Table of Contents
Table of Contents
You have a name you love, you file your trademark application, and weeks later the USPTO sends back a refusal: likelihood of confusion. It is the single most common reason trademark applications are rejected — and it is also the standard courts use to decide infringement. Understanding how examiners and judges actually weigh confusion is the difference between a name that registers cleanly and one that triggers an expensive office action or lawsuit. This article explains the test, the factors behind it, and how to assess your own risk before you file.
What “likelihood of confusion” means
Under Section 2(d) of the Lanham Act, the USPTO refuses registration of a mark that so resembles an existing registered mark that it is likely to cause confusion, mistake, or deception. The legal question is not whether two marks are identical — it is whether an ordinary consumer might believe the goods or services come from the same source, or are affiliated, sponsored, or connected. Confusion about source is the heart of it.
This is exactly the risk a trademark search is designed to catch early. Running a comprehensive trademark clearance search before filing surfaces the conflicting marks that an examiner would later cite, so you can adjust course before spending money on an application.
The DuPont factors
The USPTO and the courts analyze likelihood of confusion using the DuPont factors, a set of thirteen criteria established in the landmark 1973 decision In re E.I. du Pont de Nemours & Co. and applied through USPTO examination guidance (TMEP §1207). Not every factor applies in every case, and they are not weighted equally. The most important are usually the first two:
- Similarity of the marks in appearance, sound, connotation, and overall commercial impression.
- Similarity of the goods or services — how closely related the products are, and whether consumers would expect them from one source.
- Trade channels — whether the goods move through the same stores, websites, or distribution paths.
- Conditions of sale — impulse purchases invite more confusion than expensive, carefully considered buys.
- Fame of the prior mark — famous marks receive a wider scope of protection.
- Number and nature of similar marks already in use on similar goods, which can weaken a mark’s exclusivity.
- Actual confusion — real-world evidence that consumers have already been confused.
- Length of concurrent use without confusion, plus the variety of goods on which a mark is used, the market interface between owners, and other established facts.
The Trademark Trial and Appeal Board (TTAB) weighs whichever factors the evidence supports; federal courts apply closely related tests (such as the Polaroid factors in the Second Circuit) to decide infringement.
How the USPTO applies the test
Examiners lean heavily on the first two factors and treat them on a sliding scale. The more similar the marks, the less related the goods need to be to support a refusal — and vice versa. Two near-identical marks can be refused even for somewhat different products, while only loosely similar marks may coexist if the goods are clearly unrelated. Marks are compared in their entireties rather than dissected, and the focus is on the general recollection an average consumer carries, not a side-by-side comparison.
Crucially, confusion is judged from the perspective of the relevant consumer for those goods. A highly technical product sold to sophisticated buyers tolerates more similarity than an inexpensive consumer item bought on impulse.
How to reduce your risk before filing
Most likelihood-of-confusion refusals are avoidable with diligence up front:
- Search before you commit. A knockout search followed by a full clearance search reveals conflicts the USPTO would cite.
- Choose a distinctive mark. Coined or arbitrary marks are far easier to clear than descriptive ones.
- Define goods and services carefully. Overly broad identifications collide with more prior marks; a precise description narrows the conflict.
- Monitor after registration. Confusion runs both ways — ongoing trademark monitoring catches later filings that encroach on your mark.
If a refusal does arrive, it is not the end of the road. A well-argued response can distinguish the marks or goods, and our overview of office action response strategies walks through the options.
All thirteen DuPont factors, in order
Most summaries stop at the first two factors, which is fine for a quick read but unhelpful when you are actually arguing a refusal. The court in In re E.I. du Pont de Nemours & Co., 476 F.2d 1357 (C.C.P.A. 1973), set out thirteen, and the Trademark Trial and Appeal Board will consider any of them that the record supports:
- The similarity or dissimilarity of the marks in their entireties as to appearance, sound, connotation and commercial impression.
- 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 similarity or dissimilarity of established, likely-to-continue trade channels.
- The conditions under which and buyers to whom sales are made — “impulse” versus careful, sophisticated purchasing.
- The fame of the prior mark, measured by sales, advertising and length of use.
- The number and nature of similar marks in use on similar goods.
- The nature and extent of any actual confusion.
- The length of time during, and conditions under which, there has been concurrent use without evidence of actual confusion.
- The variety of goods on which a mark is or is not used — house mark, family mark, or single product mark.
- The market interface between the applicant and the owner of the prior mark, including consent agreements, contractual restrictions on use, assignment of the mark with its goodwill, and laches or estoppel.
- The extent to which the applicant has a right to exclude others from using its mark on its goods.
- The extent of potential confusion — whether de minimis or substantial.
- Any other established fact probative of the effect of use.
Two things are worth noticing about that list. The first is how much of it is evidentiary rather than conceptual. Factors five through twelve are not arguments you can make from the face of the application; they require proof — sales figures, advertising spend, survey evidence, third-party registrations, declarations about concurrent use. An applicant who wants to rely on them has to build a record.
The second is factor ten, the market interface. It is the factor that most often changes outcomes in practice, because it is where a consent or coexistence agreement between the two owners enters the analysis. The Federal Circuit has long treated a properly drafted consent agreement between sophisticated parties as strong evidence, on the reasoning that the businesses themselves are best placed to judge whether their customers will be confused.
Factor thirteen is the catch-all, and it is not decorative. It is where arguments that do not fit the other twelve get in — the applicant’s bad faith in adopting the mark, for instance, or a documented pattern in how consumers actually encounter the two brands.
Responding to a Section 2(d) refusal
A likelihood of confusion refusal is an opening position, not a verdict. Roughly speaking there are four routes out, and they are not mutually exclusive.
Argue the factors. The most common response attacks factors one and two directly: the marks differ in connotation or commercial impression, or the goods are unrelated in the way the cited registration actually describes them. This is won on specifics. Generic assertions that “the marks look different” do nothing; a careful comparison of meaning, the presence of a distinguishing dominant element, or evidence that the terms are weak and heavily registered by third parties can carry the day. Factor six — the number and nature of similar marks already in use — is the standard vehicle for showing a crowded field.
Amend the identification of goods. Because the Board compares the goods as written rather than as sold, narrowing an overbroad identification can remove the overlap entirely. An application covering “computer software” that is really an industrial logistics tool may be refused against a consumer gaming mark; restricting it to the actual field can dissolve the conflict. This is often the fastest and cheapest fix.
Seek a consent or coexistence agreement. Where the other owner is reachable and the real-world overlap is small, a negotiated agreement that defines each party’s lane feeds directly into factor ten. It must do more than recite that the parties do not expect confusion; the Board looks for agreements that set out why confusion is unlikely and what the parties will do if it occurs.
Appeal or let it go. If the examining attorney maintains the refusal, the route is an appeal to the TTAB. That is a genuine decision point — the cost of appealing frequently exceeds the cost of rebranding early, and a name that needs a fight to register is rarely worth the fight. Our guide to office action response strategies sets out the procedural mechanics and deadlines.
Registration and infringement: one question, two records
The same statutory question — is confusion likely — governs both whether a mark registers and whether its use infringes. The evidence is not the same, and confusing the two is a common and expensive mistake.
The USPTO and the TTAB decide registrability on the documents. They compare the applied-for mark against the cited registration and compare the goods as they are identified in the filings, regardless of how either party actually sells today. A federal court deciding infringement looks at the marketplace: real packaging, real trade dress, real channels, real purchasers, and often survey evidence. Regional circuits apply their own variants of the test — the Polaroid factors in the Second Circuit, Sleekcraft in the Ninth — though the substance tracks DuPont closely.
The two forums are not sealed off from each other. In B&B Hardware, Inc. v. Hargis Industries, Inc., 575 U.S. 138 (2015), the Supreme Court held that a TTAB finding on likelihood of confusion can have issue-preclusive effect in later infringement litigation, provided the usages adjudicated by the Board are materially the same as those before the court. The practical lesson is that a TTAB proceeding is not a low-stakes administrative skirmish to be handled casually; what is decided there may bind you later.
Frequently asked questions
Do two marks have to be identical to cause confusion?
No. Marks that merely sound alike, look alike, or convey the same meaning can be refused if the goods or services are related. The test is overall commercial impression, not identical spelling.
What are the two most important DuPont factors?
The similarity of the marks and the similarity of the goods or services. The USPTO applies them on a sliding scale, so strength on one can offset weakness on the other.
Can two similar marks legally coexist?
Yes, when the goods, services, or trade channels are different enough that consumers are unlikely to assume a common source. Long concurrent use without actual confusion supports coexistence.
Does likelihood of confusion apply to infringement too?
Yes. The same core question — whether consumers are likely to be confused about source — governs both USPTO registration refusals and trademark infringement claims in court.
How can I tell if my mark is at risk before filing?
A professional clearance search assesses your mark against existing registrations and common-law uses using the same factors an examiner would, giving you a risk read before you invest in an application.
Clear your mark with confidence
Likelihood of confusion decides whether your trademark registers and whether it can be defended. The factors are predictable, which means the risk is assessable — if you look before you file. Contact PerspireIP for a comprehensive trademark clearance search and a clear read on your likelihood-of-confusion risk before you commit to a name.
Clear Your Mark Before You File
A knockout search you run yourself catches identical marks; it rarely catches the confusingly similar ones that cause refusals. PerspireIP runs comprehensive trademark clearance searches across USPTO, state and common-law sources, each delivered with a clearance summary and a plain-English risk read — $200 for full domestic clearance (USPTO, all 50 state registers and common-law sources), or from $120 for a USPTO-plus-domain screen. A WIPO/international search is $250; custom multi-country searches are quoted. A signed written likelihood-of-confusion opinion is not part of any search tier — tell us if you need one and we will quote it separately. Search fees are flat as at October 2026, with no government charge attached to a search.