DuPont Factors: Trademark 2(d) Likelihood of Confusion

The DuPont factors are the thirteen considerations the U.S. Patent and Trademark Office uses to decide whether a proposed trademark is likely to be confused with one already registered. They come from the 1973 decision In re E. I. du Pont de Nemours & Co., 476 F.2d 1357, and every examining attorney and Trademark Trial and Appeal Board judge applies them when a Section 2(d) likelihood-of-confusion question is on the table.1TransJurLex. Application of E. I. DuPont De Nemours and Co., 476 F.2d 1357

Section 2(d) of the Lanham Act, codified at 15 U.S.C. § 1052(d), bars registration of any mark so similar to an existing registered mark that consumers would likely confuse the two.2Office of the Law Revision Counsel. 15 USC 1052 – Trademarks Registrable on Principal Register The statute doesn’t say how to measure that likelihood. DuPont filled the gap.

The Thirteen Factors

The court in DuPont laid out thirteen factors that must be considered “when of record,” while cautioning that there is “no litmus rule which can provide a ready guide to all cases.”1TransJurLex. Application of E. I. DuPont De Nemours and Co., 476 F.2d 1357

  1. Similarity or dissimilarity of the marks in appearance, sound, meaning, and overall commercial impression.
  2. Similarity or dissimilarity of the goods or services.
  3. Similarity of established trade channels.
  4. Conditions under which buyers make purchases — impulse versus careful, researched decisions.
  5. Fame of the earlier mark.
  6. Number and nature of similar marks in use on similar goods.
  7. Nature and extent of any actual confusion.
  8. Length of time the marks have coexisted without evidence of confusion.
  9. Variety of goods on which the mark is or isn’t used.
  10. Market interface between the applicant and the prior mark owner, including consent agreements.
  11. Extent to which the applicant has a right to exclude others.
  12. Extent of potential confusion — how widespread it could be.
  13. Any other established fact relevant to the effect of the marks on consumers.

Which Factors Actually Decide Most Cases

There is no scorecard. Examiners don’t tally which side “wins” more factors. It is a qualitative balancing test, and the weight given to each factor shifts with the facts of the case. Factors that lack evidence in the record get no weight at all.1TransJurLex. Application of E. I. DuPont De Nemours and Co., 476 F.2d 1357

In practice, factors 1 and 2 do the heavy lifting. The TMEP identifies them as the “key considerations” examiners always address; the rest come in only when the record supports them.3BitLaw. TMEP 1207.01 – Likelihood of Confusion Factors 11 through 13 rarely surface and are sometimes called the seldom-litigated factors. And a single factor can be dispositive on its own: the Federal Circuit confirmed in a 2026 decision that a clear lack of similarity in appearance, sound, meaning, and commercial impression can defeat a confusion claim even when other factors lean the other way.4IP UPDATE. DuPont Factors

Similarity of the Marks

This is where most refusals live or die. Examiners compare marks in their entireties across four dimensions: appearance, sound, meaning, and overall commercial impression. The test isn’t a side-by-side comparison. It’s whether someone encountering the marks at different times and places would assume the goods come from the same source.3BitLaw. TMEP 1207.01 – Likelihood of Confusion

Two marks that look nothing alike on paper can still trigger a refusal if they sound identical when spoken. KLEAR and CLEAR differ visually but are indistinguishable over the phone. The reverse also happens. In a 2026 TTAB decision, the Board found that MERCI and DANKE shared an identical meaning for chocolate products (both mean “thank you”), yet looked and sounded so different that confusion was unlikely.

The doctrine of foreign equivalents adds a wrinkle. When a mark uses a non-English word, examiners ask whether the ordinary American buyer would stop and mentally translate it. If the average consumer would take the word at face value, the doctrine doesn’t apply. If the language is commonly spoken and the translation matches an existing mark’s meaning, that overlap counts against the applicant.

Relatedness of the Goods, and the Sliding Scale

The goods or services don’t need to be identical, or even competitive, to create a conflict. They need to be related enough that consumers would assume they come from the same company. A mark on luxury handbags might conflict with a similar mark on travel accessories, because shoppers encountering both in the same store section would naturally assume one company made both.

Factors 1 and 2 operate on a sliding scale. The more similar the marks, the less similar the goods need to be to trigger a refusal. When marks are identical or nearly identical, even a loose relationship between the goods can be enough. When marks share only a slight resemblance, the goods need to be closely related before confusion becomes likely.5BitLaw. TMEP 1207.01(a) – Relatedness of the Goods or Services

Trade Channels and Buyer Sophistication

Even related goods might not cause confusion if they travel through completely different distribution networks. An industrial chemical sold through trade-specific distributors occupies a different world from a consumer cleaning product on grocery shelves, even with a similar name. Examiners look at where products are actually sold — the same retailers, specialty shops, or trade shows — to assess whether a buyer would realistically encounter both in the same purchasing context.

When an application describes goods broadly without limiting trade channels, the USPTO assumes they travel through all normal channels for those goods. That assumption tends to favor finding confusion.

Buyer sophistication cuts in the other direction. Someone spending $50,000 on medical equipment researches the manufacturer and reviews specifications. Someone grabbing a $3 snack at checkout does not. The level of care buyers exercise directly affects confusion risk. But there is a ceiling on how much this helps: even sophisticated purchasers can be confused by very similar marks, and the analysis is pegged to the least sophisticated potential purchaser in the relevant group, not the most careful one.6BitLaw. TMEP 1207.01(d)(vii) – Sophisticated Purchasers

Fame and Third-Party Use

Fame sits on a spectrum, and a mark’s position on that spectrum directly affects the scope of protection it receives. A widely recognized mark blocks a broader range of goods than an obscure one, because consumers are more likely to assume anything bearing a similar name is connected to the famous brand. Proof of fame draws on sales volume, advertising spend, length of use, media coverage, and consumer recognition. The Federal Circuit has said that a mark’s “success in the marketplace” and “popularity” must be weighed.1TransJurLex. Application of E. I. DuPont De Nemours and Co., 476 F.2d 1357 A mark can also be famous within a niche industry, which still expands its protective reach within that field.

Third-party use runs in the opposite direction. If dozens of companies in the same industry already use a similar term, the prior mark is considered weak and gets a narrower scope of protection. Consumers who regularly encounter the same word across multiple brands learn to distinguish them based on other details. Evidence of third-party use typically comes from the USPTO’s own database: an applicant facing a 2(d) refusal can pull other active registrations containing the same word or root for related goods and present those results to the examiner.3BitLaw. TMEP 1207.01 – Likelihood of Confusion

Actual Confusion, Coexistence, and Consent

Documented instances of actual confusion are some of the most persuasive evidence available. Misdirected calls, emails meant for the other company, confused social media comments, and customer service tickets from people who mixed up the brands show the problem is real rather than hypothetical.

Actual confusion isn’t required, though. The legal standard is likelihood of confusion, not proof of it. An examiner can block a registration based entirely on the other factors even when no consumer has ever been confused in practice.

Factor 8 is the flip side. Similar marks that have coexisted in the marketplace for years without generating any evidence of confusion carry weight. The TTAB must consider peaceful coexistence evidence when it is in the record, though an uncorroborated declaration from one party alone may carry limited weight.

Consent agreements fall under factor 10, and not all consent agreements are treated equally. The USPTO draws a sharp line between what it calls “naked” consent and a detailed, reasoned agreement.7BitLaw. TMEP 1207.01(d)(viii) – Consent Agreements A naked consent is essentially a letter saying “we agree to let them register” with no explanation. It carries little weight. A persuasive agreement explains the differences between the marks, describes how the goods or services differ, identifies the distinct customer bases, and outlines any steps the parties are taking to prevent confusion, such as using house marks or restricting certain marketing channels. Even a strong consent agreement doesn’t guarantee approval, but the USPTO is generally reluctant to substitute its own judgment for that of the actual marketplace competitors who have agreed they can coexist.

If You Have Received a 2(d) Refusal

When an examiner issues a likelihood-of-confusion refusal, the Office Action will identify the registered mark the examiner believes conflicts with yours and the DuPont factors supporting the refusal. You generally have three months from the issue date to respond, with an additional three-month extension available for a fee. Applications filed under the Madrid Protocol get six months with no extensions. Miss the deadline and your application is abandoned; filing fees are not refunded.8United States Patent and Trademark Office. Responding to Office Actions

Your response should address the specific factors the examiner relied on. The most common approaches:

  • Distinguish the marks in appearance, sound, meaning, and commercial impression, and gather third-party registration evidence if the shared element is weak or common in the industry.
  • Narrow your identification of goods or services to eliminate the overlap that triggered the refusal. Once narrowed, the description generally can’t be broadened back later.8United States Patent and Trademark Office. Responding to Office Actions
  • Negotiate a detailed consent agreement with the owner of the cited mark, with specific reasons why confusion is unlikely.
  • Argue buyer sophistication if your goods are expensive or sold exclusively to professional purchasers, backed by evidence of the actual purchasing conditions.

Moving your application to the Supplemental Register will not fix a 2(d) refusal. The examiner can still block registration based on likelihood of confusion, and the switch can create new complications if additional conflicting marks were filed in the interim.