Likelihood of confusion is the probability that an appreciable number of ordinary consumers will be mistaken or misled about the source, sponsorship, or affiliation of goods or services. It is not a side issue in trademark law. It is the central question in infringement suits under 15 U.S.C. § 1114 and § 1125(a), and the reason most USPTO registration refusals issue under 15 U.S.C. § 1052(d).
Courts answer it with a multifactor test. The factors are not elements, and they are not a scorecard: no plaintiff must win a majority of them, and no single factor decides a case by itself. They are a structured way to reason toward one prediction about what the buying public will think.
Every federal circuit has its own list. Three matter most.
Others carry their own names. The Third Circuit uses the Lapp factors, the Fifth Circuit calls its list the “digits of confusion,” and the Sixth, Seventh, and Eleventh Circuits each run their own variants. The split is real but shallower than the naming suggests. Strip the labels and the same handful of questions appear in every list. The genuine differences are in emphasis, in whether a factor is weighted or merely counted, and in how each circuit handles the harder confusion theories described below.
Similarity of the marks. Judged in their entireties for sight, sound, and commercial impression, not by dissecting them side by side. Real buyers do not perform a letter-by-letter comparison; they carry an imperfect memory of the senior mark into a store. This is why Seycos can conflict with Seiko and why adding a house name to an otherwise-copied term rarely cures anything.
Relatedness of the goods or services. The products need not compete. They need only be related enough that consumers might reasonably assume a common source or a licensing relationship. Related is a factual question about market expectations, not a taxonomy exercise.
Trade channels and purchaser overlap. Two identical marks can lawfully coexist if one sells industrial valves to refineries and the other sells candles at farmers markets. The internet has compressed this factor’s force, since almost everything now moves through overlapping online channels.
Strength of the senior mark. Strength has two dimensions: conceptual (where the mark sits on the fanciful-to-generic spectrum) and commercial (how much recognition it has actually earned). Exxon is fanciful and famous, so it gets a wide berth. A merely descriptive mark with modest sales gets very little.
Actual confusion. The most persuasive evidence there is, and the least often available. Misdirected calls, wrong-vendor invoices, and survey results all count. Its absence proves less than defendants argue, because confused consumers seldom announce themselves.
Intent. A junior user who deliberately picked a close mark hands the court an inference that it succeeded at what it set out to do. Good faith, by contrast, is close to worthless as a defense: a defendant can innocently confuse the public just as thoroughly as a bad actor can.
Degree of purchaser care. Buyers of $4 shampoo are inattentive. Buyers of $2 million imaging systems are not. Higher price and higher sophistication both cut against confusion, though courts note that even expert buyers can be confused when the marks are close enough.
The doctrine reaches beyond the moment a customer hands over money.
Initial-interest confusion occurs when a junior user’s mark draws a consumer’s attention through misdirection, even though the consumer understands who they are dealing with before buying. The Ninth Circuit built the theory in Brookfield Communications v. West Coast Entertainment, 174 F.3d 1036 (9th Cir. 1999), using an analogy about a misleading billboard luring drivers off the highway. It has been sharply criticized since, and the Ninth Circuit itself narrowed it in Multi Time Machine v. Amazon.com, 804 F.3d 930 (9th Cir. 2015), holding that a clearly labeled search results page confuses no reasonably prudent shopper.
Post-sale confusion looks past the buyer entirely to observers in the world. The knockoff purchaser knows exactly what they bought. The problem is everyone who sees the bag, the watch, or the car and forms a false impression about the real brand. Ferrari S.p.A. v. Roberts, 944 F.2d 1235 (6th Cir. 1991), applied this to replica car bodies and remains the standard illustration.
Reverse confusion inverts the usual harm. A small senior user is not exploited by a free rider; it is buried by one. A large junior user saturates the market until consumers assume the original owner is the imitator. Big O Tire Dealers v. Goodyear Tire & Rubber Co., 561 F.2d 1365 (10th Cir. 1977), established the theory. Applying the ordinary factors requires care, because the strength factor has to be read against the junior user’s advertising muscle rather than the senior user’s fame.
The same words describe two different proceedings, and conflating them causes real errors.
At the USPTO, an examining attorney compares your application to registered marks and prior applications on paper. The analysis is bounded by the identification of goods in the filing, not by what you sell in the real world. Broad language in your description can generate a refusal over a product you never intend to launch.
In federal court, likelihood of confusion is a fact question decided on real marketplace evidence: actual advertising, real channels, survey data, and testimony. A TTAB decision is not binding on an infringement court, although the Supreme Court held in B&B Hardware v. Hargis Industries, 575 U.S. 138 (2015), that issue preclusion can apply when the Board decided the same issue the court faces.
The Supreme Court returned to the doctrine’s primacy in Jack Daniel’s Properties v. VIP Products, 599 U.S. 140 (2023), holding that when a defendant uses a challenged term as a source identifier for its own goods, First Amendment screening tests give way and the ordinary likelihood of confusion analysis applies. Parody does not exit the test. It becomes an argument within it.
Two lessons follow from how courts actually decide these cases. First, the factors are directional, not arithmetic. Counting them up produces the wrong answer more often than not, because similarity and relatedness routinely swamp the rest. Second, most disputes never reach a factor analysis at all. They resolve on a demand letter, a coexistence agreement, or a narrowing of trade channels, precisely because both sides can read the same factors and forecast the outcome.
For the practical mechanics of responding when another business is using your name, see the linked guide below. This page defines the standard those disputes are measured against.
What are the likelihood of confusion factors? They are the multifactor tests courts and the USPTO use to decide whether two marks are too close. The USPTO and Federal Circuit use the 13 DuPont factors. The Ninth Circuit uses the 8 Sleekcraft factors. The Second Circuit uses the 8 Polaroid factors. The names and counts differ, but every version asks about mark similarity, relatedness of the goods, trade channels, strength of the senior mark, actual confusion, the defendant’s intent, and how carefully buyers shop.
Do you need proof of actual confusion to win a trademark case? No. The statutory standard is likelihood of confusion, not actual confusion, so a plaintiff can win with zero confused customers in the record. Evidence of actual confusion is powerful when it exists because it proves the point directly, but courts treat its absence as weak evidence at best, since confused consumers rarely report themselves and confusion is hard to document.
Which likelihood of confusion factors matter most? In practice the similarity of the marks and the relatedness of the goods or services do most of the work. Courts repeatedly say the factors are not a scorecard and no factor is dispositive, but a case in which the marks look and sound alike and the products sit on the same shelf will usually come out for the plaintiff even if several other factors point the other way.
What is reverse confusion in trademark law? Reverse confusion happens when a large junior user floods the market with a mark similar to a small senior user’s, so consumers come to believe the original owner is the copycat. The senior user’s identity is swamped rather than exploited. The Tenth Circuit recognized it in Big O Tire Dealers v. Goodyear in 1977, and most circuits now apply the same factors with the strength factor read against the junior user’s advertising power.
Going further: What to do when someone is using your business name .
This page is general legal information, not legal advice, and it does not create an attorney-client relationship.
The Ninth Circuit revived a small developer's SmartSync claim, holding a jury could find reverse confusion when a giant junior user swamps a senior mark.
The Second Circuit held a strong arbitrary mark on related goods made confusion likely, ordering an injunction against VIRGIN WIRELESS phone stores.
The Ninth Circuit reversed itself after Jack Daniel's, holding that Rogers cannot shield a name used as a source identifier and sending the dispute back for an ordinary likelihood-of-confusion analysis.
When two companies used the Lapp name on non-competing electrical goods, the Third Circuit produced the ten-factor framework courts still use to gauge trademark confusion.
On rehearing, a divided Ninth Circuit held that a clearly labeled list of substitute products is not trademark infringement, narrowing initial-interest confusion for the era of online search.
A unanimous Supreme Court held that a trademark defendant asserting descriptive fair use bears no burden to disprove likelihood of confusion, which remains the plaintiff's to prove.
The Second Circuit held that Google's sale and recommendation of trademarks as AdWords keyword triggers is a 'use in commerce' under the Lanham Act, reviving the suit.
Polaroid lost its 1961 infringement suit to laches, but Judge Friendly's catalogue of variables for gauging confusion became the most cited multifactor test in American trademark law.
A 1979 dispute over Slickcraft and Sleekcraft boats gave the Ninth Circuit its enduring eight-factor framework for likelihood of confusion, and a template courts still run today.
