Confusingly similar trademarks are marks similar enough that ordinary consumers might mistake one for another in the marketplace. The USPTO evaluates likelihood of confusion using the DuPont factors — a multi-factor test weighing similarity of the marks, relatedness of goods, and market overlap. No single factor controls; the examining attorney weighs the combination.
Apply the DuPont factors to your proposed mark against the closest matches in your search results — each factor either supports or undermines clearance.
The USPTO and federal courts both use DuPont; applying the framework in advance predicts how examination will evaluate the mark.
Score your top candidate against each DuPont factor relative to the two or three closest search matches.
Confusingly similar is the legal standard the USPTO applies when deciding whether one trademark is too close to another. The standard comes from 15 U.S.C. §1052(d), which allows the USPTO to refuse registration for marks “so resembling” an existing mark as to be “likely, when used on or in connection with the goods of the applicant, to cause confusion, or to cause mistake, or to deceive.”[1]
The question is not whether confusion is certain or whether any consumer has actually been confused. The question is whether confusion is likely — whether an ordinary consumer encountering both marks in their normal context would probably mistake one for the other. This probabilistic standard produces real judgment calls that the USPTO and courts handle through the DuPont factor analysis.
The DuPont factors come from In re E. I. DuPont deNemours & Co., 476 F.2d 1357 (C.C.P.A. 1973), a foundational case establishing the multi-factor test for likelihood of confusion. The factors are used both at the USPTO during examination and in federal infringement litigation.[2]
In practice, the first three or four factors dominate most USPTO examinations. Similarity of the marks, relatedness of goods, similarity of trade channels, and purchase conditions together usually determine the outcome. The other factors matter when specific facts make them relevant but are often tangential to the core analysis.
Mark similarity analysis considers four dimensions: sight (visual appearance), sound (phonetic similarity when pronounced), meaning (semantic content), and commercial impression (overall perception in context). Similarity on any of the four dimensions can support a finding of confusing similarity.
| Dimension | Example pair | Why similar |
|---|---|---|
| Sight | STARBUCK’S vs. STARBOX | Visual overlap of “STAR” plus similar length |
| Sound | LEGGO vs. LEGO | Phonetically identical despite spelling difference |
| Meaning | BREATH FRESHENER vs. MOUTH REFRESHER | Same semantic concept expressed differently |
| Commercial impression | KING BURGER vs. BURGER KING | Same overall brand impression with word order reversed |
The USPTO considers all four dimensions together. Marks that look similar but sound different may still be confusing if the visual similarity is strong enough. Marks that look different but sound identical may still be confusing based on sound alone. The cumulative impression often matters more than any single dimension, and clever spelling variations rarely defeat confusingly similar analysis when other dimensions remain close.
Relatedness of goods or services is typically the second-most-important DuPont factor after similarity of marks. Identical or highly similar marks can coexist in completely unrelated industries because consumers don’t confuse different product categories. Conversely, moderately similar marks can create confusion when the products are closely related.
“Polaris” can exist as a snowmobile brand and as an unrelated software product because the two product categories are far enough apart that consumers don’t cross the mental bridge between them. “Apple” for computers and “Apple” for music coexist within the same company now but previously coexisted separately through this same principle. The USPTO examining attorney evaluates relatedness based on how ordinary consumers perceive the products, not on technical industry classifications.[3]
Apply DuPont factor analysis in four steps: identify the closest existing marks, score each mark against the similarity dimensions, evaluate goods/services relatedness, and assess the cumulative weight. The analysis is a structured tool that predicts USPTO examination outcomes.
The analysis produces a prediction of USPTO examination outcomes. Borderline cases benefit from a trademark attorney’s review, because the factor weighting involves legal judgment that DIY analysis can get wrong. But for clear cases — either obviously fine or obviously conflicted — the DIY DuPont analysis is a reliable self-assessment tool before filing.
The DuPont factors look mathematical — thirteen factors, each scorable, producing a weighted conclusion. In practice, the analysis is judgment-heavy. Different USPTO examining attorneys weight the factors differently. Different federal courts emphasize different factors in litigation. The same facts can produce different outcomes depending on who’s applying the framework.
This judgment-heavy reality has implications for clearance. A DIY DuPont analysis produces a prediction, not a certainty. A professional trademark attorney’s analysis produces a more reliable prediction because it’s informed by experience with how examining attorneys actually apply the factors in practice. But even the best analysis cannot guarantee the outcome of a specific USPTO examination.
This is where Responsible Asset-Building treats the DuPont analysis as a tool for informed decision-making, not a test with a definitive answer. Running the analysis surfaces the strongest and weakest aspects of your filing. Addressing the weaknesses — through mark modification, class selection, or filing strategy — improves the odds of success. An educated consumer uses the framework to understand the trade-offs clearly and make the best-informed decision available.
There's no bright-line threshold. The USPTO examines the totality of the DuPont factors, and marks can be refused for surprisingly modest similarity when the goods and services overlap strongly. Identical marks in identical classes are near-certain refusals. Similar marks in closely related classes are likely refusals. Beyond that, the analysis depends heavily on specifics.
Usually not much. Courts and the USPTO treat phonetic similarity and commercial impression as more important than exact spelling. 'Nikee' for athletic shoes would be confusingly similar to 'Nike' despite the spelling difference, because it sounds the same and covers the same product category. Creative respellings rarely defeat confusingly similar analysis.
Yes, sometimes. If your target customers are sophisticated purchasers who carefully evaluate brands (B2B buyers, medical professionals, enterprise software purchasers), the USPTO may weigh the 'conditions of purchase' factor in your favor. Sophisticated buyers are assumed to exercise greater care and are less likely to be confused by similar marks. This factor rarely overcomes strong similarity plus relatedness, but it can tip borderline cases.
Unrelated products generally weigh heavily against confusingly similar analysis. The DuPont relatedness factor asks whether customers would expect the same company to sell both categories. 'Delta' can exist for airlines and 'Delta' for plumbing fixtures because no one assumes an airline also makes faucets. The unrelated goods analysis allows many similar marks to coexist across industries.
When such evidence exists, yes. Actual confusion is one of the DuPont factors, and demonstrated instances of consumer mistake can be powerful evidence. In practice, most applications lack actual-confusion evidence because the mark is new and hasn't generated enough commerce to test the question. The factor matters most in litigation between established users rather than in USPTO examination of new applications.
The examining attorney issues an office action refusing the application on likelihood of confusion grounds under 15 U.S.C. §1052(d). You have six months to respond. Possible responses include arguing that the marks aren't actually confusingly similar, showing that the goods or trade channels differ, amending the mark or class description, or abandoning the application and refiling with a modified mark.
Understand your brand, see what's worth protecting, and walk into any attorney conversation prepared. Enter your name and email once to unlock all three.