One of the most common questions I get from clients who've received an opinion letter indiciating a likelihood of confusion risk is some version of: "These marks don't look that similar to me. Why will the USPTO think there's a problem?" It's a fair question, and the answer requires understanding of how courts and the USPTO actually analyze trademark similarity. The rules are more nuanced than most people expect, and they don't always match intuition.
What follows is a plain-English walkthrough of the key legal principles governing the similarity of marks in a likelihood of confusion analysis, with real cases attached so you can see how the rules play out in practice.
The Test Is Not a Side-by-Side Comparison
This is the single most important thing to understand, and it's the one that surprises people most. When the USPTO or a court asks whether two marks are confusingly similar, they are not imagining a consumer standing at a counter with both marks in front of them at the same time. They are asking whether someone who has seen one mark and later encounters the other might mistakenly assume a connection between the two companies.
So, the proper test is whether the marks are sufficiently similar in terms of their commercial impression such that persons who encounter the marks would be likely to assume a connection between the parties. The standard accounts for the fact that memory is imperfect. Consumers don't carry mental photographs of every trademark they've seen. They carry impressions.
This matters enormously in practice. Two marks that look distinct when placed next to each other may still be confusingly similar when evaluated the way actual consumers experience them, which is sequentially and with imperfect recall. For example, in finding that the appearance of "TAKETEN" is not meaningfully distinguished from "TAKE 10!", the Federal Circuit in In re St. Helena Hospital (2014) noted that marks must be considered in light of the fallibility of memory, not on the basis of side-by-side comparison.
Marks Must Be Considered as a Whole
You cannot take a mark apart, focus on one piece, and decide the rest doesn't matter. Courts have consistently held that a mark must be evaluated in its entirety when assessing likelihood of confusion.
That said, considering a mark as a whole does not mean every element gets equal weight. Which brings us to the next principle.
Some Parts of a Mark Can Carry More Weight Than Others
Courts and the USPTO regularly give more or less emphasis to particular elements of a mark, as long as the final conclusion is based on the mark considered in its entirety. This is not a contradiction of the anti-dissection rule. It's a recognition that some parts of a mark do more of the source-identifying work than others.
The Federal Circuit confirmed in In re National Data Corp. (1985), a foundational case on this point, that there is nothing improper in giving more or less weight to a particular feature of a mark for rational reasons, provided the ultimate conclusion rests on the marks in their entireties. That case has been cited approvingly numerous times since.
The most commonly accepted reason for giving less weight to a portion of a mark is that it is descriptive or generic with respect to the relevant goods or services. If part of your mark simply describes what you sell, courts will focus on the distinctive parts when comparing your mark to a competitor's. For example, the Federal Court in Cunningham v. Laser Golf Corp. (2000) accepted that little weight could be given to "SWING" in a comparison of LASERSWING and LASER for golf clubs on the basis that it is descriptive.
Disclaimed Words Still Count
Here's one that catches a lot of applicants off guard. When the USPTO requires you to disclaim a generic or descriptive word in your mark, that word doesn't disappear from the confusion analysis. It's still part of your mark as far as consumers are concerned, because consumers have no idea what is disclaimed.
For example, in Shen Manufacturing Co. v. Ritz Hotel, Ltd. (2004), the Federal Circuit affirmed that it was proper for the TTAB to have considered the disclaimed words PARIS and HOTEL in determining that the marks RITZ PARIS RITZ HOTEL and RITZ were sufficiently different to avoid a likelihood of confusion finding.
While a disclaimed term may be given little weight, it cannot be ignored entirely.
In Word-and-Design Marks, the Words Usually Dominate
When a mark combines text and a visual design element, the verbal portion is generally considered the more important source identifier. The reasoning is practical: people are more likely to ask for a product by name than to describe its logo. A customer who wants to reorder something they liked is going to say the name, not sketch the design on a napkin.
There is an important caveat. As the Federal Circuit noted in In re Electrolyte Laboratories, Inc. (1991), there is no general rule as to whether letters or design will dominate in composite marks, and that dominance of one over the other is not automatically dispositive. The words-dominate principle is a default, not an absolute. If the design element is particularly striking and the words are weak or descriptive, the analysis may go differently.
When the Goods Are Nearly Identical, Less Similarity Is Needed
The similarity of marks does not exist in a vacuum. It interacts with the similarity of the goods or services at issue. The closer the goods, the less similar the marks need to be to create a likelihood of confusion.
If you're selling nearly identical products to the same consumers through the same channels, even a modest resemblance between marks can be enough to create a problem.
Famous Marks Get Extra Protection
The more famous a mark, the less similarity a junior mark needs to cause confusion. The Federal Circuit framed it as an inverse relationship in Century 21 Real Estate Corp. v. Century Life of America (1992): the Lanham Act's tolerance for similarity between competing marks varies inversely with the fame of the prior mark. The TTAB applied this in Harry Winston, Inc. v. Bruce Winston Gem Corp. (2014), where the fame of the Harry Winston brand was a significant factor in the analysis.
The practical implication is straightforward. If you choose a name that is even modestly similar to a well-known brand, you face a harder road than you would if the prior mark were obscure. Fame expands the zone of protection.
Similarity Is a Spectrum, Not a Binary Decision
Marks are not simply confusing or not confusing. Similarity exists on a continuum, and where a mark falls on that continuum affects how the other du Pont factors are weighed. The Federal Circuit confirmed this in In re St. Helena Hospital (2014) and the TTAB applied the same framing in North Face Apparel Corp. v. Sanyang Industry Co. (2015).
This matters because a mark that is highly similar to an existing mark may overcome other factors that might otherwise weigh against confusion. Conversely, marks that are only modestly similar may require stronger evidence of relatedness in other areas before confusion is found.
Foreign Words Get Translated
Under the doctrine of foreign equivalents, foreign words from common languages are translated into English before being compared to English word marks. If your mark is the Spanish word for a common noun, the USPTO will consider whether the English translation of that noun is confusingly similar to existing marks.
The Federal Circuit applied this doctrine in Palm Bay Imports, Inc. v. Veuve Clicquot Ponsardin (2005), one of the leading cases on the subject.
The doctrine applies to common modern languages, not obscure or dead ones. And it is not applied mechanically. If consumers in the relevant market are unlikely to recognize or translate the foreign word, the doctrine may not apply. But for Spanish, French, Italian, and other widely spoken languages, applicants should assume translation will occur.
There Is No Correct Pronunciation of a Trademark
When comparing marks for phonetic similarity, the analysis acknowledges that there is no single correct way to pronounce most trademarks. Different consumers will say the same mark differently, and the likelihood of confusion analysis accounts for the full range of reasonable pronunciations.
The Federal Circuit noted this in In re St. Helena Hospital (2014), and the TTAB has applied it in numerous cases including Edwards Lifesciences Corp. v. VigiLanz Corp. (2010). The principle goes back at least to In re Belgrade Shoe Co. (1969).
There is a sensible exception: when a mark is a recognized word, even in a foreign language, the no-correct-pronunciation principle does not apply. The TTAB drew this line in INTS It Is Not The Same, GmbH v. Disidual Clothing, LLC (2017). If consumers know how to say a word, the analysis focuses on that pronunciation.
Pure Design Marks Are Compared Visually
When neither mark contains words and neither can be spoken, the similarity analysis is necessarily visual. Courts compare the overall appearance, the impression created by the design elements, and whether the two marks would be mistaken for each other by a consumer who has seen one and later encounters the other.
An example of the TTAB applying this principle can be found in In re Medline Industries (2020). For businesses that rely primarily on logos or visual marks rather than words, this means the design itself carries the full weight of the confusion analysis.
Identical Marks on Related Goods Are Usually Dispositive
When two word marks are identical, and neither is suggestive nor descriptive of the relevant goods, the similarity factor weighs heavily in favor of confusion. For example, see In re i.am.symbolic, llc (2017), where the Federal Court affirms this plainly.
This is not surprising. If two companies use the exact same arbitrary word as their brand name in related markets, confusion is almost inevitable. The strength of the marks and the relatedness of the goods may affect how the other factors shake out, but starting with identical arbitrary marks is a very difficult position to defend.
And Sometimes One Factor Is Enough
While the likelihood of confusion analysis involves balancing multiple factors, a single factor can be dispositive when it is strong enough. The most common example is dissimilarity. If the marks are sufficiently different, no amount of relatedness in the goods or other overlapping factors will establish likelihood of confusion.
A finding that the marks are clearly dissimilar can end the analysis entirely, regardless of how the other factors might weigh.
What This Means in Practice
If you're evaluating a new brand name, choosing between marks, or responding to an Office Action citing likelihood of confusion, understanding these principles changes how you approach the problem. The question is never simply whether two marks look alike. It's whether a consumer with imperfect memory, encountering the marks at different times, would be likely to assume a connection. That's a more forgiving standard in some ways and a more demanding one in others.
If you're dealing with a likelihood of confusion refusal or trying to assess whether a name you want is available, I'm happy to take a look. Just use one of the many buttons on the site to schedule a free consultation or email me directly at jared@fivedogslaw.com.
.svg)