You're about to file your own trademark and you keep seeing the phrase "likelihood of confusion." Here's what it actually means, and what it doesn't. It is the central question in US trademark disputes: an examiner, the Board, and the courts all apply it, though it does not by itself decide whether anyone sues you. It shows up first as a letter from a USPTO examining attorney refusing your application, sometimes as a letter from another brand's lawyer, occasionally as an Amazon or Etsy listing pulled without much explanation. Every one of those traces back to the same question.
Most founders never learn how that question is answered until it is being answered against them. So let's take it apart: where the test comes from, what the 13 DuPont factors mean, why a clean registration is not a clean bill of health, and how to pressure-test your own name first.
What "Likelihood of Confusion" Actually Means #
The Lanham Act is the federal statute that governs U.S. trademarks. It uses one phrase as the bar for two very different things: refusing a registration under Section 2(d), 15 U.S.C. §1052(d) (statute), and finding infringement under Section 32, 15 U.S.C. §1114 (statute).
The question is whether an ordinary consumer, running into both marks in the real world, would likely be confused about who is behind the goods or services, or about sponsorship or affiliation between the two brands.
A few things carry most of the weight. It turns on likelihood, not actual confusion: nobody has to have actually been confused, only that confusion is probable. It is judged from the viewpoint of an ordinary purchaser, whose level of care varies with the goods, not a specialist who studies logos for a living. And "confusion" is broad, covering a buyer who thinks your product is related to or licensed by the other brand, not only one who thinks the two are identical.
That last point is why a mark can run into trouble even when the earlier owner doesn't sell exactly the same product: the question is whether buyers would assume a connection. In court, the Lanham Act covers use that is likely to cause confusion "as to the affiliation, connection, or association" of one person with another, or "as to the origin, sponsorship, or approval" of their goods (15 U.S.C. §1125(a)).
Where the Test Comes From: DuPont and the Circuit Split #
The framework the USPTO uses comes from a 1973 decision of the U.S. Court of Customs and Patent Appeals, the predecessor to today's Federal Circuit: In re E. I. du Pont de Nemours & Co., 476 F.2d 1357 (C.C.P.A. 1973). The USPTO's own examining manual sets out and applies these factors (TMEP §1207.01; full opinion via CourtListener).
That case laid out 13 factors to weigh when they are relevant. The court was explicit that there is no single litmus test. Every dispute weighs whichever factors actually matter on its facts. Some cases turn on a single factor.
Important: DuPont is not the universal court test. The DuPont factors govern the USPTO examining attorney and the Trademark Trial and Appeal Board (TTAB), whose decisions are reviewed by the Federal Circuit. A federal district court deciding an infringement suit instead applies its own circuit's likelihood-of-confusion factors. In the Ninth Circuit, for example, that is the eight-factor Sleekcraft test from AMF Inc. v. Sleekcraft Boats, 599 F.2d 341 (9th Cir. 1979) (opinion); other circuits have their own named tests (Polaroid in the Second Circuit, the Lapp factors in the Third, and so on). These lists overlap heavily with DuPont (mark similarity, relatedness of the goods, strength of the senior mark, actual confusion, trade channels) but they are not identical, and the weight a factor gets can shift a lot between the examiner's desk and a courtroom.
The 13 DuPont Factors, at a Glance #
All thirteen come from the du Pont opinion and are applied by the USPTO (TMEP §1207.01). You do not weigh them evenly, and the USPTO treats two of them as key considerations.
| # | Factor | The question it asks |
|---|---|---|
| 1 ★ | Similarity of the marks | How alike in sight, sound, and meaning? |
| 2 ★ | Similarity of goods/services | Would a buyer expect a common source? |
| 3 | Trade channels | Do both reach buyers the same way? |
| 4 | Conditions of purchase | A snap buy, or a careful decision? |
| 5 | Fame of the prior mark | Is the senior mark famous enough for a wider moat? |
| 6 | Crowded field | Are lookalikes already common for similar goods? |
| 7 | Actual confusion | Has a real buyer actually been confused? |
| 8 | Length of concurrent use | Years side by side with no trouble? |
| 9 | Variety of goods | Does the senior mark span many categories? |
| 10 | Market interface | Any consent or coexistence deal in place? |
| 11 | Right to exclude | How strong is the applicant's own claim? |
| 12 | Extent of potential confusion | How many buyers would be affected? |
| 13 | Any other fact | Anything else probative. |
The two marked with ★ are the ones the USPTO singles out: "the following two factors are key considerations in any likelihood of confusion determination" (TMEP §1207.01). The other factors still count, and as the same section notes, "any single factor may control a particular case." But it's why, when you compare your name to an existing mark, it pays to start with those first two questions.
Why USPTO Approval Doesn't Bind a Federal Court #
This is the most expensive misunderstanding in the whole field. Three separate forums apply a likelihood-of-confusion analysis, and they do not always land in the same place: partly because, as above, a district court uses its own circuit's factors rather than DuPont.
- USPTO examining attorney. Pre-registration, just you and the examiner. The 13 DuPont factors, but on limited evidence.
- Trademark Trial and Appeal Board (TTAB). Post-publication opposition or cancellation, and appeals from examiners' refusals. Both sides brief it, the evidence is fuller, the DuPont factors still apply, but there is no jury and no damages.
- Federal district court. Infringement suits with full discovery, injunctions, and money on the table: decided under the forum circuit's own likelihood-of-confusion factors.
A mark can register cleanly and still lose an infringement suit later. A court sees things the examiner never did: real evidence of confused customers, actual marketing, how the goods move through the market. Common law rights built purely through use can beat a federal registration in court, and a court can weigh evidence of a mark's strength in the marketplace that an examiner never saw.
A clean registration lowers your risk and strengthens your position. It is not a guarantee no one will sue, and it is not legal clearance to launch.
Where Confusion Actually Bites #
Likelihood of confusion reaches your business in three ways. For a self-filer they differ a lot in how often they hit and how much they cost, and the dollar figures below are illustrative, not typical; real costs and frequency vary case by case.
The one you'll most likely meet: a Section 2(d) office action #
A USPTO examining attorney refuses your application because they found a prior mark they consider confusingly similar. The USPTO calls likelihood of confusion "the most common reason for refusing registration" (USPTO), so if you file, plan for the possibility.
The cost is real but survivable. Your filing fee, several hundred dollars per class, is non-refundable, and you wait months to respond. Hiring an attorney to write the response adds to the cost. A note on doing it yourself: an applicant whose domicile is outside the United States must be represented by a U.S.-licensed attorney at the USPTO (USPTO: do I need an attorney?). This is the cheapest of the three outcomes, and the one a good pre-filing search is built to reduce.
The tail risk: a TTAB opposition or cancellation #
Your mark publishes for opposition, or it is already registered, and a third party files to block or cancel it at the TTAB. This is less common than an office action. When it lands, defending it is a litigated proceeding with its own rules and costs, and it can take a long time, with your brand in limbo meanwhile. Figures vary widely by case.
The uncommon, high-stakes one: a federal lawsuit #
An existing owner sues you for infringement under the Lanham Act. The plaintiff doesn't even need a registration, since common law rights can be enough. Defense costs can run high and a losing defendant's potential exposure can include the defendant's profits, the plaintiff's damages, and, in some cases, an enhancement of the award. Each of those has its own conditions and limits under 15 U.S.C. §1117: for example, any increase in damages "shall constitute compensation and not a penalty," and recovery of profits is subject to principles of equity. None of it is automatic.
The rough shape of the list is that the most common outcome is the cheap one and the costly outcomes are less frequent, but every branding choice you make, the name and the design, is a quiet bet on which side of the confusion line you land.
How to Self-Test Your Own Mark #
Before you spend money on filing or design, run this quick check against every conflicting mark your search turns up. It's a first screen, not a legal opinion.
- Sight. Put both marks side by side and squint. Could they pass for the same source?
- Sound. Say both out loud at normal speed. In a noisy room, could someone hear one as the other? The USPTO does exactly this: it refused registration of XCEED for agricultural seed as confusingly similar to the registered X-SEED, and the Federal Circuit affirmed, noting there is no "correct" pronunciation of a coined mark to hide behind (In re Viterra Inc., 671 F.3d 1358 (Fed. Cir. 2012)).
- Meaning. What does each name evoke? Do they point at the same idea?
- Goods. Are the goods or services related: would a buyer expect one company behind both? (A Nice class number helps you organize a search; it's the relatedness of the goods, not the class number itself, that drives a conflict.)
- Fame. Is the prior mark famous beyond its registered goods? If so, its protection widens.
- Channel. Where will buyers actually encounter your mark, and does the other brand live there too?
If several of these point toward similarity, treat it as a real confusion risk: rework the name or get a professional clearance opinion before you file. There's no magic number of "yes" answers that passes or fails: the factors are weighed together, not counted.
Red Flags: When to Stop and Reconsider #
- A prior mark that sounds phonetically identical to yours for the same or related goods.
- A famous mark using similar wording, even for very different goods.
- Similar wording on a product sold through the exact channels you're targeting.
- A senior owner with a documented habit of opposing similar filings, which you can check in TTAB records.
- Signs of actual confusion already, like customers asking whether the two brands are connected.
- A conflicting mark used across many categories, since broad use can mean broad protection.
Frequently Asked Questions #
What's the difference between likelihood of confusion at the USPTO vs. in court? #
Closely related tests, different evidence and different stakes. The USPTO and the TTAB apply the 13 DuPont factors, mostly on the marks and goods as written. A federal court applies its own circuit's multifactor test (such as the Ninth Circuit's Sleekcraft factors), with full discovery, real marketing materials, and actual-confusion data. The factor lists overlap heavily, but they aren't identical, and a mark can clear the USPTO and still lose in court.
Is "likelihood" the same as "more likely than not"? #
No, and it's worth separating two different things. The substantive standard is that confusion must be probable, not merely possible; actual, proven confusion is not required. That's different from the burden of proof: in a civil infringement suit, the plaintiff generally must prove the elements, including likelihood of confusion, by a preponderance of the evidence. The Ninth Circuit's model jury instructions are one example: the plaintiff "has the burden of proving each of the following elements by a preponderance of the evidence," including use "in a manner that is likely to cause confusion," and a preponderance means "more probably true than not true" (Ninth Circuit Manual of Model Civil Jury Instructions, Instructions 1.6 and 15.6).
Does it matter if I didn't know about the other mark? #
Lack of knowledge is not a complete defense to infringement. Intent may be relevant under the applicable circuit's confusion factors (the Ninth Circuit's Sleekcraft test, for example, includes the defendant's intent in selecting the mark; AMF Inc. v. Sleekcraft Boats) and may also affect remedies (15 U.S.C. §1117), but a good-faith adopter can still infringe.
Can two identical marks coexist if they're in different industries? #
Often yes, sometimes no. It hinges on whether the goods or services are related enough that buyers would expect one company behind both, not on whether they happen to sit in different Nice classes. Relatedness also needs evidence. In In re Coors Brewing Co., the Federal Circuit agreed that the two BLUE MOON marks were similar but reversed and remanded the Board's decision refusing BLUE MOON for beer over a registration for restaurant services: applying the "something more" standard, it held that evidence that restaurants serve beer was not enough on its own to show the two were related (343 F.3d 1340 (Fed. Cir. 2003)). A strong or famous mark may receive broader protection, but the scope depends on the applicable claim and the evidence; a dilution claim has its own conditions, including wide recognition by the general consuming public of the United States (15 U.S.C. §1125(c)).
Who decides likelihood of confusion? #
At the USPTO, an examining attorney during review. At the TTAB, a panel of three administrative trademark judges. In federal court, a judge on motions, and at trial a judge or, if one is demanded, a jury, applying the local circuit's standard, which overlaps heavily with DuPont but is its own test.
If I get a cease & desist letter citing likelihood of confusion, am I cooked? #
Not necessarily. A cease & desist letter is the sender's opinion that confusion exists, and plenty of them overreach. Your response depends on the strength of the senior mark, the similarity of the marks and goods, and evidence that you adopted the name in good faith (relevant to intent and remedies, though not a complete defense). Don't ignore it, don't panic-rebrand, and get a written legal opinion.
How can I reduce my risk of a likelihood-of-confusion problem? #
Search before you file, in order: the USPTO register for your name and its phonetic variants; a visual logo search if you have a design element; a common-law sweep across Google and social platforms; and an attorney clearance opinion if real money is on the line. A clean search lowers your risk. It is not a clearance, a guarantee, or insurance.
Related reading #
- How to Check If a Brand Name Is Trademarked in the US, a pre-filing wordmark search guide.
- How to Check If a Logo Is Trademarked: A US Search Guide, a visual mark search guide.
- Brand Launch Trademark Checklist: 12 Pre-Launch Checks, for the full pre-launch review.
So what do you actually do? #
"Likelihood of confusion" sounds vague, but in practice the analysis keeps returning to two questions: are the marks similar, and are the goods or services similar? Lean the wrong way on those and you're exposed across all three forums, each applying its own version of the test at a different price. Lean the right way (names differing in sight, sound, and meaning, and goods that a buyer wouldn't trace to one source) and your risk drops sharply.
The search and the self-test exist to surface conflicts while they're still cheap to fix, before they turn into an office action or something worse. Run the test on yourself before an examiner runs it for you.
This is general information, not legal advice. Confusion analysis is fact-specific, and the outcome of any case depends on its facts. For questions about a mark, a refusal, a cease & desist letter, or a lawsuit, talk to a qualified trademark attorney.
External resources #
- USPTO Trademark Search, to search the federal register.
- USPTO: Responding to Office Actions, on what to do after a 2(d) refusal.
- Trademark Trial and Appeal Board, on opposition and cancellation procedures.
- USPTO TMEP §1207, examining-attorney guidance on likelihood of confusion.
