Likelihood of Confusion

The standard does not require that consumers actually be confused, only that confusion is reasonably probable. It applies to marks that are similar in appearance, sound, or meaning and that are used on related goods or services.

How the likelihood of confusion works

The U.S. Patent and Trademark Office (USPTO) and federal courts evaluate the likelihood of confusion using a set of factors established in In re E.I. du Pont de Nemours & Co. (1973). These are commonly called the DuPont factors. Not all factors carry equal weight in every case, and examiners or courts apply them based on the specific circumstances.

The most heavily weighted factors are:

• Similarity of the marks, how alike the marks are in appearance, sound, and meaning
• Relatedness of the goods or services, whether the products or services are sold in the same market or to the same consumers
• Strength of the existing mark, how distinctive or well-known the senior mark is
• Channels of trade, whether the goods or services are sold through the same outlets or to the same customer base
• Conditions of purchase, whether buyers exercise care or make impulse decisions

The USPTO examiner weighs these factors collectively. No single factor is automatically decisive, though similarity of the marks and relatedness of goods are typically the most influential.

Why the likelihood of confusion matters

For anyone registering a trademark, likelihood of confusion is the most common type of substantive refusal issued by the USPTO.

If a pending mark is deemed too similar to an already-registered mark in the same or related product category, the application will receive an office action citing Section 2(d) of the Lanham Act. Approximately 1 in 5 trademark applications receive this type of refusal.

Beyond registration, the likelihood of confusion is the central question in trademark infringement litigation, where plaintiffs succeed in approximately 55% of cases, and verdicts can reach $190 million or more.

A trademark owner who believes another party is using a confusingly similar mark can pursue legal action based on this standard, even without proving actual consumer confusion. About 3,000 trademark lawsuits are filed in U.S. federal courts each year.

Understanding this standard is essential before investing in a brand name, logo, or slogan. A thorough trademark search conducted before filing can identify potential conflicts early and reduce the risk of a refused application or infringement claim—particularly as applicants from Mainland China file over 20% of all U.S. trademark applications, increasing register density, given that nearly half of all U.S. trademark applications are rejected.

Common examples of the likelihood of confusion

The following scenarios illustrate how this standard applies in practice.

• Similar names in the same industry: A new software company applies to register "Appify" as a trademark. The USPTO found a registered mark "Appifi" for similar software products. The marks are phonetically close, and the goods are identical; the likelihood of confusion is high.
• Different industries, same mark: A clothing brand called "Falcon" and a restaurant called "Falcon" may coexist without confusion because the goods and services are unrelated, and consumers are unlikely to think they share a common source.
• Famous marks and dilution: A startup using "Gooogle" for any product category would likely face a likelihood-of-confusion challenge given the strength and recognition of the Google mark, even if the goods differ.
• Stylized logos vs. word marks: Two companies using the word "Crest", one for toothpaste, one for financial services, may not create confusion because the industries are distinct and consumers exercise care in financial decisions.

Key characteristics of the likelihood of confusion standard

The standard is consumer-centric. The relevant question is how an ordinary, reasonably prudent consumer would perceive the marks, not how a trademark attorney or the business owner would.

Confusion need not be certain or even likely in the majority of cases. Courts have found infringement where even a small percentage of consumers were likely to be confused, particularly when the goods or services are high-stakes or the marks are highly similar.

The standard also encompasses several types of confusion: point-of-sale confusion (at the time of purchase), post-sale confusion (after the transaction), and reverse confusion (in which a junior user's extensive use leads consumers to believe the senior user is the infringer).

Likelihood of confusion vs. trademark dilution

Likelihood of confusion and trademark dilution are distinct legal concepts, though both protect trademark owners. Likelihood of confusion focuses on consumer deception, the risk that buyers will mistake one brand for another. Dilution, by contrast, applies only to famous marks and addresses harm to the mark's distinctiveness or reputation, regardless of whether consumers are confused. A small business owner is far more likely to encounter the likelihood of confusion issues during the registration process than dilution claims, which are reserved for nationally recognized brands.

Considerations when evaluating potential conflicts

Before filing a trademark application, conducting a comprehensive trademark search, covering both the USPTO database and common law uses, is the most effective way to identify potential likelihood of confusion issues, especially as the USPTO received nearly 765,000 trademark applications in FY 2024 alone.

If the USPTO issues a Section 2(d) refusal based on likelihood of confusion, applicants can respond by arguing that the marks are sufficiently different, that the goods or services are distinct, or that the existing mark is weak due to widespread third-party use. These arguments require careful legal analysis and supporting evidence. The TTAB affirmance rate is approximately 90% for likelihood of confusion refusals on appeal.

Trademark applicants should also be aware that the likelihood of confusion is evaluated at the time of filing, not based on future plans to expand into different markets. Claiming that two marks operate in different industries today does not guarantee protection if the goods or services later overlap.

Related terms and next steps

Understanding the likelihood of confusion connects directly to several other foundational trademark concepts. A common-law trademark can create prior rights that affect a likelihood-of-confusion analysis, even without federal registration. The concept of use in commerce is also relevant, as trademark rights and the potential for confusion arise from actual use in the marketplace. Before choosing a business name, a business name availability search can surface potential conflicts at an early stage.

For those preparing to file, working with a trademark attorney to conduct a clearance search and evaluate the likelihood of confusion risks before submitting an application can significantly improve the chances of approval; less than 50% of self-filed applications make it past examination without a refusal.

FAQs about the likelihood of confusion

What is the difference between the likelihood of confusion and actual confusion?

Actual confusion refers to documented instances in which consumers were already misled, such as misdirected calls, emails, or purchases, whereas the likelihood of confusion is a predictive legal standard that asks whether confusion is reasonably probable, regardless of whether it has occurred yet. Courts can find infringement based on the likelihood of confusion alone, meaning a trademark owner need not wait for real-world harm before taking action.

Can two trademarks coexist if they are in different trademark classes?

Filing in different trademark classes does not automatically preclude a likelihood-of-confusion finding, because the USPTO evaluates the actual relatedness of the goods or services, not just their classification. Two marks registered in different classes can still conflict if consumers would reasonably expect the products to come from the same source or if the goods are sold through overlapping channels.

How does the strength of a mark affect the likelihood of confusion analysis?

A highly distinctive or commercially well-known mark receives broader protection, meaning a junior mark need not be as similar to trigger a likelihood of confusion finding. Conversely, a weak mark, one that is descriptive or widely used by third parties in the same industry, receives a narrower scope of protection, and the USPTO or a court may require a higher degree of similarity before finding a conflict.

Does registering a trademark in a different state protect against a likelihood of confusion refusal?

State trademark registration has no bearing on a federal likelihood-of-confusion analysis, and the USPTO does not consider state registrations when evaluating pending applications. Common law rights established through actual use in commerce, however, can affect the analysis; an unregistered mark used consistently in a specific geographic market may still create prior rights that a later applicant must contend with.

When does the USPTO evaluate the likelihood of confusion, at the time of filing or at the time of registration?

The USPTO evaluates the likelihood of confusion based on the record as it exists at the time of filing, not at the time a registration ultimately issues. This means that future plans to rebrand, narrow the product line, or enter new markets are not considered. The analysis turns on how the marks and goods or services compare at the moment the application is examined.

How is the likelihood of confusion evaluated when the marks look different but sound the same?

The similarity analysis covers appearance, sound, and meaning independently, and a strong phonetic resemblance alone can be sufficient to support a likelihood-of-confusion finding, even when the marks look distinct on paper. Because consumers frequently encounter and recall marks by hearing them spoken, in conversation, advertising, or over the phone, the USPTO treats sound as one of the most practically significant dimensions of similarity.