Genericness, Genericide, and Acquired Distinctiveness: Where Trademark Protection Begins and Ends

Summary. Trademark protection has two borders, and most expensive brand disputes are fought on one of them. At the bottom border sits genericness: a term that names the product itself (aspirin, escalator, thermos) belongs to everyone, whether it was born generic or died into genericness through the owner's own neglect, and no registration, however old, can save it. At the upper border sits acquired distinctiveness, the showing under Section 2(f) of the Lanham Act that lets a merely descriptive term, a product shape, a color, or a sound cross into protection because the public has learned to read it as a brand. This article explains the primary significance test, the difference between genericness ab initio and genericide, the Supreme Court's rejection of a per se rule in USPTO v. Booking.com, the Federal Circuit's line from Marvin Ginn through Princeton Vanguard, Cordua, and Royal Crown, and the Teflon and Thermos survey formats and their critiques. It then walks through the Converse factors, the five-year presumption, why sales figures alone rarely prove secondary meaning, the special rules for product configuration and color, and the brand usage disciplines that keep a successful name from becoming a noun. Two worked examples (a software product whose name is becoming a verb, and a furniture maker seeking 2(f) protection for a chair design) show how the doctrine plays out in practice.

In 1921 Bayer owned the word "aspirin." By the end of that year it owned it only when speaking to pharmacists; to the people who actually swallowed the tablets, Judge Learned Hand found, the word had simply become the name of the drug. Bayer had sold ASPIRIN tablets to the public without ever giving it another word for acetylsalicylic acid, so the public called the drug aspirin. A century later a company called Booking.com persuaded the Supreme Court that a term built entirely from a generic word and a generic suffix could nonetheless be a brand, because consumers said so. Between those two cases lies the whole subject of this article: what makes a word (or a shape, or a color) belong to one seller, and what makes it belong to everyone.

The short answer

Trademark law protects signs that tell buyers where a product comes from. It refuses to protect signs that tell buyers what the product is. A term is generic when its primary significance to the relevant public is the name of a type of product or service rather than a particular source of it, and a generic term is unprotectable no matter how much money has been spent on it, no matter how long it has been registered, and no matter how unfair the result feels to the company that coined it. Some terms are generic from birth (ab initio); others are born as brands and die into the language through a process the cases call genericide, usually because the owner used the mark as a noun, let competitors and journalists do the same, and never gave the public an alternative word.

At the other end of the spectrum, a term that merely describes the product (CREAMY for yogurt, STEELBUILDING.COM for steel buildings sold online) is not protectable as a trademark either, but it is not doomed. If the owner can prove that the buying public has come to associate the term primarily with one source, the term has acquired secondary meaning, which the Lanham Act calls acquired distinctiveness and codifies in Section 2(f), 15 U.S.C. § 1052(f). That showing is made with evidence: length and exclusivity of use, advertising that teaches consumers to treat the term as a brand, sales, unsolicited press, copying by competitors, and, when the term is highly descriptive or the "mark" is a product shape, a consumer survey. Five years of substantially exclusive and continuous use may be accepted as prima facie evidence, but the USPTO and the courts are free to demand more, and for highly descriptive terms they routinely do.

The two doctrines are mirror images. Genericness asks whether the public has stopped (or never started) hearing the term as a brand. Acquired distinctiveness asks whether the public has started hearing a descriptive term as a brand. Both questions turn on the same thing, consumer perception, and both are proved with the same kinds of evidence. That is why one article covers both ends.

The distinctiveness spectrum and why its two ends matter

Judge Friendly's opinion in Abercrombie & Fitch Co. v. Hunting World, Inc., 537 F.2d 4 (2d Cir. 1976) sorted word marks into five categories of increasing strength: generic, descriptive, suggestive, arbitrary, and fanciful. The Supreme Court adopted the ladder in Two Pesos, Inc. v. Taco Cabana, Inc., 505 U.S. 763, 768 (1992), and every trademark examiner, Board panel, and district judge in the country still uses it. The top three categories are inherently distinctive: a suggestive mark (COPPERTONE for sunscreen) requires a small mental leap, an arbitrary mark (APPLE for computers) is an ordinary word used out of context, and a fanciful mark (KODAK) is invented. All three are protectable from the first sale.

The bottom two rungs are where the money is spent. A descriptive term immediately conveys a quality, feature, function, ingredient, or purpose of the goods, and it can become a mark only through acquired distinctiveness. A generic term is the common name of the goods themselves, or of a category or key aspect of them, and it can never become a mark. The statute enforces that structure in two places. Section 2(e)(1), 15 U.S.C. § 1052(e)(1), bars registration of a mark that is "merely descriptive" of the goods, and Section 2(f) then lets the applicant cure that bar with proof that the mark "has become distinctive." Genericness has no cure. Section 2(f) opens with the words "except as expressly excluded," and a generic term is excluded by the definition of "trademark" itself in 15 U.S.C. § 1127: a trademark is a word or device used "to identify and distinguish" one seller's goods from another's and "to indicate the source." A word that names the product cannot do that job.

The policy behind the two rules is competition, not fairness to the first user. If one bakery could own the word "bread," every other bakery would need a longer and less efficient phrase to tell customers what it sells. Descriptive terms are treated more gently because they are not the only way to name the product; the law lets a company own BEST BUY once the public reads it as a store rather than a sales pitch, and then lets competitors go on using "best buy" descriptively under the fair use defense in 15 U.S.C. § 1115(b)(4). That defense is the subject of its own cluster; see Trademark Fair Use: Descriptive Use, Nominative Use, and Expressive Works After Jack Daniel's. For the rest of the spectrum and how examiners place a mark on it, see Trademark Overview: Substantive Standards for Protection and Selecting a Protectable Trademark Checklist: A Practical Checklist.

One more structural point. Everything below applies to the three kinds of "mark" the Lanham Act recognizes: words, trade dress (packaging and product design), and the non-traditional signs (color, sound, scent). But the categories do not map the same way onto all of them. The Supreme Court held in Wal-Mart Stores, Inc. v. Samara Bros., 529 U.S. 205 (2000) that product design is never inherently distinctive, and in Qualitex Co. v. Jacobson Products Co., 514 U.S. 159 (1995) that a color can be a mark only on a showing of secondary meaning. For those signs there is no suggestive or arbitrary rung; the owner starts at "descriptive" and must climb with 2(f) evidence, and the Abercrombie ladder is replaced by a two-question inquiry: is it functional, and has it acquired distinctiveness. We return to that in the trade dress section below.

Genericness: the primary significance test

The modern test for genericness has a statutory home. When Congress amended the Lanham Act in 1984, it wrote into the cancellation provision, 15 U.S.C. § 1064(3), that "the primary significance of the registered mark to the relevant public rather than purchaser motivation shall be the test for determining whether the registered mark has become the generic name of goods or services." The same amendment added to § 1127 the rule that a mark is not generic "solely because such mark is also used as a name of or to identify a unique product or service." Both sentences were written to overrule a specific case: in Anti-Monopoly, Inc. v. General Mills Fun Group, 684 F.2d 1316 (9th Cir. 1982), the Ninth Circuit had held MONOPOLY generic on the strength of a survey showing that buyers wanted the game, not the maker. Congress thought that confused what consumers wanted with what the word meant, and the primary significance test now governs in every forum.

The Federal Circuit turned the statutory sentence into a two-step inquiry in H. Marvin Ginn Corp. v. International Ass'n of Fire Chiefs, Inc., 782 F.2d 987 (Fed. Cir. 1986): first, identify the genus (the category) of goods or services at issue; second, ask whether the relevant public understands the term primarily to refer to that genus. In Ginn the Board had cancelled the registration of FIRE CHIEF for a magazine on the theory that the title named the magazine's target audience. The Federal Circuit reversed: the genus was magazines directed to the field of fire fighting, nothing in the record showed the relevant public calling that class of publications "fire chief," and at most the term was descriptive. That decision supplies the structure of every USPTO genericness refusal and every TTAB cancellation on that ground. See TMEP § 1209.01(c)(i).

Three things about the test deserve emphasis because they decide cases.

The genus is defined by the goods in the application or registration, not by the owner's marketing. If the identification says "restaurant services," the question is whether the term names restaurant services or a key aspect of them; the applicant does not get to narrow the genus to "upscale restaurants serving Latin American food" after the fact. Conversely, a challenger cannot inflate the genus to make a term look generic. The fight over the genus is frequently the whole case.

The relevant public is the actual or potential purchaser, not the trade and not the lexicographer. For a prescription drug the public may be physicians; for a consumer product it is ordinary shoppers; for a business-to-business component it is the engineers and buyers who specify it. Judge Hand's split in Bayer Co. v. United Drug Co., 272 F. 505 (S.D.N.Y. 1921), between consumers (for whom "aspirin" was generic) and the trade (for whom it still meant Bayer) was the first careful recognition that the same word can be a brand to one audience and a noun to another, and that the injunction can be tailored accordingly.

"Primary" means primary, not exclusive. A term can have both a generic and a brand meaning; the question is which one dominates in the mind of the relevant public. That is why mixed evidence usually defeats a genericness refusal. In In re Merrill Lynch, Pierce, Fenner & Smith, Inc., 828 F.2d 1567 (Fed. Cir. 1987), the Board had refused CASH MANAGEMENT ACCOUNT as generic for brokerage services on the strength of press clippings; the Federal Circuit found that the same clippings showed substantial recognition of the term as Merrill Lynch's and reversed, instructing the Board to resolve reasonable doubt in the applicant's favor, since anyone harmed by the registration can still oppose it or petition to cancel.

The Ninth Circuit phrases the same inquiry as the "who are you / what are you" test: a mark answers the buyer's question "who are you? where do you come from? who vouches for you?", while a generic term answers "what are you?" The formulation traces to Professor McCarthy by way of Official Airline Guides, Inc. v. Goss, 6 F.3d 1385 (9th Cir. 1993), and the court applied it in Filipino Yellow Pages, Inc. v. Asian Journal Publications, Inc., 198 F.3d 1143 (9th Cir. 1999), which affirmed summary judgment that FILIPINO YELLOW PAGES was generic for a telephone directory aimed at Filipino-Americans and which also confirmed, for that circuit, that the party asserting an unregistered mark bears the burden of proving it is not generic once the defendant raises the issue.

Generic from birth versus genericide: two roads to the same place

A term can be generic in two ways, and the distinction matters for how the case is argued even though the legal result is identical.

Genericness ab initio means the term named the category before the applicant ever adopted it, or named it from the moment the category came into existence. "Pretzel crisps" for pretzel crackers, "churrascos" for a restaurant serving that style of grilled meat, "cash management account" (as the Board saw it) for a brokerage account that manages cash. The evidence is dictionaries, trade usage, competitor usage, and the applicant's own descriptive use. The owner's advertising spend is almost irrelevant, because no amount of advertising can convert the name of a thing into a brand for that thing. The one famous exception that proves the rule is Booking.com, discussed below, where the Court accepted that even a generic.com term can be perceived as a brand; but the Court's premise was that the public actually perceives it that way, which is a question of fact, not a reward for marketing.

Genericide is the death of a mark that was once distinctive. The classic graveyard is well known: aspirin, cellophane (DuPont Cellophane Co. v. Waxed Products Co., 85 F.2d 75 (2d Cir. 1936)), escalator (lost by Otis in a 1950 Patent Office cancellation proceeding in which Otis's own advertising and patents had used "escalator" as the name of the machine), thermos, trampoline, yo-yo, dry ice, linoleum, and more recently "murphy bed," which the Second Circuit found generic in Murphy Door Bed Co. v. Interior Sleep Systems, Inc., 874 F.2d 95 (2d Cir. 1989), relying on an earlier Patent Office refusal to register the term, dictionaries that list "Murphy bed" as the standard name for a bed that folds into a wall, and decades of newspaper and magazine usage. The mechanism is captured in the abandonment definition in 15 U.S.C. § 1127: a mark is deemed abandoned "when any course of conduct of the owner, including acts of omission as well as commission, causes the mark to become the generic name for the goods or services." The statute is explicit that doing nothing counts.

The thermos case is the best teaching example because the court described the owner's mistakes in detail. In King-Seeley Thermos Co. v. Aladdin Industries, Inc., 321 F.2d 577 (2d Cir. 1963), the Second Circuit affirmed a finding that "thermos" had become the generic name for vacuum-insulated bottles. The company had itself used the word generically in early advertising, had taken decades to begin policing dictionaries and the press, and by the time it did, the word had entered ordinary speech in lower case. The remedy is instructive: Aladdin could use "thermos" in lower case, could not use "Thermos" with a capital T or the phrase "genuine thermos," and had to accompany the word with its own house mark. The court was protecting whatever residual brand significance remained while refusing to let King-Seeley withdraw a common noun from the language.

The survivors show the other side. Xerox spent decades running advertisements asking the public to "photocopy" rather than "xerox" a document; Johnson & Johnson rewrote a jingle so that it sings "I am stuck on Band-Aid brand" rather than "stuck on Band-Aid"; Kimberly-Clark places "Kleenex brand tissues" on every package. None of those companies could stop people from using the words loosely in conversation, and none needed to. What they needed was a record showing that the primary significance of the word to purchasers, when it mattered, remained the brand. Verb use, noun use, and lower-case use by the public are evidence; they are not the test. That is the holding of Elliott v. Google, discussed next.

One structural difference between the two roads: a mark that was generic ab initio was never valid, so the registration was void from the start and a challenger need only prove what the word meant when it was registered. A mark that suffered genericide was valid once, which means an infringement that occurred before the tipping point was an infringement, and the challenger must prove that the public's primary understanding has changed. In practice, both challenges are brought the same way: an ex parte refusal during examination, an opposition or cancellation before the Trademark Trial and Appeal Board (TTAB), or a counterclaim or affirmative defense in litigation. For the mechanics of a cancellation petition, see Petitioning to Cancel a Trademark Registration: A Practical Checklist and Trademark Cancellation in Federal Litigation: A Practical Guide.

The leading genericness cases, explained

USPTO v. Booking.com: no per se rule for generic.com

The USPTO had refused to register BOOKING.COM for travel reservation services on the theory that "booking" is generic for that service and adding ".com" to a generic term produces another generic term, much as adding "Company" to a generic word produces nothing protectable under the nineteenth-century Goodyear rule. The district court, after the applicant submitted a Teflon-format survey showing that about three-quarters of respondents treated BOOKING.COM as a brand name, found the term descriptive with acquired distinctiveness, and the Fourth Circuit affirmed in Booking.com B.V. v. USPTO, 915 F.3d 171 (4th Cir. 2019). The Supreme Court affirmed as well in USPTO v. Booking.com B.V., 591 U.S. 549 (2020).

Justice Ginsburg's opinion makes three points that now control every "generic.com" and most compound-term cases. First, whether a term is generic depends on its meaning to consumers, so a compound term is generic only if the term as a whole signifies a class of goods or services to the relevant public. Second, because only one entity can hold a given domain name at a time, consumers may well understand "booking.com" to refer to a single website, which is a source, and that understanding is exactly what the primary significance test asks about. Third, the USPTO's fear that registration would let the owner monopolize "booking" was answered by the weakness of the mark: a descriptive mark that has barely crossed into distinctiveness has a narrow scope, and the fair use defense protects competitors' descriptive use. The Court was careful to say that it was not holding generic.com terms distinctive as a class; it rejected a per se rule in both directions.

Justice Breyer's dissent is worth reading for the survey critique: he argued that Teflon-format surveys measure whether consumers recognize a term as "a brand," which is not the same as whether the term is the common name of the services, and that a dominant seller can always teach consumers to treat the generic name as its brand. The majority acknowledged that surveys must be designed with care and that dictionaries, usage by consumers and competitors, and other sources of meaning remain relevant. Practitioners should treat that paragraph as an invitation: after Booking.com, the quality of the evidence of consumer perception decides the case.

Elliott v. Google: verb use is not genericide

David Elliott registered hundreds of domain names containing "google," lost them in a UDRP proceeding, and then sued to cancel the GOOGLE registrations as generic, arguing that the public uses "google" as a verb meaning "to search the internet." The Ninth Circuit affirmed summary judgment for Google in Elliott v. Google, Inc., 860 F.3d 1151 (9th Cir. 2017). Two holdings matter. First, genericness is always evaluated with respect to a particular type of good or service; the question was whether the public understands "google" primarily as the name for internet search engines, not whether it uses the word as a verb for the act of searching. Second, verb use is not automatically generic use. A speaker who says "I googled it" may mean "I used the Google search engine" (what the court called discriminate use, consistent with brand significance) or may mean "I searched the internet with whatever engine" (indiscriminate use). Elliott had evidence of verb use but had not shown that the public primarily understands the word as the generic name for search engines, and Google's Teflon survey showed that more than ninety percent of respondents classified GOOGLE as a brand. The court also noted that no competitor needed the word "google" to describe its own search engine, which supported the inference that the word had not become the only available name for the category.

Elliott is the case to cite when a client's product name is becoming a verb. It does not say verb use is harmless; it says verb use is evidence to be weighed and that the mark owner can win with a survey and a record of brand-conscious use.

In re Cordua Restaurants: a term for a key aspect of the services

Cordua owned a registration for CHURRASCOS as a word mark for restaurant services and applied to register a stylized version. The examiner refused the new application as generic, and the Federal Circuit affirmed in In re Cordua Restaurants, Inc., 823 F.3d 594 (Fed. Cir. 2016). The court held that "churrascos" (a style of grilled meat and, by extension, a restaurant serving it) names a key aspect of the genus "restaurant services," and a term that names a key aspect or a sub-category of the identified services is generic for those services even if it does not name the whole genus. The court also held that the prior registration of the word mark did not bind the Office on the later application; each application is examined on its own record, and a registered mark can be challenged as generic at any time. Cordua is the authority an examiner will cite to refuse a term like GLUTEN-FREE BAKERY for bakery services on the ground that it names what the services are about, and it explains why "pizzeria," in the court's own example, is generic for restaurant services even though it names only one kind of restaurant.

Princeton Vanguard v. Frito-Lay: the term as a whole, and surveys count

Princeton Vanguard sold PRETZEL CRISPS, a flat pretzel cracker, and held a Supplemental Register registration plus a pending Principal Register application. Frito-Lay opposed and petitioned to cancel on genericness, and in 2014 the Board agreed, reasoning largely from the component words: "pretzel" was generic for pretzels and "crisps" was generic for crackers, so the compound was generic too. The Federal Circuit vacated in Princeton Vanguard, LLC v. Frito-Lay North America, Inc., 786 F.3d 960 (Fed. Cir. 2015). It held that the Board had applied the wrong test by analyzing the words separately: whether the term is a compound word or a phrase, the Ginn inquiry is always what the term as a whole means to the relevant public, and evidence about the component words is relevant only as a step toward that question. The court also stressed that direct evidence of how the public perceives a term, such as consumer surveys and testimony, is preferable to indirect evidence, and it could not see why the Board had analyzed the words one at a time when the record, including competing Teflon surveys (one reporting 55 percent brand responses, the other 41 percent), spoke to the term as a whole. On remand the Board again found PRETZEL CRISPS generic, this time on an analysis of the whole term, and held in the alternative that the 2(f) evidence fell short: Frito-Lay North America, Inc. v. Princeton Vanguard, LLC, Opp. No. 91195552 (T.T.A.B. Sept. 6, 2017) (precedential). The owner then took the second round to district court under 15 U.S.C. § 1071(b), and the Fourth Circuit held in Snyder's-Lance, Inc. v. Frito-Lay North America, Inc., 991 F.3d 512 (4th Cir. 2021) that it could do so even though it had chosen the Federal Circuit the first time. The lesson is procedural as much as substantive: a genericness decision that skips the "term as a whole" step, or waves away a survey, will not survive appeal.

Royal Crown v. Coca-Cola: ZERO, sub-genus, and the burden on highly descriptive terms

Coca-Cola applied to register a family of marks containing ZERO (COKE ZERO, SPRITE ZERO, and so on) for soft drinks, claiming acquired distinctiveness in ZERO under Section 2(f). Royal Crown opposed, arguing that ZERO is generic for zero-calorie beverages, or at minimum so highly descriptive that Coca-Cola's evidence could not establish secondary meaning. The Board dismissed the oppositions. In Royal Crown Co. v. Coca-Cola Co., 892 F.3d 1358 (Fed. Cir. 2018), the Federal Circuit vacated and remanded on two grounds that are now routine citations. First, the Board had framed the genus too narrowly; it should have asked whether ZERO names a key aspect of the goods (calorie content) or a sub-category (zero-calorie soft drinks), which is enough for genericness under Cordua. Second, when a term is not generic but is highly descriptive, the Board must make an express finding about the degree of descriptiveness before weighing 2(f) evidence, because the burden of proving acquired distinctiveness rises with descriptiveness. Sales figures and advertising dollars that would suffice for a mildly descriptive term may fall well short for a term at the edge of genericness.

The older anchors: Kellogg, Bayer, and the Teflon and Thermos surveys

Two Supreme Court and district court decisions supply the vocabulary everyone still uses. In Kellogg Co. v. National Biscuit Co., 305 U.S. 111 (1938), the Court held that "shredded wheat" was the generic name of a pillow-shaped biscuit whose patents had expired, that the pillow shape was functional and likewise free for all to use, and that Kellogg's only obligation was to use reasonable care to identify its own product as Kellogg's. Justice Brandeis's opinion is the origin of the rule that the public pays for a patent monopoly with the right, after expiry, to make the product and call it by its name. Bayer, discussed above, supplied the primary significance test in its original form ("What do the buyers understand by the word for whose use the parties are contending?") and the idea that the answer may differ by audience.

The two survey formats take their names from later cases. The Thermos format comes from the district court decision affirmed in King-Seeley, American Thermos Products Co. v. Aladdin Industries, Inc., 207 F. Supp. 9 (D. Conn. 1962): respondents were asked, in effect, what they would call a container that keeps liquids hot or cold, and roughly three-quarters answered "thermos." The Teflon format comes from E.I. DuPont de Nemours & Co. v. Yoshida International, Inc., 393 F. Supp. 502 (E.D.N.Y. 1975), in which respondents were first taught the difference between a brand name and a common name with examples, then asked to classify a list of terms (including TEFLON) as brand or common; a substantial majority called TEFLON a brand, and the mark survived. The formats, their critiques, and how to choose between them are covered in the survey section below.

How the USPTO and the courts actually decide genericness

The evidence that counts

Because the test is consumer understanding, any competent evidence of what the relevant public thinks is admissible, and the cases keep returning to the same categories. Dictionary entries (and their absence) carry real weight; a listing of the term in lower case as a common noun is powerful for the challenger, and a listing that marks the term as a trademark is powerful for the owner. Media and trade usage is next: how newspapers, trade journals, and websites use the term, and especially whether they use it to refer to the category without the owner's product in mind. Competitor usage matters because a term that competitors need in order to describe their own goods is probably generic, and a term that no competitor uses (as in Elliott) is probably not. The owner's own usage is often decisive; an applicant whose specimen says "our churrascos are grilled over oak" has told the examiner what the word means. Consumer testimony and, above all, surveys round out the list. Restatement (Third) of Unfair Competition § 15 and 2 McCarthy on Trademarks and Unfair Competition ch. 12 collect the categories and the cases.

The USPTO adds a practical rule from the internet age: the examining attorney may rely on evidence of how a term is used on websites and in online publications, and the Board treats a properly introduced web page as evidence of what the page says and how the term is used, though not as proof of the truth of its contents. An applicant responding to a genericness refusal should expect a record built from search results and should answer it with evidence rather than argument.

Burdens and standards of proof

The standard of proof depends on where the fight is and who registered what.

The "clear evidence" line comes from Merrill Lynch and was repeated in In re Hotels.com, L.P., 573 F.3d 1300 (Fed. Cir. 2009), and Cordua. For years the Board read it as a heightened standard akin to clear and convincing proof, and applicants argued it that way. In 2022 the USPTO issued Examination Guide 1-22 taking the position that "clear evidence" does not mean clear and convincing evidence and that, as with other substantive refusals, the examining attorney must establish a reasonable predicate for the refusal with relevant evidence, after which the applicant must answer it. Check the current TMEP § 1209.01(c)(i) before briefing the point, because that is where the Office states the standard its examiners apply. The practical consequence is that applicants can no longer win a genericness appeal simply by calling the examiner's evidence thin; they must answer it. We flag the point because older briefs and older secondary sources still describe the standard as heightened.

In litigation, the allocation turns on the register. A Principal Register registration shifts the burden to the challenger. A Supplemental Register registration carries no presumption of validity at all (see 15 U.S.C. § 1094 and the discussion in Common-Law Rights, the Supplemental Register, and the Principal Register: A Practical Guide), so the owner of PRETZEL CRISPS on the Supplemental Register stood in the same position as the owner of an unregistered term. For unregistered terms the plaintiff ordinarily carries the burden on every element, including that the term is not generic, as Filipino Yellow Pages held. The Second Circuit carved out an exception in Murphy Door Bed: where the claim is that the public expropriated a term the plaintiff coined (genericide), rather than that the term was generic from the start, the defendant bears the burden of proving genericness even if the mark is unregistered. The ab initio versus genericide distinction therefore decides who loses a close case.

Genericness is a question of fact, and summary judgment is possible

Genericness is a fact question reviewed for clear error after a bench trial under Fed. R. Civ. P. 52(a)(6) and for substantial evidence on appeal from the Board. But courts do resolve it under Fed. R. Civ. P. 56 when the record is one-sided, as in Elliott (summary judgment for the owner, on a record of verb use and little else) and Filipino Yellow Pages (summary judgment for the accused, on a record the owner could not rebut), and a court asked for a preliminary injunction will deny it if the term looks generic, as the Fourth Circuit did for "crab house" in Hunt Masters, Inc. v. Landry's Seafood Restaurant, Inc., 240 F.3d 251 (4th Cir. 2001). The usual pattern is that a challenger with dictionary entries, trade usage, and a competent survey survives summary judgment, and an owner with a strong Teflon survey and a record of brand-conscious usage wins it.

Surveys: Teflon, Thermos, and their critiques

A genericness survey measures whether the relevant public understands a term as a common name or a brand. Two formats dominate, and each has a known weakness that the other side's expert will exploit. For survey design across all trademark issues, see Trademark Consumer Survey Design: A Practical Checklist and Consumer Survey Expert Methodology in Trademark Cases: A Practical Guide; for the admissibility fight, see Daubert Challenges to Consumer Survey Experts in Trademark Litigation: A Practical Guide.

Teflon format. Respondents are screened for membership in the relevant purchasing public, taught the distinction between a brand name (a name used by one company) and a common name (a name shared by all makers of a type of product) with neutral examples, tested on their understanding with a short quiz, and then asked to classify a list of terms, including the mark at issue and a mix of known brands and known common names, as brand, common, or don't know. The test terms double as controls: a respondent who calls REFRIGERATOR a brand is noise. The output is a percentage of respondents who treat the disputed term as a brand. Courts have accepted brand-classification rates in the sixties and above as strong evidence of non-genericness; a figure near fifty percent is contestable; a figure in the thirties is evidence for the challenger.

The critique, pressed hardest by Justice Breyer in Booking.com and by many academic commentators, is that "brand" and "common name" are not mutually exclusive in the respondent's mind. A respondent may know that Thermos is "a brand" because one company sells thermoses under that name, while still using "thermos" as the only word for the product. The Teflon question measures recognition of a trademark claim rather than primary significance. A well-designed survey addresses this with a careful preamble (brand means a name used by one company to identify its products, common name means a name that any company making the product can use) and with controls chosen to resemble the disputed term in structure.

Thermos format. Respondents are placed in a purchasing scenario and asked, open-ended, what they would call the product or what they would ask for in a store, with follow-up questions about whether they know any other names and any brand names for it. The output is the share of respondents who volunteer the disputed term as the name of the product. The format's weakness is the opposite of Teflon's: a dominant brand will be volunteered as a name for the product even by respondents who know it is a brand, so a high "thermos" response may show market share rather than genericness. Follow-up questions ("Is that the name of the product or the name of a brand?") mitigate the problem, and a careful expert will present both the raw and the adjusted figures.

Which to use. Owners usually prefer Teflon because it produces a brand percentage; challengers often prefer Thermos because it produces a usage percentage. Because the Federal Circuit in Princeton Vanguard called direct evidence such as surveys preferable to indirect evidence, in a contested matter the parties will often each run one, and the case becomes a battle over universes, controls, and preambles. Two design points recur in the opinions: the universe must be the actual or potential purchasers of the goods identified in the application or registration (not the general public, and not the trade unless the trade is the purchaser), and the survey must test the term as the public encounters it, which for a compound term means the whole term and not its parts.

When a survey will not help. In Hunt Masters the plaintiff, which ran the Charleston Crab House, offered a consumer survey to show that "crab house" was not generic. The Fourth Circuit held the survey irrelevant: the plaintiff had not coined the term, which already named a class of restaurants that serve crabs (just as "ale house" and "coffee shop" name classes), so there was no genericide question for a survey to answer. Whether that categorical rule survives Booking.com, which made consumer perception the test for every term, is open to argument, and the Board said much the same about Teflon surveys for terms that are not inherently distinctive in its 2017 PRETZEL CRISPS decision. The practical lesson holds either way: a survey will not rescue a term that dictionaries, competitors, and the trade already use as a category name. Conversely, when the record is mixed and the parties are well funded, the absence of a survey is a conspicuous gap that fact finders notice.

Cancellation at any time: Section 14(3), incontestability, and Park 'N Fly

A generic registration is never safe. 15 U.S.C. § 1064(3) permits a petition to cancel "at any time" if the registered mark "becomes the generic name for the goods or services, or a portion thereof, for which it is registered," and the same subsection covers functionality, abandonment, and fraud. The five-year window that cuts off most cancellation grounds under § 1064(1) does not apply. Nor does incontestability. 15 U.S.C. § 1065(4) provides that "no incontestable right shall be acquired in a mark which is the generic name for the goods or services or a portion thereof," and 15 U.S.C. § 1115(b) makes the conclusive evidentiary effect of an incontestable registration expressly subject to the conditions in § 1065. A court hearing an infringement case may order cancellation itself under 15 U.S.C. § 1119.

The contrast with descriptiveness is sharp and is the holding of Park 'N Fly, Inc. v. Dollar Park & Fly, Inc., 469 U.S. 189 (1985). Park 'N Fly held an incontestable registration for airport parking services; Dollar Park and Fly argued the mark was merely descriptive and therefore invalid. The Supreme Court held that an incontestable registration cannot be challenged on the ground that the mark is merely descriptive, because Congress listed the defenses that survive incontestability and descriptiveness is not among them. The Court was explicit that genericness remained available at any time; in that case the district court had found after a bench trial that the mark was not generic, and the Ninth Circuit had left that finding alone. The practical rule: once a descriptive mark has been registered on the Principal Register (whether under 2(f) or by an examiner's mistake) and has become incontestable, the only way to attack its distinctiveness is to prove that it is generic, which is a much harder showing. Challengers who could have proved lack of secondary meaning in year four are often stuck in year six. For the incontestability mechanics, see Maintaining Trademark Registrations.

A note on the phrase "or a portion thereof" in both sections: a registration can be partially cancelled, or restricted, where part of the mark is generic. The usual tool during examination is a disclaimer of the generic matter; after registration, a petition under § 1064 may seek to cancel as to goods for which the term is generic while leaving the rest intact.

Generic marks and trade dress: shapes, colors, and the Wal-Mart rule

The vocabulary of genericness applies awkwardly to product shapes, because a shape does not "name" anything. The law reaches the same result through two other routes. First, product design is never inherently distinctive. Wal-Mart Stores, Inc. v. Samara Bros., 529 U.S. 205, 213 (2000) held that the design of a product (there, a line of children's garments with appliqués) can be protected as trade dress only on proof of secondary meaning, reasoning that consumers do not ordinarily read product features as source indicators and that allowing suits over unproven design trade dress would chill competition. Packaging may still be inherently distinctive under Two Pesos, and the Court instructed that close cases be classified as product design. Second, a shape that is functional is unprotectable regardless of distinctiveness under 15 U.S.C. § 1052(e)(5) and TrafFix Devices, Inc. v. Marketing Displays, Inc., 532 U.S. 23 (2001), which applied the Inwood definition (a feature is functional if it is essential to the use or purpose of the article or affects its cost or quality) from Inwood Laboratories, Inc. v. Ives Laboratories, Inc., 456 U.S. 844, 850 n.10 (1982).

The Board and the Federal Circuit also use "generic" in a looser sense for product configurations that are so common in the industry that no amount of evidence could make them source-identifying: a basic guitar body outline, a standard pill shape, a plain rectangular bottle. In Stuart Spector Designs, Ltd. v. Fender Musical Instruments Corp., 94 U.S.P.Q.2d 1549 (T.T.A.B. 2009), the Board sustained oppositions to Fender's 2(f) applications for the outlines of the Stratocaster, Telecaster, and Precision Bass bodies despite half a century of use and enormous sales, because the shapes had been copied so widely for so long that the public saw them as the shape of an electric guitar rather than as Fender's. Whether one calls that "generic" or "no acquired distinctiveness," the result is the same, and the lesson for product-design applicants is the one in the worked example at the end of this article: a design that competitors have used freely for years is almost never rescued by advertising figures. The Federal Circuit has confirmed that this use of "generic" has teeth: in Sunrise Jewelry Manufacturing Corp. v. Fred S.A., 175 F.3d 1322 (Fed. Cir. 1999), it held that "generic name" in § 1064(3) reaches trade dress, so even an incontestable registration for a product design (there, a metallic nautical rope design for watches and jewelry) can be cancelled at any time if the design is generic for the goods.

Color follows Qualitex: a single color applied to goods can be registered on the Principal Register only with proof of acquired distinctiveness, and the Federal Circuit's decision in In re Owens-Corning Fiberglas Corp., 774 F.2d 1116 (Fed. Cir. 1985), which accepted pink for fiberglass insulation on a record of decades of advertising built around the color itself (more than $42 million in consumer advertising over a decade, with Pink Panther campaigns urging buyers to "think pink" and "put your house in the pink"), remains the model of what a successful color record looks like. See TMEP § 1202.05. For the full trade dress treatment, see The Evolution of Trade Dress Protection for Product Design and Configuration: A Comprehensive Case Law Analysis and Trade Dress Protection Toolkit: From Product Design and Packaging to Registration, Functionality, and Enforcement.

Acquired distinctiveness under Section 2(f): the statute, the routes, and the leading cases

Section 2(f) is a door, not a reward for effort. It provides that, except for matter barred by subsections (a) through (d), (e)(3), and (e)(5), "nothing in this chapter shall prevent the registration of a mark used by the applicant which has become distinctive of the applicant's goods in commerce." 15 U.S.C. § 1052(f). Read the exclusions closely, because they define what evidence can and cannot fix. Proof of acquired distinctiveness cures a merely descriptive refusal under § 2(e)(1), a primarily geographically descriptive refusal under § 2(e)(2), and a surname refusal under § 2(e)(4), and it is the only route to registration for product designs, single colors, and most slogans. It cannot cure functionality under § 2(e)(5), deceptive matter or a false suggestion of connection under § 2(a), likelihood of confusion under § 2(d), or (for marks that became distinctive after December 8, 1993) geographic deceptive misdescriptiveness under § 2(e)(3). And it cannot cure genericness at all: as Ginn put it, a generic term is "the ultimate in descriptiveness" and is "incapable of acquiring de jure distinctiveness." The USPTO treats a 2(f) claim for generic matter as legally insufficient no matter how much evidence comes with it (TMEP § 1212.02(i)).

What the applicant must actually prove

"Has become distinctive" means that in the minds of the relevant public the primary significance of the term or design is the producer rather than the product. The formulation is the same in every forum: Coach Services, Inc. v. Triumph Learning LLC, 668 F.3d 1356 (Fed. Cir. 2012) (Board), Zatarains, Inc. v. Oak Grove Smokehouse, Inc., 698 F.2d 786 (5th Cir. 1983) (litigation), and Restatement (Third) of Unfair Competition § 13. Consumers do not need to know the owner's corporate name; an association with a single source, even an anonymous one, is enough. The Federal Circuit applied that "anonymous source" rule in Tone Brothers, Inc. v. Sysco Corp., 28 F.3d 1192 (Fed. Cir. 1994), where survey respondents associated an unlabeled spice container with one company they could not name. Restatement § 13 makes the same point (a designation may identify a source "whether known or anonymous"), and two chapters of McCarthy carry most of the doctrine: 2 McCarthy on Trademarks and Unfair Competition ch. 15 (secondary meaning) and the survey sections of ch. 32.

Three ways to make the showing at the USPTO

The Trademark Rules give an applicant three routes, and they are cumulative rather than exclusive. See 37 C.F.R. § 2.41(a) and TMEP §§ 1212.04 to 1212.06.

The Dial-A-Mattress decision, In re Dial-A-Mattress Operating Corp., 240 F.3d 1341 (Fed. Cir. 2001), is also a genericness case worth knowing: the Board had called the mnemonic generic because "mattress" is generic, and the court reversed because no evidence showed the public calling shop-at-home mattress retailers "1-888-M-A-T-R-E-S-S." The mark was descriptive, and the prior registration carried it across.

Who bears the burden, and why it grows with descriptiveness

The applicant bears the burden of proving acquired distinctiveness in examination, and it keeps the ultimate burden of persuasion in an opposition too. In Yamaha International Corp. v. Hoshino Gakki Co., 840 F.2d 1572 (Fed. Cir. 1988), which upheld Hoshino's 2(f) registrations for guitar peg-head shapes, the court treated a 2(f) claim as something like an affirmative defense: the opposer need only produce enough evidence or argument that the Board could find the showing inadequate, and the applicant must then carry the day. In litigation the plaintiff carries the burden for an unregistered mark; for a registered mark the presumption under 15 U.S.C. § 1057(b) shifts both the burden of production and the burden of persuasion to the challenger, but only for infringement that began after the registration issued.

How much evidence is enough depends on how descriptive the matter is. "The applicant's burden of showing acquired distinctiveness increases with the level of descriptiveness; a more descriptive term requires more evidence of secondary meaning." In re Steelbuilding.com, 415 F.3d 1293 (Fed. Cir. 2005). Royal Crown turned that sliding scale into a procedural requirement: the Board must make an express finding on the degree of descriptiveness before it weighs the 2(f) evidence. At the far end of the scale some matter cannot be rescued at all. In In re Boston Beer Co., 198 F.3d 1370 (Fed. Cir. 1999), the brewer of Samuel Adams showed more than $10 million a year in advertising and roughly $85 million in annual sales under THE BEST BEER IN AMERICA, and the court still held the phrase so laudatory, and so widely used by others, that it could not function as a mark.

Two procedural consequences follow. First, a 2(f) claim is treated as a concession that the matter is not inherently distinctive (TMEP § 1212.02(b)), so an applicant with a plausible suggestiveness argument should make it first and claim 2(f) only in the alternative (TMEP § 1212.02(c)), preserving both for appeal. Second, the examining attorney can accept a 2(f) claim as to only part of a mark (a "2(f) in part" claim under TMEP § 1212.02(f)), which lets an applicant register a composite while claiming acquired distinctiveness only in the descriptive element, and the alternative is a disclaimer under 15 U.S.C. § 1056.

The Supplemental Register as a waypoint

An applicant that cannot yet prove acquired distinctiveness can often register a descriptive term, a surname, or a nonfunctional design on the Supplemental Register under 15 U.S.C. § 1091, which is reserved for matter "capable of distinguishing" the applicant's goods. A Supplemental registration carries no presumption of validity and cannot become incontestable, but it lets the owner use the ® symbol, it will be cited against later confusingly similar applications, and it is expressly not an admission that the mark lacks distinctiveness (15 U.S.C. § 1095). Two traps: an intent-to-use applicant cannot amend to the Supplemental Register until it files an acceptable allegation of use, and the amendment resets the effective filing date to that filing (37 C.F.R. § 2.75(b)); and a Madrid § 66(a) application cannot go to the Supplemental Register at all. Generic matter is not "capable of distinguishing," so the Supplemental Register is no refuge from a genericness refusal, as PRETZEL CRISPS learned when Frito-Lay attacked its Supplemental registration and its 2(f) application in the same proceeding.

The leading acquired distinctiveness cases, explained

Zatarains: the same company, the same box, opposite results. Zatarain's sold two batter mixes in nearly identical packaging. FISH-FRI had been in use since 1950; the company sold 916,385 cases between 1964 and 1979, advertised it, and offered two surveys in which 23 and 28 percent of New Orleans respondents named Zatarain's FISH-FRI when asked about coatings for frying fish. The Fifth Circuit thought the surveys thin but upheld the finding of secondary meaning because the circumstantial evidence pointed the same way. CHICK-FRI, used since 1968 without comparable evidence, had no secondary meaning, and its registration was cancelled. And even FISH-FRI's protection did not stop competitors from describing their own products as fish fry, because the defendants' descriptive use was a fair use. Secondary meaning is proved mark by mark, and it never takes the words out of the language.

Converse: when secondary meaning must exist, and how to weigh it. Converse registered the midsole design of the Chuck Taylor All Star (the stripes, toe cap, and diamond-patterned toe bumper) in 2013 and asked the International Trade Commission to exclude imitations. In Converse, Inc. v. International Trade Commission, 909 F.3d 1110 (Fed. Cir. 2018), the Federal Circuit vacated and remanded, holding that a trade dress owner must prove secondary meaning as of each accused infringer's first use, that the registration's presumption operates only from the registration date forward, and that uses more than five years before the relevant date matter only if they likely shaped consumer perception at that date. It also restated the test as six factors: (1) association by actual purchasers, typically measured by surveys; (2) length, degree, and exclusivity of use; (3) amount and manner of advertising; (4) amount of sales and number of customers; (5) intentional copying; and (6) unsolicited media coverage. Those six factors now organize most Board and Federal Circuit opinions, and the regional circuits use close variants.

Steelbuilding.com and Louisiana Fish Fry: highly descriptive terms need more than a declaration. Steelbuilding.com won on genericness (the Board had defined the genus too narrowly) and lost on acquired distinctiveness. Its evidence included an internet poll on a related website that did not count participants or stop anyone from voting repeatedly, and the court agreed it "does not even remotely follow the precepts of standard trademark name-recognition polls." In In re Louisiana Fish Fry Products, Ltd., 797 F.3d 1332 (Fed. Cir. 2015), the applicant had used LOUISIANA FISH FRY PRODUCTS for some thirty years and spent more than $2.4 million advertising it over three years, but every figure related to the composite mark, not to the highly descriptive phrase FISH FRY PRODUCTS for which it claimed distinctiveness. The court also held that the Board was within its discretion to refuse a five-year declaration as prima facie proof for a term that descriptive.

Coach Services and Real Foods: what the Board can and cannot ignore. In Coach Services the Board found COACH descriptive for test-preparation books but accepted the publisher's 2(f) showing without a survey, and the Federal Circuit agreed no survey was required. It vacated anyway, because the Board had refused to consider third-party "coach" titles published after the application date (in an opposition, acquired distinctiveness is tested as of the decision, not the filing date) and had relied on marketing documents the sponsoring witness could not authenticate. In Real Foods Pty Ltd. v. Frito-Lay North America, Inc., 906 F.3d 965 (Fed. Cir. 2018), CORN THINS and RICE THINS failed: they were highly descriptive, the applicant had done little advertising, its sales were "not insignificant" but not high, other sellers used THINS, and the opposer's survey showed only about ten percent of respondents associating CORN THINS with one company.

Proving secondary meaning: evidence that works and evidence that fails

The six Converse factors are not a scorecard. Fact finders ask one question (does the relevant public see this as a brand?) and treat each category of evidence as a window into that question. The table below summarizes how each window usually looks from the bench.

Look-for advertising beats advertising budgets

Advertising matters only if it teaches consumers to treat the matter as a brand. The First Circuit put it bluntly in Yankee Candle Co. v. Bridgewater Candle Co., 259 F.3d 25 (1st Cir. 2001): advertising that merely features the product "is no more probative of secondary meaning than are strong sales"; to count, it "must direct the consumer to those features claimed as trade dress." Owens-Corning's pink insulation is the model of the opposite, because the campaign was about the color. The same logic explains Louisiana Fish Fry: $2.4 million spent promoting a composite mark said nothing about how consumers read the descriptive phrase inside it. Courts also discount advertising that buries the claimed matter beneath a prominent house mark, and the Eighth Circuit has said so expressly for trade dress (Aromatique, Inc. v. Gold Seal, Inc., 28 F.3d 863 (8th Cir. 1994)).

Why sales figures so often fail

Sales prove that people like the product. They do not prove that people read the word or shape as a brand, and the Federal Circuit has said for decades that strong demand for a product usually reflects the product's desirability rather than secondary meaning. In Cicena Ltd. v. Columbia Telecommunications Group, 900 F.2d 1546 (Fed. Cir. 1990), a hit see-through telephone had been on the market for about eighteen months, which the court said pointed "strongly away" from secondary meaning, and competitors' copying showed a desire to sell an appealing product, not to trade on goodwill. Boston Beer is the cautionary tale at scale. Sales help when they come with context: market share, repeat purchase rates, growth that tracks the brand campaign, and sales of products where the claimed matter is the only branding visible.

Unsolicited media, copying, and the circuit split on intent

Independent press that calls a design "the brand's signature" or uses a term as the owner's name is persuasive because the owner did not pay for it. Copying is more contested. Most circuits treat deliberate copying as relevant, but they divide over whether the copier must have intended to confuse buyers or trade on goodwill. The Tenth Circuit requires proof of that intent in product design cases because competitors usually copy features for their utility or appeal (Craft Smith, LLC v. EC Design, LLC, 969 F.3d 1092 (10th Cir. 2020)); the Ninth Circuit, in a 2023 furniture case, held that deliberate copying can support secondary meaning without separate proof of intent to confuse (Jason Scott Collection, Inc. v. Trendily Furniture, LLC, 68 F.4th 1203 (9th Cir. 2023)). Know your circuit before you build a case on a knockoff.

Surveys, declarations, and the anonymous source

A secondary meaning survey shows respondents the claimed matter stripped of other branding and asks whether they associate it with one company or more than one, with a control that removes the claimed feature. Courts generally treat a majority association as strong evidence, figures in the twenties and thirties as weak but sometimes enough alongside other proof (Zatarains), and figures near ten percent as evidence against the claim (Real Foods). Timing matters: Royal Crown questioned the weight of a survey taken five years before the decision, and Converse ties the inquiry to each infringer's first use. Declarations from customers help if they are numerous and in the declarants' own words; boilerplate statements signed by distributors or employees are routinely discounted. Experts design surveys, not counsel; the Ninth Circuit in Elliott approved excluding two surveys that the plaintiff's lawyer had designed and run, and admissibility runs through Fed. R. Evid. 702. For the deposition side of that fight, see Expert Depositions: Taking Apart the Opinion, Protecting Your Own Expert, and Building the Daubert Record.

Timing: the five-year presumption and the relevant date

The five-year statement in § 2(f) is a convenience for the Office, not an entitlement, and the TMEP warns that it will often be insufficient for highly descriptive matter, product designs, and colors (TMEP § 1212.05(a)). In litigation the relevant date is the date each defendant began its accused use (Converse); in an opposition it is the date of decision (Coach Services). That difference cuts both ways. An owner suing a long-established imitator must prove that secondary meaning existed years ago, with evidence from that period. An applicant in a slow opposition gets the benefit of fame that grows while the case is pending, but also bears the burden of third-party uses that appear while it is pending.

Keeping a brand from becoming a noun: usage discipline that works

The statute tells owners where the danger lies. Under 15 U.S.C. § 1127, a mark is abandoned when "any course of conduct of the owner, including acts of omission as well as commission," causes it to become generic. The cases then reward owners who can show they fought the drift. In Elliott the Ninth Circuit said that Google's policing of dictionary publishers weighed against genericide, citing Filipino Yellow Pages and King-Seeley for the converse point that an owner's failure to police supports a finding of genericness. An owner cannot control what people say at dinner, and the law does not ask it to. It can control its own usage, its partners' usage, and much of what competitors and publishers print, and those are the sources a tribunal reads.

The practices that actually move the evidence are few and unglamorous:

1. Pair the mark with a generic name, every time. TALLYO expense-splitting app, XEROX copier, KLEENEX tissues. The generic name is the antidote to genericide. The test is the term's primary significance to the relevant public, and the Seventh Circuit has added that genericide rarely happens until competitors cannot compete effectively without using the word, a showing that a well-established generic name makes much harder. The Elliott plaintiffs lost because the evidence showed the public still understood GOOGLE as the name of one company's search engine, with "search engine" doing the generic work; Bayer lost because the public knew no other word for the drug.
2. Never use your own mark as a noun, verb, plural, or possessive. An owner's own generic use is the most damaging evidence in the record. In In re Gould Paper Corp., 834 F.2d 1017 (Fed. Cir. 1987), the Federal Circuit found SCREENWIPE generic for screen-cleaning cloths largely because the applicant's own materials used it as the name of the product.
3. Make the mark look like a brand. Capitalize it or set it in a distinctive style, and use ™ before registration and ® after. The symbol creates no rights, but it signals brand status to readers and preserves damages under 15 U.S.C. § 1111.
4. Say "brand" where it helps. Kimberly-Clark's "Kleenex brand tissues" and the "I am stuck on Band-Aid brand" jingle are not vanity; they are evidence that the owner taught the public the difference.
5. Write usage rules into contracts. Licensees, resellers, marketplace partners, and API developers should be bound to the usage guidelines, and licensing should carry real quality control. See Trademark Licensing and Quality Control: How Naked Licensing Kills a Brand and Brand Licensing Toolkit: License Terms, Quality Programs, and Audit Rights.
6. Police competitors first, publishers second, and the public last. Competitor usage is the evidence that wins genericness cases; in Schwan's IP, LLC v. Kraft Pizza Co., 460 F.3d 971 (8th Cir. 2006), competitors' use of "brick oven" to name a style of frozen pizza supported the finding that the term was generic. Polite correction letters to editors and dictionary publishers come next. Threatening individual consumers is unnecessary and usually counterproductive.
7. Keep the file. Copies of corrections sent and answered, dictionary correspondence, enforcement letters, and a periodic Teflon survey baseline become exhibits when a challenger arrives. The owner who can show years of policing and a stable brand percentage looks like Google. The owner who cannot looks like King-Seeley.

A ready-to-adapt set of usage guidelines, with drafting notes, is in the Distinctiveness Toolkit: Secondary Meaning Declarations, Genericness Surveys, and Brand Usage Guidelines.

Strategy, costs, and practical realities for each side

If you own the mark or the application

Start upstream. A suggestive or coined mark avoids most of this article, and a clearance search should flag descriptive candidates before money is spent on them; see How to Conduct a Comprehensive Trademark Clearance Search. If the business insists on a descriptive name, plan for acquired distinctiveness from launch: run look-for advertising, keep dated copies of every ad, track sales by mark, and keep an unsolicited-media file.

Facing a § 2(e)(1) refusal, argue suggestiveness first, claim 2(f) in the alternative, aim every piece of evidence at the matter actually claimed, and keep the Supplemental Register in reserve. The procedure, deadlines, and evidence package are covered step by step in Proving Acquired Distinctiveness and Defeating a Genericness Challenge: A Practical Guide and Responding to a Trademark Office Action Checklist: A Practical Checklist.

Once registered, a descriptive mark has one great prize waiting: incontestability. File the Section 15 declaration as soon as the mark has five years of continuous post-registration use, because after Park 'N Fly a challenger can no longer attack it for lack of secondary meaning, only for genericness and the other § 1064(3) grounds. In litigation, prove secondary meaning as of each defendant's first use, commission the survey early enough to fix it if the pilot goes badly, and assume the defendant will counterclaim for cancellation under § 1119 on genericness grounds.

If you are challenging the mark or defending an infringement claim

Timing is everything. Before registration, a letter of protest can put dictionary, media, and competitor evidence in front of the examining attorney at almost no cost; see Expungement, Reexamination, and Letters of Protest: The Trademark Modernization Act in Practice. After publication, an opposition under 15 U.S.C. § 1063 must be filed within thirty days or an extension. For five years after registration, a petition to cancel can attack descriptiveness and the lack of secondary meaning; after that, only genericness and the other grounds in § 1064(3) remain.

In litigation, plead genericness as an affirmative defense and as a counterclaim for cancellation; the pleading mechanics are in Answering a Federal Complaint: Denials, Affirmative Defenses, Counterclaims, and Judgment on the Pleadings. Take discovery into the owner's own usage (the Gould problem), the history of the term, and every competitor who uses it. Define the genus from the identification of goods in the registration, which is what the Federal Circuit requires, as in Magic Wand, Inc. v. RDB, Inc., 940 F.2d 638 (Fed. Cir. 1991), rather than from the owner's marketing.

Then ask whether genericness is the right fight. If the real point is that the client used words in their ordinary descriptive sense, descriptive fair use under § 1115(b)(4) is narrower, cheaper, and does not require destroying the registration. The Second Circuit sent exactly that defense to a jury in Tiffany & Co. v. Costco Wholesale Corp., 971 F.3d 74 (2d Cir. 2020), where Costco said it used "Tiffany" to describe a style of ring setting. See Asserting or Defeating a Trademark Fair Use Defense: A Practical Guide. Remember, too, that winning on genericness does not always end the case: a first user may still have an unfair competition claim that requires the newcomer to distinguish itself, as in Blinded Veterans Ass'n v. Blinded American Veterans Foundation, 872 F.2d 1035 (D.C. Cir. 1989), where the D.C. Circuit held "blinded veterans" generic but remanded for consideration of passing off.

What it costs and how long it takes

Electronic discovery in a Board proceeding is not a rounding error. In the PRETZEL CRISPS dispute, the applicant told the Board it had spent roughly $200,000 on electronic discovery in a single year, in a ruling on a motion to compel that the Board designated precedential: Frito-Lay North America, Inc. v. Princeton Vanguard, LLC, 100 U.S.P.Q.2d 1904 (T.T.A.B. 2011). That case began in 2009 and was still being litigated in the Fourth Circuit in 2021. A litigation-grade survey with a report and deposition commonly costs tens of thousands of dollars and can exceed six figures, and each side usually needs a rebuttal expert as well. For managing the Board stages, see TTAB Practice Toolkit: Oppositions, Cancellations, and the Trademark Trial and Appeal Board from Pleading to Appeal.

Common mistakes

Owners make these:

• Using the mark as the name of the product in their own ads, packaging, help pages, or app store listing, which hands the challenger its best exhibit.
• Launching a new category without a generic name, so the public has to use the brand as the noun. That is the aspirin and thermos story.
• Claiming 2(f) by reflex, conceding that the mark is not inherently distinctive when a suggestiveness argument was available.
• Proving the wrong thing: sales and advertising for a composite or house mark offered to show distinctiveness of a descriptive element (Louisiana Fish Fry), or ads that show a product design without directing attention to it (Yankee Candle).
• Running a do-it-yourself survey or website poll (Steelbuilding.com; the counsel-designed surveys excluded in Elliott).
• Ignoring competitors' generic use for years, then discovering that it is the centerpiece of a cancellation petition.
• Forgetting the relevant date: in litigation, secondary meaning must predate each infringer's first use (Converse).

Challengers make these:

• Dissecting a compound term and declaring it generic because each part is, instead of proving what the whole means to the public (Princeton Vanguard; In re American Fertility Society, 188 F.3d 1341 (Fed. Cir. 1999), which held that the Office must prove the meaning of a phrase as a whole and did not prove SOCIETY FOR REPRODUCTIVE MEDICINE generic).
• Defining the genus by the owner's advertising rather than by the goods identified in the registration (Magic Wand; Steelbuilding.com).
• Treating verb use as proof of genericide (Elliott).
• Missing the five-year window for descriptiveness challenges and then discovering Park 'N Fly.
• Choosing a genericness counterclaim when descriptive fair use would have done the job at a fraction of the cost.

Two worked examples

A software product name that is becoming a verb

Suppose Juniper Labs launched TALLYO in 2019, an app that splits group expenses and settles them through payment services, and registered the mark in 2020 for "downloadable software for tracking and splitting shared expenses." By 2026 the app has millions of users, people say "just tallyo me for the pizza," and a technology columnist refers to "the flood of tallyo apps." A competitor, Ledgerline, starts advertising itself as "the tallyo for teams." When Juniper objects, Ledgerline petitions the Board to cancel the registration under § 1064(3).

The petition will be decided on four questions. First, the genus is set by the registration: software for tracking and splitting shared expenses (Magic Wand). Second, "tallyo me" is evidence, not the test. Under Elliott, the question is whether the public primarily understands "tallyo" as the name of that category of software, and a user who says "tallyo me" usually means "pay me through Tallyo," which is brand use. Third, competitor usage will decide the case. Ledgerline's own tagline is self-serving, but if several other apps describe themselves as "a tallyo," the record starts to look like Schwan's. Fourth, Juniper's own house is not in order: its 2022 campaign said "Tallyo it!" and its help center explains "how to tallyo a bill." That is the Gould problem, and it must be fixed now, because the Board will measure public understanding on the record at trial, not as of the registration date.

Juniper's plan writes itself. It adopts usage rules ("the TALLYO app," "split it with TALLYO"), supplies a generic name ("expense-splitting app") and uses it everywhere, rewrites the help center, sends correction requests to the columnist and to app store listings that use "tallyo" for rival apps, and sends firm letters to the competitors. It commissions a Teflon survey from an independent expert, with a universe of people who have used an app to split expenses in the last year and controls that mix real brand names with common names such as "payment app" and "expense tracker." Ledgerline will answer with a Thermos-style survey ("what would you call an app that splits a restaurant bill?"), and Juniper's expert will point out that respondents who volunteer "Tallyo" mostly identify it as a brand on follow-up. On a record like that, with a strong brand percentage and little competitor use, Juniper should keep its registration, and the burden of proof (the registration's presumption puts it on Ledgerline) is a further advantage.

One more point favors Juniper. Ledgerline cannot fall back on descriptive fair use, because a coined word like TALLYO has no ordinary descriptive meaning to use "otherwise than as a mark"; its only non-brand meaning is the generic one Ledgerline is trying, and probably failing, to prove. If Ledgerline keeps using "the tallyo for teams," it is the infringer.

A furniture maker seeking 2(f) registration for a chair design

Suppose Holloway & Finch, a forty-person North Carolina workshop, has made the ARDEN lounge chair since 2009. Its signature is a steam-bent ash frame in which the rear legs sweep forward into the armrests in one unbroken curve. A design patent on the frame issued in 2011 and expired in 2025, and within months importers were selling close copies online, several listed as "Arden-style lounge chair." Holloway applies to register the frame configuration on the Principal Register under § 2(f) and plans to sue the largest importer.

The examining attorney will start where Wal-Mart requires: a product design is never inherently distinctive, so the application will be refused as a nondistinctive configuration unless the 2(f) showing succeeds (TMEP § 1202.02(b)(i)), and the examiner will ask functionality questions first. Those follow the factors in In re Morton-Norwich Products, Inc., 671 F.2d 1332 (C.C.P.A. 1982): is there a utility patent, does the owner's advertising tout utilitarian advantages, are alternative designs available, and is the design cheaper to make? Holloway has no utility patent, plenty of alternative lounge chair designs exist, and steam-bending is more expensive than straight-cut legs. But its catalog says the curve "cradles the lower back," and that sentence will be quoted back to it under TrafFix. It should stop using it.

On acquired distinctiveness, sixteen years of use is a good start and a five-year statement alone is not enough for a product design. The advertising record is mixed: for most of the chair's life the ads showed the chair and sold the HOLLOWAY & FINCH name, which Yankee Candle says is not enough, but since 2023 a "Follow the Arden line" campaign has shown the curve in silhouette, which is look-for advertising. Design magazines have called the curve "the Arden's signature." The importers' "Arden-style" listings are evidence of deliberate copying that trades on recognition, strong evidence in the Ninth Circuit after Jason Scott Collection and weaker in the Tenth Circuit after Craft Smith. And Holloway commissions a secondary meaning survey of recent buyers of chairs priced above $1,000, showing an unbranded photograph of the frame to a test cell and a straight-legged version to a control cell. A net association figure in the forties, combined with the advertising, media, and copying evidence, gives the examiner something to accept.

The importer will make two arguments worth anticipating. It will say that sixteen years of "exclusive" use proves only that the design patent kept competitors out, so sales during that period reflect the product's appeal rather than brand meaning. And it will invoke Converse: in the lawsuit Holloway must prove secondary meaning as of the importer's first sale in 2025, without help from a registration that will not issue until later, and as the owner of unregistered trade dress it bears the burden of proving nonfunctionality under 15 U.S.C. § 1125(a)(3). Holloway's best answers are the post-2023 look-for campaign, the survey, and the word mark: listings that call a knockoff "Arden-style" use Holloway's ARDEN mark, which is a simpler infringement claim. If the 2(f) claim fails, the Supplemental Register remains available for a nonfunctional configuration, and the next model should get its own design patent; the comparison is laid out in Design Patents vs. Trade Dress: Two Ways to Protect How a Product Looks. The final caution is the Fender lesson: had dozens of makers produced similar curved-leg chairs for decades, no amount of evidence would have helped.

Frequently asked questions

Can a generic term ever become a trademark? No. Evidence that the public associates a generic term with one seller (sometimes called de facto secondary meaning) does not make it protectable, because protecting it would let one company own the name of the thing. A stylized or design version can be registered, with the generic word disclaimed, only if the design creates a commercial impression separate from the word. The stylized BUNDT for cake mix in In re Northland Aluminum Products, Inc., 777 F.2d 1556 (Fed. Cir. 1985) did not, and the court added that once a term is found generic, evidence of secondary meaning cannot change the result.

People use our brand as a verb. Are we losing it? Not because of verb use alone. Elliott holds that verb use is evidence to be weighed, and that the question is whether the public uses the word as the name of the product category. What you should worry about is competitors and the press using your mark as the category name, and your own marketing doing the same.

Does a registration protect us from a genericness challenge? It helps but does not protect. A registration shifts the burden to the challenger, but a petition to cancel on genericness grounds can be filed at any time, and incontestability does not bar it (15 U.S.C. § 1065(4)).

How long does it take to acquire distinctiveness? There is no fixed period. The USPTO may accept five years of substantially exclusive and continuous use as prima facie proof for modestly descriptive terms. Eighteen months pointed "strongly away" from secondary meaning in Cicena, while an intense launch campaign built on look-for advertising can work faster than decades of quiet use.

Do we need a survey? Not as a matter of law; Coach Services and Yamaha both accepted 2(f) showings without one. As a practical matter, for highly descriptive terms, product designs, colors, and any contested litigation, the absence of a survey is a gap the other side will exploit.

What survey percentage is enough? A majority association is strong evidence. Lower figures can help alongside other proof (Zatarains upheld a finding built partly on 23 and 28 percent), and figures around ten percent cut against the claim (Real Foods). For genericness, Teflon brand percentages in the sixties and above, as in Booking.com and Elliott, are strong evidence for the owner.

Can adding .com make a generic term protectable? Sometimes. Booking.com rejected a rule that "generic.com" terms are always generic, but before it the Federal Circuit had held HOTELS.COM, MATTRESS.COM, and LAWYERS.COM generic on their records, as in In re 1800Mattress.com IP, LLC, 586 F.3d 1359 (Fed. Cir. 2009) and In re Reed Elsevier Properties Inc., 482 F.3d 1376 (Fed. Cir. 2007). After Booking.com the answer turns on evidence of consumer perception, and even a winning applicant gets only a narrow mark.

Is the Supplemental Register worth it? Usually, as a waypoint. It gives the ® symbol, a registration the USPTO will cite against later applicants, and a federal record of the claim. It gives no presumption of validity, cannot become incontestable, and does not stop a genericness challenge.

Practical takeaways

• Genericness and acquired distinctiveness are the same question asked from opposite ends: what does the relevant public understand the term to mean?
• A generic term can never be a mark, and a genericness challenge is never too late; descriptiveness challenges expire when a registration becomes incontestable.
• Define the genus from the identification of goods, and analyze compound terms as a whole.
• Verb use is evidence, not the test. Competitor use, owner use, and the existence of another common name usually decide genericide cases.
• Acquired distinctiveness must be proved for the matter actually claimed, at the right date, with evidence aimed at consumer perception: look-for advertising, contextualized sales, unsolicited media, copying, and a well-designed survey.
• The more descriptive the term, the heavier the burden; at the extreme, no evidence will do.
• Product designs and colors always need secondary meaning and must also clear functionality.
• Owners should treat brand usage discipline as evidence-building and keep the file; challengers should consider letters of protest early and descriptive fair use as a cheaper alternative to cancellation.

This document is general information, not legal advice for any particular situation; the law changes, and the facts matter.

Related Documents

Related Articles

• Trademark Overview: Substantive Standards for Protection: the distinctiveness spectrum and the other requirements for protection.
• Trademark Overview: The Subject Matter of Trademark Law: words, trade dress, colors, and other signs that can serve as marks.
• Trademark Fair Use: Descriptive Use, Nominative Use, and Expressive Works After Jack Daniel's: the defense that protects competitors' descriptive use of a mark with acquired distinctiveness.
• Expungement, Reexamination, and Letters of Protest: The Trademark Modernization Act in Practice: putting genericness evidence before an examiner before registration.
• Common-Law Rights, the Supplemental Register, and the Principal Register: A Practical Guide: what each register gives and withholds.
• Consumer Survey Expert Methodology in Trademark Cases: A Practical Guide: Teflon, Thermos, and secondary meaning survey design.
• Daubert Challenges to Consumer Survey Experts in Trademark Litigation: A Practical Guide: keeping your survey in and the other side's out.
• The Evolution of Trade Dress Protection for Product Design and Configuration: A Comprehensive Case Law Analysis: Two Pesos, Wal-Mart, TrafFix, and their progeny.
• Trademark Cancellation in Federal Litigation: A Practical Guide: using § 1119 counterclaims to cancel a registration in court.
• Trademark Licensing and Quality Control: How Naked Licensing Kills a Brand: the other way owners lose marks through their own conduct.
• Expert Depositions: Taking Apart the Opinion, Protecting Your Own Expert, and Building the Daubert Record: deposing and defending survey experts.

Related Guides

• Proving Acquired Distinctiveness and Defeating a Genericness Challenge: A Practical Guide: the step-by-step companion to this article.
• Asserting or Defeating a Trademark Fair Use Defense: A Practical Guide: the alternative to a genericness counterclaim.
• Design Patents vs. Trade Dress: Two Ways to Protect How a Product Looks: choosing protection for product designs.
• Maintaining Trademark Registrations: Section 8 and Section 15 filings, including incontestability.

Related Checklists

• Acquired Distinctiveness and Genericness Evidence Checklist: A Practical Checklist: every category of evidence, for both sides.
• Trademark Consumer Survey Design: A Practical Checklist: universe, controls, and questions.
• Responding to a Trademark Office Action Checklist: A Practical Checklist: answering § 2(e)(1) refusals on time.
• Selecting a Protectable Trademark Checklist: A Practical Checklist: avoiding the problem at the naming stage.
• Petitioning to Cancel a Trademark Registration: A Practical Checklist: the mechanics of a § 1064 petition.

Related Toolkits

• Distinctiveness Toolkit: Secondary Meaning Declarations, Genericness Surveys, and Brand Usage Guidelines: templates for the evidence and the usage rules discussed here.
• Trade Dress Protection Toolkit: From Product Design and Packaging to Registration, Functionality, and Enforcement: product configuration from filing to enforcement.
• TTAB Practice Toolkit: Oppositions, Cancellations, and the Trademark Trial and Appeal Board from Pleading to Appeal: running the Board proceeding where most genericness fights happen.
• Brand Licensing Toolkit: License Terms, Quality Programs, and Audit Rights: writing usage rules into licenses.