.

Phase 4 — Account for the flavor of confusion

The factors are the same; the kind of confusion you allege changes how you read them and whom you must survey. Identify the flavor explicitly.

WHY / traps. Pleading and proving the wrong flavor wastes the case. Reverse confusion is the most commonly missed theory and the one that most punishes a mechanical analysis. Initial-interest and post-sale theories change the relevant audience — and therefore the survey universe — from purchasers to searchers or to the observing public. Name the flavor in Phase 1, and let it discipline the rest.

Phase 5 — Decide whether to commission a survey

Surveys are the only way to put admissible numbers on consumer perception, and in a close case they can be decisive. They are also expensive, slow, and — if botched — a gift to the other side. Choose deliberately.

WHY / traps. A mediocre survey is often worse than no survey: it hands the opponent a Daubert target under Federal Rule of Evidence 702 and a number to dismantle. As a rough field guide, net confusion around 15% or higher generally supports a finding of likely confusion; below roughly 10% tends to negate it; the 10-15% band is a contested gray zone where design quality decides everything. For methodology, see designing a trademark survey that survives and building a bulletproof consumer survey; for the attacks you must anticipate, Daubert challenges to survey experts. The Federal Judicial Center's Reference Manual on Scientific Evidence (Reference Guide on Survey Research) is the standard playbook on both sides.

Phase 6 — Weigh holistically and write the conclusion

Now integrate. The output of this phase is a dated, evidence-cited memorandum that a stranger could follow and a court could respect.

A memo skeleton you can reuse

A defensible likelihood-of-confusion memo tends to follow the same spine:

  1. Question presented — the marks, goods, forum, posture (registrability vs. use), and theory of confusion.
  2. Controlling test — the named factor list and its source case.
  3. Facts — the senior and junior marks, goods, channels, customers, and any actual-confusion or third-party-use evidence, each with a citation to the record.
  4. Factor-by-factor analysis — direction and weight for each, with interactions noted.
  5. Survey — design, format, universe, control, and net result, if any.
  6. Conclusion — a holistic prediction naming the load-bearing factors, a candor about closeness, and any doubt-resolution rule applied.
  7. Recommendations — clear / modify / abandon / negotiate a consent or coexistence agreement / proceed to enforcement, with next steps.

WHY / traps. The conclusion is a prediction, not arithmetic, and it must be written to survive a hostile reader. The biggest documentation failures are conclusory verdicts with no weighting, and analyses that quietly assume the registrability and infringement questions are interchangeable. When the call is genuinely close, that is itself a finding — say so, and let it drive a risk-calibrated recommendation rather than false confidence.

How the two scenarios come out

Scenario A (forward). A fanciful, incontestable, famous senior mark; a near-identical junior mark; adjacent goods on shared shelves; ordinary, low-care buyers; and an industry trend that makes brand extension plausible. Strength, similarity, and proximity — the heavy hitters — all point the same way, and an Eveready survey would likely confirm it. This is a strong likelihood-of-confusion case, the kind where a senior user should consider a cease-and-desist letter and, if needed, a motion to stop the launch. For Hollow & Vine, the lesson runs the other way: this is a name to abandon at clearance, not defend at trial.

Scenario B (reverse). Identical marks and identical goods, but the action is in the strength factor read backward: Meridian's saturation, not the senior user's modest fame, is what creates confusion, and Meridian's lack of intent to deceive does not save it. A practitioner who applied forward-confusion logic — "the small company's mark is weak, so no confusion" — would blow the case. Spotting the reverse-confusion flavor in Phase 1 is what makes the rest of the analysis come out right.

Common mistakes (a quick audit)

The factors at a glance

| Factor | Core question | Usual weight | Watch for | |---|---|---|---| | Strength of senior mark | How distinctive and recognized is it? | Heavy | Reverse confusion flips to junior's strength; third-party use erodes it | | Similarity of marks | Same sight, sound, or meaning as a whole? | Heavy; can be dispositive | Anti-dissection; no side-by-side; foreign equivalents | | Relatedness / proximity | Would buyers assume a common source? | Heavy | Competition not required; sliding scale with similarity | | Trade / marketing channels | Do the same buyers meet both? | Moderate | "Both online" proves little | | Purchaser care | How careful is the buyer? | Moderate | High care rarely rescues near-identical marks | | Bridging the gap / expansion | Will the senior user grow into the space? | Moderate | Objective evidence only; not secret intent | | Actual confusion | Did real confusion happen? | Heavy if present | Absence weak unless long overlapping coexistence | | Intent / good faith | Did the junior user aim to exploit goodwill? | Variable | Knowledge ≠ bad faith; irrelevant motive in reverse confusion | | Quality of junior goods | Does quality affect reputation/source belief? | Minor | Cuts both ways | | Circuit-specific (DuPont, etc.) | Crowded field, concurrent use, consent, etc. | Variable | Empty factors are neutral, not adverse |

Primary authority

Verify the controlling circuit's exact factor list and current case law before relying on this analysis.

Related Resources

This checklist is general information, not legal advice. Likelihood of confusion is intensely fact-specific and varies by circuit and forum. For a clearance opinion, an enforcement decision, or litigation, consult qualified trademark counsel — including Rightsy's virtual trademark attorneys — about your particular facts.

Read this article on Rightsy