The DuPont Factors: How Courts Decide Likelihood of Confusion

In 1973, the Trademark Trial and Appeal Board issued its decision in In re E.I. du Pont de Nemours & Co., 476 F.2d 1357 (C.C.P.A. 1973), establishing a 13-factor test for likelihood of confusion that has governed trademark law for over half a century. Every Section 2(d) refusal the USPTO issues, every opposition proceeding at the TTAB, and every federal trademark infringement claim in most circuits runs through some version of this framework.

Yet the DuPont factors are widely misunderstood. Practitioners treat them as a checklist when they function more like a weighted algorithm. Examiners focus heavily on two or three factors while nominally acknowledging all thirteen. Courts in different circuits apply their own variations — the Second Circuit's Polaroid factors, the Ninth Circuit's Sleekcraft test — that overlap with DuPont but diverge in meaningful ways.

This guide breaks down all 13 factors, explains which ones carry real weight, walks through landmark cases, and shows how modern clearance data maps to each factor in the analysis.

The Origin of the DuPont Test

The case itself was straightforward. DuPont sought to register RALLY for an all-purpose spray cleaner. The examiner refused registration under Section 2(d) of the Lanham Act, citing a prior registration for RALLY for a combination polish, cleaner, and wax for automobiles. The question was whether consumers would be confused between a general household cleaner and an automotive cleaning product bearing the same name.

The Court of Customs and Patent Appeals (predecessor to the Federal Circuit) reversed the refusal, holding that the goods were sufficiently different and traveled through different trade channels. In doing so, Judge Rich articulated 13 factors to be considered "when relevant" in determining likelihood of confusion. The court was explicit that not all factors are relevant in every case, and that the weight given to each factor depends on the circumstances.

This flexibility is both the test's strength and its source of unpredictability.

All 13 DuPont Factors

Factor 1: Similarity of the Marks

The similarity or dissimilarity of the marks in their entireties as to appearance, sound, connotation, and commercial impression.

This is consistently the most important factor. The TTAB and examining attorneys evaluate marks across four dimensions:

• Appearance — Visual similarity of the words, letters, or design elements
• Sound — Phonetic similarity when spoken aloud
• Connotation — The meaning or idea conveyed by each mark
• Commercial impression — The overall impression the mark makes on consumers in the marketplace

The marks must be compared in their entireties, not dissected into components. However, certain elements may be given more or less weight. Generic or descriptive terms within a mark are typically discounted, while distinctive or dominant elements receive greater attention.

Key case: In In re Viterra Inc., 671 F.3d 1358 (Fed. Cir. 2012), the Federal Circuit held that marks must be considered in their entireties but that "there is nothing improper in stating that, for rational reasons, more or less weight has been given to a particular feature of a mark, provided the ultimate conclusion rests on a consideration of the marks in their entireties."

Practical note: The TTAB frequently finds likelihood of confusion even when marks are not identical. In Bridgestone Americas Tire Operations v. Federal Corp., 673 F.3d 1330 (Fed. Cir. 2012), the court upheld confusion between POTENZA and POTENZA X — the addition of a single letter was insufficient to distinguish the marks for related goods.

Factor 2: Similarity of the Goods or Services

The similarity or dissimilarity and nature of the goods or services as described in an application or registration or in connection with which a prior mark is in use.

The second most critical factor. Goods and services need not be identical — they need only be "related" in a manner that consumers would assume they originate from the same source. The question is whether consumers encountering the goods or services under similar marks would believe the same entity is responsible for both.

The TTAB regularly relies on third-party registrations covering both types of goods to establish relatedness. If multiple registrations exist that cover both software and consulting services, for example, the TTAB may find that these goods are related because the marketplace demonstrates they commonly originate from the same source.

Key case: In In re Davey Products Pty Ltd., 92 USPQ2d 1198 (TTAB 2009), the Board found that cosmetic products and hair care products were related because consumers would expect them to come from the same source, even though they are technically different products.

Factor 3: Similarity of Trade Channels

The similarity or dissimilarity of established, likely-to-continue trade channels.

This factor examines where and how the goods or services reach consumers. Products sold in the same retail environments, through the same distribution networks, or to the same customer base are more likely to cause confusion.

When the goods are described broadly in the application or registration without limitation to specific trade channels, the TTAB presumes that the goods travel through all normal channels of trade for those goods. This presumption frequently works against applicants who argue their goods are sold in different stores or through different distribution methods.

Factor 4: Conditions of Purchase

The conditions under which and buyers to whom sales are made, i.e., "impulse" vs. careful, sophisticated purchasing.

Expensive goods purchased by sophisticated buyers after careful deliberation are less likely to cause confusion than inexpensive impulse purchases. A consumer buying a $50,000 industrial machine presumably exercises more care than someone grabbing a $3 snack at a convenience store.

However, the TTAB applies this factor cautiously. Even sophisticated purchasers can be confused by highly similar marks on related goods. And the Board has repeatedly held that when the identification of goods contains no price limitation, it must be assumed that the goods include both expensive and inexpensive items.

Factor 5: Fame of the Prior Mark

The fame of the prior mark (sales, advertising, length of use).

Fame is a "dominant" factor that can overwhelm other considerations. A famous mark receives a broader scope of protection, meaning confusion may be found even with less similar marks or less related goods. The TTAB evaluates fame based on sales volume, advertising expenditure, length and exclusivity of use, market share, and consumer recognition studies.

The Federal Circuit has recognized several tiers: a mark can be famous for likelihood of confusion purposes without meeting the higher threshold for dilution under Section 43(c). The TTAB evaluates fame on a spectrum from very strong to very weak.

Key case: In Bose Corp. v. QSC Audio Products Inc., 293 F.3d 1367 (Fed. Cir. 2002), the court noted that "fame of the prior mark is a dominant factor in the likelihood of confusion analysis" and that "famous marks enjoy a wide latitude of legal protection."

Factor 6: Number and Nature of Similar Marks on Similar Goods

The number and nature of similar marks in use on similar goods.

This is one of the most underutilized defense factors. If numerous third parties use similar marks for similar goods, consumers may be accustomed to distinguishing between them based on subtle differences. A "crowded field" of registrations can narrow the scope of protection afforded to any single mark.

However, the evidence must show actual use, not merely registration. The TTAB has held that third-party registrations alone, without evidence of use, are not sufficient to establish a crowded field. They are, however, relevant to show that a term has a recognized descriptive or suggestive meaning in the industry.

Factor 7: Nature and Extent of Any Actual Confusion

The nature and extent of any actual confusion.

Evidence of actual confusion is powerful but not required. The TTAB has consistently held that a likelihood of confusion can be established without any evidence of actual confusion — particularly when the marks have not yet been used concurrently in the marketplace.

Conversely, evidence of actual confusion is the "best evidence" of likelihood of confusion. Consumer surveys, misdirected communications, misdirected deliveries, and testimony from consumers who were actually confused all carry significant weight.

Key case: In Majestic Distilling Co. v. Brandon, 315 F.3d 1311 (Fed. Cir. 2003), the Federal Circuit noted that "the absence of actual confusion evidence is rarely significant" in TTAB proceedings because the marks may not yet coexist in the marketplace.

Factor 8: Length of Concurrent Use Without Confusion

The length of time during and conditions under which there has been concurrent use without evidence of actual confusion.

This factor can favor the applicant if both marks have coexisted in the marketplace for a significant period without evidence of confusion. However, the weight given to this factor depends on the extent of marketplace exposure. If both parties operate in small, geographically separate markets, the absence of confusion may simply reflect the absence of consumer overlap rather than the absence of confusing similarity.

Factor 9: Variety of Goods on Which a Mark Is Used

The variety of goods on which a mark is or is to be used.

If the prior mark is used on a wide variety of goods, consumers may expect the mark owner to expand into new product categories — making confusion more likely when a junior user adopts a similar mark for yet another product category. This factor reflects the "natural zone of expansion" concept.

Factor 10: Market Interface Between the Parties

The market interface between applicant and the owner of a prior mark — such as assignment, license, or consent agreements, or the like.

Evidence of a relationship between the parties, such as a prior licensing arrangement, a consent agreement, or an assignment, is relevant. A legitimate consent agreement between the parties can weigh heavily against a finding of confusion. The TTAB generally respects consent agreements unless they fail to address the likelihood of confusion concern.

Factor 11: Applicant's Right to Exclude Others

The extent to which applicant has a right to exclude others from use of its mark on its goods.

If the applicant holds a valid registration for the same or similar mark for other goods, this factor weighs in the applicant's favor. It reflects the idea that the applicant has already been vetted by the USPTO for confusing similarity with prior marks.

Factor 12: Extent of Potential Confusion

The extent of potential confusion, i.e., whether de minimis or substantial.

This factor asks whether any confusion that might occur would be limited or widespread. A finding that confusion would affect only a tiny fraction of consumers weighs against a finding of likelihood of confusion. However, the TTAB rarely resolves cases on this factor alone.

Factor 13: Any Other Established Fact Probative of the Effect of Use

Any other established fact probative of the effect of use of the mark applied for on the state of mind of purchasers or users.

This is the catch-all factor, allowing consideration of any evidence relevant to confusion that does not fit neatly into the other twelve factors. It can include evidence of intent to trade on another's goodwill, evidence of bad faith, or evidence relating to the specific market dynamics of the industry.

Which Factors Actually Matter Most

Not all DuPont factors are created equal. Empirical analysis of TTAB decisions reveals a clear hierarchy.

The Federal Circuit has explicitly acknowledged this hierarchy. In In re Guild Mortgage Co., 912 F.3d 1376 (Fed. Cir. 2019), the court stated that "the first two DuPont factors — the similarity of the marks and the similarity of the goods or services — are typically the most important."

In practice, if the marks are highly similar and the goods are related, the applicant faces an uphill battle regardless of what the remaining factors show. Conversely, if the marks are clearly dissimilar, the analysis often ends there without detailed consideration of other factors.

The Polaroid Test: The Second Circuit's Variation

The Second Circuit does not use the DuPont factors. Instead, it applies the eight-factor test from Polaroid Corp. v. Polarad Electronics Corp., 287 F.2d 492 (2d Cir. 1961), decided twelve years before DuPont:

1. Strength of the senior mark
2. Similarity of the marks
3. Proximity of the products
4. Likelihood of the senior mark holder bridging the gap
5. Evidence of actual confusion
6. Defendant's good faith in adopting the mark
7. Quality of the defendant's product
8. Sophistication of the buyer

The Polaroid test overlaps substantially with DuPont but includes two notable differences. First, the "bridging the gap" factor — whether the senior user is likely to expand into the junior user's market — has no direct counterpart in DuPont (though it relates to Factor 9). Second, the "good faith" factor explicitly considers the junior user's intent, which DuPont addresses only through the catch-all Factor 13.

Key case: In Starbucks Corp. v. Wolfe's Borough Coffee Inc., 588 F.3d 97 (2d Cir. 2009), the Second Circuit found no likelihood of confusion between STARBUCKS and CHARBUCKS for coffee products, despite the marks' phonetic similarity. The court emphasized that the Polaroid factors require a holistic analysis and that no single factor is dispositive.

Other Circuit Variations

Each federal circuit has its own named test, though the substance is largely similar:

Despite the different names and factor counts, the core inquiry is always the same: are consumers likely to be confused about the source of the goods or services? The most consistently important factors across all circuits are mark similarity, goods/services relatedness, and the strength of the senior mark.

The Sleekcraft Test (9th Circuit)

The Ninth Circuit's Sleekcraft test — from AMF Inc. v. Sleekcraft Boats, 599 F.2d 341 (9th Cir. 1979) — is particularly influential because of the volume of trademark litigation in California. Its eight factors are:

1. Strength of the mark
2. Proximity of the goods
3. Similarity of the marks
4. Evidence of actual confusion
5. Marketing channels used
6. Type of goods and degree of consumer care
7. Defendant's intent in selecting the mark
8. Likelihood of expansion of product lines

The Ninth Circuit has applied Sleekcraft extensively in internet-related cases, including disputes over domain names, metatags, and keyword advertising. In Network Automation Inc. v. Advanced Systems Concepts Inc., 638 F.3d 1137 (9th Cir. 2011), the court held that the internet context may alter the relative importance of the factors — particularly reducing the weight of marketing channels (since everyone is online) while increasing the importance of consumer sophistication.

How Examining Attorneys Apply DuPont in Practice

At the USPTO, examining attorneys apply the DuPont factors during ex parte examination when evaluating Section 2(d) refusals. Understanding their practical approach is essential for anticipating and responding to refusals.

The Examiner's Workflow

1. Search the register. The examiner searches the USPTO database for marks that are similar in appearance, sound, or meaning to the applied-for mark, focusing on marks in the same or related classes.

2. Evaluate mark similarity (Factor 1). The examiner compares the marks in their entireties, considering appearance, sound, connotation, and commercial impression.

3. Evaluate goods/services relatedness (Factor 2). The examiner reviews the identification of goods/services in both the application and the cited registration. Broad identifications are construed to encompass all goods within their scope.

4. Presume overlapping trade channels (Factor 3). Unless the identifications contain specific limitations, the examiner presumes the goods travel through all normal trade channels to all normal purchasers.

5. Issue the refusal. If Factors 1-3 support confusion, the examiner issues a Section 2(d) refusal. Factors 4-13 are typically addressed only if the applicant raises them in response.

This means the initial refusal often rests on just three factors, with the remaining ten available as arguments for the applicant's response.

Mapping Clearance Data to DuPont Factors

Modern trademark clearance tools generate data that maps directly to the DuPont analysis. Understanding this mapping transforms clearance from a pass/fail exercise into a predictive risk assessment.

Factor 1: Mark Similarity Data

Trademark search algorithms evaluate similarity across the same dimensions that courts consider:

• Phonetic algorithms (Metaphone, Double Metaphone) map to the "sound" dimension
• String similarity metrics (Levenshtein distance, Jaro-Winkler) map to "appearance"
• Semantic analysis maps to "connotation" and "commercial impression"

Signa's API returns similarity scores across all three dimensions, allowing counsel to quantify the similarity argument rather than relying on subjective assessment alone.

Factor 2: Goods/Services Relatedness

The relatedness of goods can be assessed by examining:

• Class overlap — Whether the marks are in the same or related Nice classes
• Coordinated classes — USPTO maintains a list of classes that are considered related for examination purposes
• Third-party registration analysis — Whether other registrations cover both categories of goods, suggesting the marketplace treats them as related

Signa's search results include class information and goods descriptions that enable rapid assessment of Factor 2 exposure.

Factor 3: Trade Channel Analysis

Trade channels can be inferred from the identification of goods and services. Broad identifications imply broad trade channels. The same clearance data that informs Factor 2 — specifically the goods/services descriptions — also informs the trade channel analysis.

Factor 5: Fame Assessment

Fame can be estimated through:

• Registration history — Marks with long registration histories and multiple renewals suggest established use
• Portfolio breadth — Owners with registrations across many classes suggest commercial significance
• Filing activity — Active enforcement (evidenced by recorded assignments, opposition filings) suggests a vigilant brand owner

Factor 6: Crowded Field Evidence

The number of similar marks on the register for similar goods is directly quantifiable through search data. A clearance search that returns dozens of live registrations incorporating a similar term for similar goods provides strong crowded field evidence under Factor 6.

Building a DuPont Analysis from Clearance Results

When evaluating a potential conflict identified during clearance, structure your analysis as follows:

Step 1: Quantify Mark Similarity

Run the candidate mark against the cited mark using phonetic, visual, and semantic similarity algorithms. Document the scores. A Levenshtein distance of 1-2 (e.g., SIGNA vs. SIGNIA) combined with a high phonetic similarity score indicates significant Factor 1 risk.

Step 2: Map Goods Relatedness

Compare the goods/services descriptions. Identify whether they fall in the same class, related classes, or unrelated classes. Search for third-party registrations covering both categories to assess whether the goods are considered related in the marketplace.

Step 3: Assess the Strength of the Cited Mark

Evaluate whether the cited mark is inherently distinctive (fanciful, arbitrary, suggestive) or weak (descriptive, generic elements). Check for evidence of fame — portfolio breadth, registration duration, enforcement activity.

Step 4: Search for Crowded Field Evidence

Count the number of live registrations incorporating the same or similar term for the same or related goods. If the field is crowded, this narrows the cited mark's scope of protection.

Step 5: Synthesize

Weight the factors according to the hierarchy established by the Federal Circuit: mark similarity and goods relatedness first, fame and crowded field second, all others as relevant. Assign a risk level and document the basis for your assessment.

Common Mistakes in DuPont Analysis

Overweighting Differences

Applicants frequently focus on the differences between marks rather than the similarities. The legal standard is whether the marks are similar enough to cause confusion — not whether a careful, side-by-side comparison reveals differences. The TTAB has repeatedly held that "marks must be considered in light of the fallibility of memory" — consumers rarely encounter the marks simultaneously.

Ignoring the Breadth of Identifications

An identification of "clothing" in Class 25 covers everything from socks to wedding dresses. An identification of "software" in Class 9 covers everything from video games to enterprise ERP systems. The TTAB will not read limitations into a broad identification, even if the applicant's actual products are narrow.

Treating the Factors as a Scorecard

DuPont is not a majority-wins test. A party cannot prevail simply by arguing that seven of thirteen factors favor them. A single dominant factor — like the fame of the prior mark — can override all others.

Neglecting the Crowded Field Argument

Factor 6 is one of the strongest defensive arguments available, but practitioners frequently fail to develop it. Conducting a thorough search for third-party registrations in the same space and presenting that evidence can meaningfully narrow the scope of protection afforded to the cited mark.

How Signa Supports DuPont Analysis

Signa's trademark search API is designed to generate the data that DuPont analysis requires. When you run a clearance search through Signa, you receive:

• Multi-dimensional similarity scoring — Phonetic, visual, and semantic similarity metrics that map directly to Factor 1's four dimensions of comparison
• Comprehensive goods/services data — Full identification text and class information for every result, enabling Factor 2 and Factor 3 analysis
• Owner portfolio data — Registration counts, class breadth, and filing history that inform the Factor 5 fame assessment
• Result volume metrics — The total number of similar marks returned for similar goods, providing quantitative Factor 6 crowded field evidence
• Coverage across 200+ offices — Results from USPTO, EUIPO, WIPO, UKIPO, CNIPA, and hundreds of other jurisdictions, ensuring your DuPont analysis accounts for the global landscape

The goal is not to replace legal judgment — no algorithm can do that. The goal is to ensure your legal judgment is informed by complete, accurate, and systematically organized data. A DuPont analysis built on comprehensive clearance data is more defensible, more predictable, and more efficient than one built on manual search and subjective assessment alone.

---

This guide is intended as an educational resource and does not constitute legal advice. The DuPont factors are applied on a case-by-case basis, and outcomes depend on the specific facts of each situation. Always consult with a qualified trademark attorney for matters involving likelihood of confusion analysis.

Sources:

• In re E.I. du Pont de Nemours & Co., 476 F.2d 1357 (C.C.P.A. 1973)
• Polaroid Corp. v. Polarad Electronics Corp., 287 F.2d 492 (2d Cir. 1961)
• AMF Inc. v. Sleekcraft Boats, 599 F.2d 341 (9th Cir. 1979)
• USPTO — Trademark Manual of Examining Procedure (TMEP) §1207.01: Likelihood of Confusion
• TTAB Manual of Procedure (TBMP) §300: Ex Parte Appeals
• In re Viterra Inc., 671 F.3d 1358 (Fed. Cir. 2012)
• In re Guild Mortgage Co., 912 F.3d 1376 (Fed. Cir. 2019)
• Starbucks Corp. v. Wolfe's Borough Coffee Inc., 588 F.3d 97 (2d Cir. 2009)
• Network Automation Inc. v. Advanced Systems Concepts Inc., 638 F.3d 1137 (9th Cir. 2011)