A trademark is a strange kind of property. You cannot pick it up, fence it off, or lock it in a vault. It exists only in the minds of consumers, as a learned association between a symbol and a source. That makes it both wonderfully durable—a well-tended mark can outlive every patent and copyright a company owns—and dangerously fragile, because the very act that creates a trademark (using it to identify your goods) is the act that, done carelessly, can destroy it. A patent dies on a schedule no owner can stop. A trademark dies only when its owner kills it.

This article is Part 3 of our four-part trademark overview. Part 1 asked what can be a trademark in the first place—words, logos, colors, sounds, and trade dress—and where the statutory bars draw the line. Part 2 examined the substantive standards a designation must meet to qualify for protection: distinctiveness, secondary meaning, use in commerce, and priority. Here in Part 3 we get practical. We cover the two halves of a brand owner's life: first, obtaining rights—how rights arise through use, how federal registration works from application to certificate, the gulf between the Principal and Supplemental Registers, and how to reach beyond U.S. borders through the Madrid Protocol—and second, exploiting rights through licensing, assignment, and financing, while avoiding the classic ways an owner can throw a perfectly good mark away. Part 4 then covers infringement and the related rights—likelihood of confusion, dilution, cybersquatting, and remedies.

If you read nothing else, read this: in trademark law, the goodwill is the thing. Registration certificates, license agreements, and assignment deeds are all just paperwork wrapped around an intangible asset—the reputation a mark carries in the marketplace. Almost every rule below, from the quality-control requirement to the prohibition on "assignments in gross," exists to keep the paperwork honest to that underlying reality. Keep your eye on the goodwill and the doctrine snaps into focus; lose sight of it and the rules look arbitrary.

For the cross-references: see Part 1 on what can be a mark, Part 2 on the substantive standards, and Part 4 on enforcement. For the plain-English on-ramps, Trademark Basics and The Trademark Process are good companions, and the Trademark Registration Guide is the practical walkthrough.

The Two Sources of Trademark Rights: Use and Registration

The single most important—and most counterintuitive—fact about American trademark law is this: you do not need to register a trademark to own one. Unlike a patent, which springs into existence only when the U.S. Patent and Trademark Office grants it, a trademark right is born the moment you use a distinctive mark in commerce to identify the source of your goods or services. The law calls these unregistered rights "common law" rights, and they are real, enforceable property—you can sue on them, license them, and pass them to your heirs.

This is the use-based system, and it sets the United States apart from most of the world. In a great many countries, trademark rights flow from registration on a first-to-file basis: whoever runs to the trademark office first generally wins, even if a competitor has been using the mark for years. The American rule is, broadly, first to use. Priority—the legal rank that decides who beats whom in a conflict—generally goes to the senior user, the party that used the mark in commerce first, not to whoever filed an application first. (Part 2 of this series develops priority in detail; see Trademark Overview--Substantive Standards for Protection. The classic doctrine sorting out two innocent users in different regions is the Tea Rose-Rectanus doctrine, which we examine through the lens of Stone Creek v. Omnia.)

So why does anyone bother to register? Because common law rights, while real, are limited and hard to prove. They reach only as far as the geographic area where the mark is actually used and known, plus a modest natural "zone of expansion." A bakery in Tucson that has never registered its name has rights in Tucson and perhaps the surrounding region—but it cannot stop an unrelated, good-faith bakery from opening under the same name in Portland. Common law rights also leave the owner with the burden of proving, in any dispute, exactly when use began, how far it spread, and that the mark is distinctive. That is expensive, fact-intensive litigation; the documentary spadework alone is daunting (see Proving Trademark Use and Priority).

Federal registration solves much of this. It is a bargain the Lanham Act offers: bring your mark to the USPTO, survive examination, and receive a bundle of powerful procedural and substantive advantages that common law use alone can never confer. We catalog those benefits below. But hold onto the foundational point, because it echoes throughout the rest of this article: registration confirms and strengthens rights; it does not create them. Even an expired or cancelled registration leaves the owner's underlying common law rights intact, because those rights came from use, not from the certificate. See Trademark Rights Under Common Law for a fuller treatment.

State law sits underneath all of this

Trademarks live in a layered legal world. At the federal level, the governing statute is the Lanham Act, 15 U.S.C. §§ 1051–1141n, administered by the USPTO, with the agency's regulations at 37 C.F.R. §§ 2.2–7.41 and its day-to-day examining standards collected in the Trademark Manual of Examining Procedure (TMEP). But underneath the federal layer sits a parallel system of state trademark law: every state recognizes common law unfair-competition principles, and most states maintain their own trademark statutes and registers. A state registration is generally available only for a mark already used in commerce within that state; it offers some procedural and evidentiary advantages in state-court litigation—but it is no substitute for federal registration, and it does not confer nationwide rights. For any business with ambition beyond a single locality, the federal register is the real prize. State registration is best understood as a low-cost supplement, not a strategy.

Before You File: Selecting and Clearing a Mark

Obtaining protection begins long before any form is submitted. It begins with two decisions that quietly determine whether the whole enterprise will succeed: which mark to adopt, and whether anyone else already owns it.

Choosing a strong, registrable mark

Not every mark is created equal, and not every mark is even capable of protection. Part 1 of this series lays out the "distinctiveness spectrum" in full (see Trademark Overview--The Subject Matter of Trademark Law), but the short version matters enormously for the application stage. Marks fall along a continuum from generic (the common name of the product—"Email" for email, never protectable), through descriptive (describing a feature—"Creamy" for yogurt, protectable only after it acquires secondary meaning), to suggestive, arbitrary, and fanciful (the inherently distinctive categories the law protects most readily and registers most easily).

The practical counsel here is the same counsel a seasoned trademark lawyer gives every excited founder: the marketing instinct and the legal instinct point in opposite directions. Marketing loves a descriptive name because it tells the customer instantly what the product does. Trademark law disfavors exactly that name, because a mark that merely describes the goods is weak, hard to register, and easy for competitors to crowd. The strongest marks—think KODAK (fanciful, an invented word) or APPLE for computers (arbitrary, a real word with no connection to the goods)—say nothing about the product at all, which is precisely why the law gives them the broadest protection. A coined or arbitrary mark is, in the long run, the cheapest insurance a brand can buy: it costs a little imagination up front and saves a fortune in enforcement and refusals later. When the impulse to be descriptive is strong, raise the tension early; see When to Trademark Your Brand and Trademarking Your Own Name for the practical angles.

When you file, you will also fix two things that frame the entire application:

  • Mark format. You must declare whether you are registering a standard character mark (the words themselves, in any font, size, or color—the most flexible protection), a stylized or design mark (a particular logo or visual presentation), or a non-traditional mark such as a sound or color mark. A standard-character registration protects the words however displayed; a design registration protects the specific design. Many owners file both, to cover the name and the logo independently.
  • Identification of goods and services. You must describe, precisely, the goods and services the mark will identify. This is not a formality—it defines the scope of your rights, your fees, and the universe of conflicting marks you must clear against. Overbroad or vague descriptions draw refusals; the USPTO's online ID Manual of pre-approved descriptions is the safest starting point. Because goods and services are sorted into 45 international classes and you pay per class, the identification also drives cost; see USPTO Trademark Classes--A Guide to the Nice Classification.

Clearance: search before you leap

The second pre-filing task is clearance—investigating whether your chosen mark conflicts with someone else's prior rights. This is where a great many brand-building projects come to grief, because the senior user's rights (recall: first to use) can defeat your registration and, worse, expose you to an infringement suit even after you have sunk a fortune into packaging, signage, and advertising. Adopting a mark without clearing it is the trademark equivalent of building a house without checking the title.

A proper clearance program proceeds in stages. It usually starts with a quick "knock-out" search—a preliminary scan of the USPTO database and obvious sources to weed out blatant, identical conflicts before more money is spent. If the mark survives that, the owner commissions a full availability search, a comprehensive review covering federal registrations and applications, state registrations, common law uses (business names, domain names, social media handles, industry directories), and sometimes design and phonetic-equivalent searches. Counsel then analyzes the results—not just identical hits but confusingly similar marks on related goods, which is the standard the USPTO and the courts actually apply.

Where the search turns up a worrying mark, the owner has options short of abandonment: investigating whether the other party is actually still using its mark, negotiating a consent or coexistence agreement that carves out non-overlapping lanes of use, or, in the right case, petitioning to cancel a conflicting registration that has gone unused. This is detailed work with its own dedicated treatment; see How to Conduct a Comprehensive Trademark Clearance Search. The payoff is twofold: you avoid adopting a mark you cannot keep, and a documented, attorney-supervised clearance search can later serve as a powerful shield against a claim of willful infringement, as discussed in The Shield of Good Faith.

Worked example (hypothetical). Acme Roasters, a coffee startup, falls in love with "MORNING GLORY" for its beans. A knock-out search reveals an existing federal registration for "MORNING GLORY" covering tea, plus a common law café in another state using "Morning Glory Coffee." Counsel advises that tea and coffee are "related goods" likely to draw a Section 2(d) likelihood-of-confusion refusal, and that the café is a senior user with regional rights. Acme has three realistic paths: pick a different name (cheapest in the long run), seek a coexistence agreement with the tea company and the café (possible but slow, and never guaranteed), or proceed and gamble (the worst option). Good clearance turned a potential six-figure rebrand-under-fire into a five-figure naming decision made calmly at the outset.

The Federal Application and Examination Process, Step by Step

Suppose the mark is strong and the clearance is clean. Now you file. The federal process is a defined pipeline with predictable stages, and understanding the map prevents missed deadlines—which, in trademark practice, are frequently fatal. A companion piece walks through the mechanics screen by screen; see How to File a Trademark Application with the USPTO and the step-by-step Federal Trademark Application Checklists and USPTO Trademark Application Checklists.

Do you need an attorney?

One threshold rule reshaped the landscape and surprises many applicants. Since 2019, any applicant whose domicile is outside the United States or its territories must be represented by a U.S.-licensed attorney before the USPTO (37 C.F.R. § 2.11). The rule was the agency's response to a flood of fraudulent and improper foreign-origin filings. U.S.-domiciled applicants are not required to use counsel, but—given the deadlines, the substantive refusals, and the per-class costs of getting it wrong—the agency itself "strongly encourages" hiring a trademark attorney. The cost of competent counsel is small next to the cost of a botched identification or a missed Statement of Use deadline.

Choosing a filing basis

Every U.S. application rests on a filing basis, and the most common two are foundational:

  • Use in commerce—Section 1(a) (15 U.S.C. § 1051(a)). This basis is for a mark already in use in interstate commerce. The applicant must specify the dates of first use and submit a specimen—real-world evidence of the mark as actually used in the marketplace (a product label, packaging, a webpage where the goods can be purchased). "Use in commerce" means bona fide use in the ordinary course of trade, not token use manufactured just to support a filing (15 U.S.C. § 1127). The USPTO has grown notably stricter about specimens, rejecting mockups, digitally altered images, and "ornamental" uses that do not function as source identifiers.

  • Intent to use—Section 1(b) (15 U.S.C. § 1051(b)). This basis is for a mark the applicant has a bona fide intention to use but has not yet used. The intent-to-use ("ITU") application, added by the Trademark Law Revision Act of 1988, lets a business stake an early claim before launch—a major strategic advantage, because the eventual filing date can establish constructive use priority back to the application date. But an ITU applicant cannot obtain a registration on intention alone; the mark must actually be put into use, proven by a Statement of Use, before the registration issues. The "bona fide intent" requirement has teeth: it must be a genuine, documented intention to use the mark in the ordinary course of trade, not a mere wish to reserve a name. Part 2 develops the bona-fide-intent standard further.

Two other bases serve specific populations: Section 44 for applicants relying on a foreign application or registration under the Paris Convention, and Section 66(a) for requests to extend an international registration to the United States through the Madrid Protocol (discussed below). One important constraint: an application to register a mark on the Supplemental Register must be based on actual use, not intent to use—because the Supplemental Register exists for marks not yet distinctive enough to serve as source identifiers, and intent alone cannot demonstrate the capacity to distinguish.

Filing and the formalities

Applications are filed electronically through the USPTO's Trademark Center (the successor to the legacy TEAS forms). The filing fee is a processing fee—nonrefundable, and owed whether or not a registration ever issues. (Nearly everything submitted becomes part of the public record, including the applicant's name and address, so plan accordingly.) In January 2025 the USPTO overhauled its fee structure, retiring the old TEAS Plus / TEAS Standard tiers in favor of a single base application fee per class, with surcharges for using free-form (non-ID-Manual) goods-and-services descriptions and for unusually long descriptions. The practical lesson of the new structure is the same lesson the ID Manual has always taught: use the pre-approved language and keep your descriptions tight, or pay extra. Fees change, so confirm current amounts on the USPTO site before filing.

After filing, the single most valuable habit an applicant can develop is monitoring. Check the application's status through the Trademark Status and Document Retrieval (TSDR) system every three to four months. The USPTO communicates by issuing documents with hard deadlines, and a missed deadline ordinarily means abandonment. Diligence here is not optional housekeeping; it is the difference between a registration and a dead file.

USPTO examination

Once the application clears minimum filing requirements, it is assigned a serial number and routed to an examining attorney—a USPTO lawyer who scrutinizes the application for two things: compliance with the formal requirements, and the substantive eligibility of the mark for registration. The examiner runs a search for conflicting marks and reviews the written application, the drawing, and any specimen. This takes time; examination backlogs mean the first action often arrives many months after filing.

If the examiner finds a problem, the applicant receives an Office Action—a letter explaining each refusal or requirement. Office Actions come in two flavors. Procedural objections are technical: a deficient specimen, an unclear identification, a missing disclaimer of unregistrable matter. Substantive refusals go to whether the mark can be registered at all. The most common substantive refusal by far is Section 2(d)—likelihood of confusion with a prior registered or pending mark—followed by Section 2(e) refusals for marks that are merely descriptive, geographically descriptive, or primarily merely a surname. (Two famous other Section 2 bars—the prohibitions on "immoral or scandalous" marks and on "disparaging" marks—were struck down as unconstitutional under the First Amendment in Iancu v. Brunetti, 139 S. Ct. 2294 (2019), and Matal v. Tam, 137 S. Ct. 1744 (2017). The road to Tam runs through the saga recounted in Washington Redskins Trademark Troubles.)

The applicant generally has three months to respond to an Office Action, with a single three-month extension available for a fee. (This shortened window replaced the old flat six-month deadline under a rule that took effect in December 2022 for most applications—yet another reason the calendar matters.) Responding to refusals is its own craft: amending the identification, submitting substitute specimens, arguing against a 2(d) refusal by distinguishing the goods or the marks, or—where a descriptive-mark refusal is in play—claiming acquired distinctiveness under Section 2(f). Where a 2(d) refusal rests on a prior registration, the applicant can sometimes clear the path with a consent agreement from the prior owner. The strategic toolkit overlaps with patent prosecution; for the cousin discipline, see Responding to Patent Office Actions.

If the examiner is not satisfied, a final refusal issues. At that point the applicant who wants to fight on may appeal to the Trademark Trial and Appeal Board (TTAB), the USPTO's administrative tribunal. The TTAB also hosts oppositions and cancellations—proceedings that look and feel a lot like federal litigation, complete with discovery; for that flavor see Discovery Practice in TTAB Trademark Proceedings.

Publication and opposition

If the examiner raises no objection, or every objection is overcome, the mark is approved for publication in the Trademark Official Gazette, the USPTO's weekly publication. (Marks bound for the Supplemental Register are not published.) Publication opens a 30-day window in which any party who believes it would be damaged by the registration may file a Notice of Opposition—or request more time to oppose—before the TTAB. An opposition is the marketplace's chance to object before the certificate issues; it is litigation in miniature, and it is one more reason a clean clearance search matters so much, because a well-resourced senior user watching the Gazette will pounce on a conflicting mark.

If no one opposes, or the opposition fails, the next step depends entirely on the filing basis—and here the two main paths diverge.

The two roads to a certificate

For a use-based application (Section 1(a)), and for applications based on a foreign registration (Section 44) or a Madrid extension (Section 66(a)), the road is short: after the opposition period passes with no successful challenge, the USPTO registers the mark and issues a certificate of registration. Use was already proven up front, so nothing remains but to issue the registration and start the post-registration maintenance clock.

For an intent-to-use application (Section 1(b)), there is one more crucial hurdle, because the applicant has not yet shown actual use. Instead of a certificate, the USPTO issues a Notice of Allowance—roughly eight weeks after publication. A Notice of Allowance is not a registration; it is an invitation to finish the job. From its date, the applicant has six months to either:

  1. begin using the mark in commerce and file a Statement of Use (SOU), with a specimen and a fee; or
  2. file a Request for an Extension of time to file the SOU.

The applicant may file successive extension requests, but the law caps the total: no more than five extensions, for a maximum of three years from the Notice of Allowance, after which the application dies if use has not begun. Miss the deadline entirely and the application is abandoned—though a petition to revive is available within two months of the abandonment date if the failure was unintentional. When the SOU is filed, the examiner reviews it (and can issue further Office Actions on the same response cycle). If the use is acceptable and no objection remains, the registration issues, typically within about two months.

The ITU mechanism is one of the most useful tools in trademark practice, but it is also one of the most deadline-laden. A docketing failure between Notice of Allowance and Statement of Use is among the more common—and most avoidable—ways to lose a mark a business has spent years preparing to launch.

The Principal Register and the Supplemental Register

The USPTO maintains two federal registers, and understanding the difference is essential, because they confer very different rights. Treating them as interchangeable is a classic novice error.

The Principal Register is the main event. It is reserved for marks that are either inherently distinctive (suggestive, arbitrary, and fanciful) or that have acquired distinctiveness—secondary meaning—through use and promotion. Registration on the Principal Register confers the full slate of Lanham Act benefits, which is why it is the goal of nearly every serious filing:

  • A legal presumption of ownership of the mark and of the registrant's exclusive right to use it nationwide on the listed goods and services (15 U.S.C. §§ 1057(b), 1115(a)). This shifts the burden of proof in litigation—an enormous advantage that turns the registrant's evidentiary burden into the challenger's.
  • Constructive nationwide use as of the application's filing date (15 U.S.C. § 1057(c)), which extends priority across the whole country regardless of where the owner has actually used the mark, cutting off later good-faith adopters in remote regions—the very gap that limits common law rights.
  • Constructive notice to the world of the registrant's claim (15 U.S.C. § 1072), which forecloses the "good faith" defense a junior user might otherwise raise.
  • The right to use the federal registration symbol ® (using it without a registration can itself be a problem).
  • A basis for federal-court jurisdiction, the ability to record the registration with U.S. Customs and Border Protection to block infringing imports, a hook for fighting cybersquatters, and—after five years of continuous use—a path to incontestability (discussed below).

The Supplemental Register is the consolation prize, and a genuinely useful one. It exists for marks that are not yet distinctive enough for the Principal Register—typically descriptive marks that have not yet acquired secondary meaning—but that are nonetheless capable of distinguishing the applicant's goods or services. A Supplemental registration is modest: it confers the right to use the ® symbol, places the mark in the USPTO's searchable database (so it can block later confusingly similar applications), and gives the owner access to federal court. But it does not carry the Principal Register's evidentiary presumptions—no presumption of validity, ownership, or exclusive nationwide rights—and a Supplemental registration can never become incontestable. It is, however, a strategic foothold: if a Supplemental-registered mark later acquires distinctiveness through years of use and promotion, the owner can file a new application to register the now-stronger mark on the Principal Register. Many descriptive brands begin on the Supplemental Register and graduate to the Principal Register once consumers come to recognize them as source identifiers.

Worked example (hypothetical). Brightline Analytics adopts "DATACLEAR" for data-cleaning software. The examiner refuses Principal registration under Section 2(e)(1) as merely descriptive—"DATACLEAR" describes software that clears up data. Brightline cannot yet prove secondary meaning because the brand is new. Rather than abandon the application, Brightline amends to the Supplemental Register, secures the ® symbol and a database entry that will block confusingly similar later filers, and keeps building the brand. Four years and a lot of advertising later, surveys show consumers associate "DATACLEAR" specifically with Brightline. It files a fresh Principal Register application claiming acquired distinctiveness under Section 2(f)—and graduates.

Reaching Across Borders: The Paris Convention and the Madrid Protocol

Trademark rights are territorial. A U.S. registration protects a mark in the United States and nowhere else. A company selling—or planning to sell—internationally must secure rights country by country, and the two treaties that smooth the way are central to any global brand strategy. (For the bigger picture, see International Trademark Protection within a broader online strategy.)

The Paris Convention for the Protection of Industrial Property gives an applicant a priority right: file in one member country, and you have six months to file in other member countries while claiming the original filing date as your effective date in each. That six-month head start can be decisive in first-to-file jurisdictions, where a single day's delay can hand a squatter your name.

The more powerful tool is the Madrid Protocol, administered by the World Intellectual Property Organization (WIPO). Madrid lets a trademark owner file a single international application, based on a "home" application or registration (for U.S. owners, a U.S. application or registration filed through the USPTO), and then designate any number of the 100-plus Madrid member countries in which protection is sought—all through one filing, in one language, with one set of fees, managed through one central WIPO registration. Compared with hiring local counsel to file separately in each country, it is a tremendous administrative simplification.

Madrid comes with trade-offs every practitioner must flag. First, central attack: for the first five years, the international registration depends on the underlying home (basic) application or registration. If the home filing is refused, cancelled, or abandoned within that period, the international registration—and every national designation hanging from it—falls with it. (After five years the international registration becomes independent. And even after a central attack, the owner can "transform" the dropped designations into separate national applications, preserving the priority date, at additional cost.) Second, each designated country still examines the extension under its own law: Madrid is a filing convenience, not a guarantee of registration, and a designation can be refused in any country for that country's own substantive reasons. When the United States is the designated country (an incoming Section 66(a) extension), the USPTO examines it like a domestic application, and the resulting registration has its own renewal track under Section 71 rather than the usual Section 8.

For most growing brands, the strategic question is not whether to think internationally but when—and the Madrid Protocol, layered on a solid domestic U.S. registration, is the workhorse that makes affordable multinational protection possible.

Keeping What You Earned: Maintenance and Incontestability

Obtaining a registration is the beginning, not the end. A federal registration is not a "set it and forget it" asset; it must be actively maintained, and the maintenance regime is unforgiving of missed deadlines. This article touches the essentials; the dedicated treatment is Maintaining Trademark Registrations.

A registration can endure forever—in perpetual, renewable ten-year terms—but only if the owner does two things: keeps using the mark, and files the required maintenance documents on time. Specifically:

  • A Declaration of Use under Section 8 must be filed between the fifth and sixth year after registration, attesting (with a specimen) that the mark is still in use, or that any nonuse is excusable. This is the first big checkpoint, and it culls a large share of registrations whose owners have stopped using the mark or stopped paying attention.
  • A combined Section 8 Declaration and Section 9 Renewal must be filed between the ninth and tenth year, and again every ten years thereafter, to keep the registration alive.
  • For Madrid-based Section 66(a) registrations, the equivalent ongoing declaration is filed under Section 71.

Miss these deadlines (there are short grace periods, for extra fees) and the registration is cancelled or expires—permanently. There is no revival; the only remedy is to start over with a brand-new application, with no guarantee the mark will register a second time. Because of recurring concerns about "deadwood"—registrations covering goods the owner no longer sells—the USPTO runs a post-registration audit program that randomly demands additional proof of use, and the agency has steadily tightened proof-of-use scrutiny as part of a broader campaign to clean up the register. The lesson is blunt: claim only the goods and services you actually use, and keep your specimens honest.

Incontestability: the gold standard

After five consecutive years of continuous use of a Principal Register mark following registration, the owner may file a Declaration of Incontestability under Section 15 (15 U.S.C. § 1065). Incontestability is the strongest status American trademark law offers: it makes the registration conclusive evidence of the owner's exclusive right to use the mark on the listed goods and services, subject only to a short list of statutory exceptions (such as genericness, functionality, fraud, and abandonment). An incontestable mark cannot be challenged on the ground that it is "merely descriptive"—a defense that would otherwise be available against many registrations. The Supreme Court confirmed in Park 'N Fly, Inc. v. Dollar Park & Fly, Inc., 469 U.S. 189 (1985), that incontestable status can be wielded offensively, as a sword in an infringement suit, not merely defensively. The Supplemental Register, by contrast, can never reach incontestable status—one more reason the Principal Register is the goal.

Exploiting the Mark: Licensing, Selling, and Borrowing Against It

A trademark is property, and like other property it can be put to work. The three principal ways to monetize a mark are licensing it (renting out the right to use it), assigning it (selling it outright), and using it as collateral (borrowing against it). Each is governed by special rules that exist, once again, to protect the goodwill the mark represents. Get the rules wrong and you do not merely make a bad deal—you can destroy the mark itself. This is where trademark monetization diverges sharply from its cousins; for the very different world of patent deals, compare How to License Your Patent--From Valuation to Term Sheet, and for the software analog, Drafting Software License Agreements.

Trademark licensing and the cardinal rule: quality control

A trademark license is permission to use a mark, granted by the owner (the licensor) to another party (the licensee) on agreed terms—exclusivity, territory, duration, fees, royalties, and, above all, quality standards. Licensing is everywhere in the modern economy: franchises, brand extensions, character merchandising, co-branding, and the "intercompany" licenses that let a corporate group hold its marks in one entity and license them to operating affiliates. A trademark license can be recorded with the USPTO, but recordation is not required for the license to be valid.

Here is the rule that governs everything: a trademark licensor must exercise reasonable control over the nature and quality of the goods or services the licensee provides under the mark. This is the quality-control requirement, and it is not a drafting nicety—it is structural to what a trademark is. A trademark guarantees consumers a consistent source and a consistent standard. When an owner licenses the mark to others but exercises no control over what they do with it, the mark stops guaranteeing anything; the public is deceived into thinking quality is consistent when it is not. The law's response is severe.

A license without adequate quality control is called a "naked license," and granting one can constitute abandonment of the mark—not just as to that one licensee, but potentially across the board, destroying the owner's rights against the whole world. The Lanham Act supplies the mechanism: a mark is abandoned when, "through conduct of the owner," it "loses its significance as a mark" (15 U.S.C. § 1127), and uncontrolled licensing is the textbook example.

Three cases anchor the doctrine, and every practitioner should know them:

  • Dawn Donut Co. v. Hart's Food Stores, Inc., 267 F.2d 358 (2d Cir. 1959), is the foundational authority. The Second Circuit explained that the Lanham Act, by defining "related company" use to require that the owner control the "nature and quality" of the goods, places "an affirmative duty upon a licensor . . . to take reasonable measures to detect and prevent misleading uses of his mark by his licensees." A licensor who tolerates uncontrolled use risks losing the mark, because the public would be deprived of its assurance of consistent quality. (Confusingly, Dawn Donut is also the source of the unrelated "Dawn Donut rule" on geographically remote, non-competing use under registered-mark priority—but for our purposes it is the quality-control case.)

  • Barcamerica International USA Trust v. Tyfield Importers, Inc., 289 F.3d 589 (9th Cir. 2002), is the modern cautionary tale and a vivid teaching case. The owner of the "Leonardo Da Vinci" wine mark licensed it but exercised essentially no control over the licensee's wine—it relied on the winemaker's reputation, conducted no inspections, sampled the wine only casually, and could point to no contractual quality standards it actually enforced. The Ninth Circuit held the license "naked" and found the mark abandoned, stripping the owner of its rights. The lesson is the whole doctrine in a sentence: where a licensor "fails to exercise adequate quality control over the licensee," the trademark "ceases to function as a symbol of quality and a controlled source." A bare contractual right to inspect is not enough; the licensor must show it actually controlled quality.

  • Eva's Bridal Ltd. v. Halanick Enterprises, Inc., 639 F.3d 788 (7th Cir. 2011), is the sharpest articulation of all, and the most quotable. Judge Easterbrook, writing for the Seventh Circuit, confronted a family that had licensed the "Eva's Bridal" name to relatives but imposed no quality standards whatsoever. The opinion cuts to the bone: the licensor "had no standards" and exercised "no control," and "the lack of control means that 'Eva's Bridal' had become a naked license—and that the trademark's owner had abandoned the mark." Easterbrook's framing is the one to remember: the question is not whether a licensor controls quality in some idealized sense but whether it has any standards at all and enforces them. "How much control is enough?" he asked, and answered that there is no precise formula—but zero is never enough.

The practical takeaways for any licensing program are concrete. The license agreement should set explicit quality standards for the goods or services; reserve the licensor's right to inspect, sample, and approve; and—critically—the licensor must actually exercise those rights, keeping records that prove it. Courts permit a licensor to rely on a trusted licensee's own quality controls in narrow circumstances (the Ninth Circuit recognized as much in FreecycleSunnyvale v. Freecycle Network, 626 F.3d 509 (9th Cir. 2010), even as it found the reliance there inadequate to save the mark), but they treat such reliance skeptically. The safest course is to draft strong controls and use them. A well-drafted "pro-licensor" agreement builds the quality-control machinery into the deal; the worst-drafted is the friendly handshake that nobody ever revisits.

Worked example (hypothetical). Heritage Brands owns the registered mark "OAKMONT" for premium leather goods and licenses it to three regional manufacturers. For two of them, Heritage sets written specifications (leather grade, stitching standards, defect tolerances), audits factories twice a year, and approves samples before each season. For the third—a friend's company—Heritage signs the same paper but never inspects anything, never reviews a sample, and looks the other way as quality slides. Years later, when Heritage sues an infringer, the defendant digs up the third license in discovery and argues naked licensing and abandonment of the entire "OAKMONT" mark. The two well-controlled licenses do not save Heritage; the question is whether the uncontrolled one caused the mark to lose its significance. Heritage's casual third license—paper without practice—has put its whole brand at risk. The fix would have cost a few site visits a year.

The hidden trap: the "accidental franchise"

There is a second danger in trademark licensing that sophisticated companies routinely overlook, and it cuts in the opposite direction from naked licensing. Exercise too much control over a licensee's business, charge a fee, and weave the licensee into a unified marketing system, and you may have created a franchise—whether you meant to or not, and whether or not your contract says "this is not a franchise."

Under federal and most state franchise laws, a franchise exists when three elements coexist: (1) a trademark license; (2) a marketing system or "community of interest"—the licensor's significant assistance with or control over the licensee's method of operation, such that the public regards the outlets as a unified concept; and (3) a required fee. Crucially, a contract clause disclaiming franchise status is not dispositive; courts look to oral promises, course of dealing, and the parties' actual practices (see, e.g., Purugganan v. AFC Franchising, LLC, 2021 WL 723916 (D. Conn. Feb. 24, 2021); Unlimited Prepaid, Inc. v. Air Voice Wireless LLC, 2018 WL 6303852 (C.D. Cal. 2018)). The list of companies caught by surprise is long and distinguished—even the national Girl Scouts organization was held subject to a state dealership statute when it tried to terminate a local chapter (Girl Scouts of Manitou Council, Inc. v. Girl Scouts of the U.S.A., Inc., 549 F.3d 1079 (7th Cir. 2008)).

The accidental franchise is a genuine dilemma, because quality control is exactly the kind of operational involvement that helps establish a franchise relationship. The trademark law wants you controlling the licensee; the franchise law penalizes you for controlling too much. The resolution is to calibrate: exercise the quality control the trademark law demands (standards plus enforcement directed at the goods or services), while avoiding the broader operational dominion, prescribed marketing plans, and mandatory fees that trip the franchise wire—or, if a franchise is unavoidable, comply with franchise-disclosure and registration laws on purpose. Counsel structuring a serious licensing program should run the franchise analysis deliberately rather than discover it in litigation after a terminated licensee cries foul.

Selling the mark: assignments and the anti-assignment-in-gross rule

An owner can also sell a mark outright through an assignment. Both registered and unregistered marks can be assigned, and an assignment of a registered mark must be in writing (15 U.S.C. § 1060). Recording the assignment with the USPTO is not required to make the transfer effective, but it is strongly advisable, because recordation gives notice and protects the buyer against later claims by a subsequent purchaser.

But trademark assignments carry a rule with no real analog in patent or copyright law: an assignment must transfer the mark together with the goodwill the mark symbolizes. A purported transfer of the mark alone, stripped from its goodwill, is an "assignment in gross"—and it is invalid. The buyer in an assignment-in-gross acquires nothing; worse, the seller may be deemed to have abandoned the mark by severing it from the business and goodwill that gave it meaning. The reason traces straight back to the opening theme: the mark is not a thing in itself; it is a symbol of the goodwill, and a symbol cut loose from what it symbolizes is empty.

The deeper doctrine asks whether the assignee continues to use the mark on substantially the same goods or services, so that consumers' settled expectations about what the mark stands for are not disrupted. In Sugar Busters LLC v. Brennan, 177 F.3d 258 (5th Cir. 1999), the Fifth Circuit explained that an assignment is in gross—and invalid—if the assignee uses the mark on a "different good[]" such that the public would be deceived, because the goodwill did not actually travel with the mark. The Second Circuit applied the same logic in Marshak v. Green, 746 F.2d 927 (2d Cir. 1984), where a purported sale of the rights to a musical group's name without the accompanying performers or business was held an invalid assignment in gross. The point is not that the paperwork must recite the magic words "together with the goodwill" (though it always should); it is that the substance of the transfer must carry the goodwill over, typically by transferring the associated business assets or by the assignee continuing the same line of goods or services.

Two further wrinkles matter. First, intent-to-use applications are subject to a special anti-trafficking rule (15 U.S.C. § 1060(a)(1)): an ITU application generally cannot be assigned before the applicant has filed its Statement of Use (before actual use begins), except to a successor to the applicant's ongoing and existing business to which the mark pertains. Congress added this restriction to stop the warehousing and sale of mark "reservations" that were never tied to a real business; an improper assignment of an ITU application can invalidate the resulting registration. Second, recall the practical interplay with licensing: a sloppy assignment, like a sloppy license, can result in abandonment. The transactional discipline here is exacting, which is why even short-form assignments are drafted with care.

Worked example (hypothetical). Nimbus Software sells its "SKYVAULT" cloud-storage brand to Acme Corp. for a tidy sum. The contract assigns "the trademark SKYVAULT and Registration No. ___" and stops there—no goodwill language, no transfer of customer lists, codebase, or the storage business itself. Acme then slaps "SKYVAULT" on an unrelated line of garden hoses. Later, when Acme tries to enforce the mark, a competitor argues the original sale was an assignment in gross: the mark was severed from its goodwill and applied to entirely different goods, deceiving consumers who knew SKYVAULT as cloud storage. A court could well find the assignment invalid and the mark abandoned—a total loss of an asset Acme paid real money for. Two fixes would have saved the deal: assigning the goodwill and underlying business assets along with the mark, and Acme continuing to use SKYVAULT on cloud storage rather than hoses.

Using the mark as collateral

Finally, a valuable mark can be pledged as collateral to secure financing—an increasingly common move as intangible assets dominate corporate balance sheets. Here the law gets technical, and the technicality trips up even sophisticated lenders.

Start with the division of labor. Creation and attachment of a security interest in a trademark is governed by state UCC Article 9: a trademark is a "general intangible," and the interest attaches in the ordinary way through a security agreement that grants the interest and adequately describes the collateral. Perfection—the step that establishes the lender's priority over other creditors—is also governed by Article 9, accomplished by filing a UCC-1 financing statement in the right state office. The federal trademark statute, unlike the Copyright Act, contains no provision addressing the creation or perfection of security interests, so courts have held it does not preempt the UCC on this point. Recordation under the UCC is therefore sufficient to perfect a security interest in a federally registered trademark or pending application.

So why record anything with the USPTO at all? Because USPTO recordation provides protection against subsequent bona fide purchasers and mortgagees of the mark, and it keeps the agency's chain-of-title records straight if the mark is later sold or licensed. The belt-and-suspenders best practice is therefore to do both: file the UCC-1 to perfect, and record with the USPTO to guard against good-faith buyers. One drafting nuance from practice: lenders typically record a separately executed short-form security agreement (not the full loan documents, and not a mere copy of the UCC-1) with the USPTO, because the recorded instrument must specifically identify each registration or application encumbered, and filing the full agreement would needlessly disclose the entire deal on a public record. A lender that records only with the USPTO and skips the UCC-1 may find itself unperfected and subordinated in the borrower's bankruptcy—a costly mistake that flows directly from misunderstanding which sovereign governs which step.

Three Ways to Lose a Mark

We have met the loss doctrines in passing, but they deserve a clear, consolidated statement, because the most common way brand owners lose marks is not through some clever competitor's attack—it is through their own conduct. Three routes are the heavy hitters.

First, abandonment through nonuse. A U.S. trademark must be used to survive. Under 15 U.S.C. § 1127, a mark is abandoned when its use is discontinued with intent not to resume use. Both elements are required—discontinuance and intent not to resume—but the statute provides a crucial evidentiary shortcut: three consecutive years of nonuse is prima facie evidence of abandonment, shifting the burden to the owner to prove an intent to resume use within the reasonably foreseeable future. Vague hopes of someday reviving a brand do not suffice; the owner must show concrete plans. The Second Circuit drove the point home in Silverman v. CBS, Inc., 870 F.2d 40 (2d Cir. 1989), holding that CBS had abandoned the "Amos 'n' Andy" radio marks after roughly two decades of deliberate nonuse, because an indefinite intention to keep a property "warehoused" against some hypothetical future revival is not an intent to resume use within the meaning of the statute. This is why "trademark maintenance" is not just paperwork—it is actually using the mark.

Second, abandonment through uncontrolled (naked) licensing, covered in detail above. When an owner licenses without quality control, the mark "loses its significance as a mark" and is abandoned. Dawn Donut, Barcamerica, and Eva's Bridal are the touchstones.

Third, genericide. A mark can be loved to death. If a mark's primary significance to the relevant public shifts from identifying a source to naming the product category itself, the mark becomes generic and protection evaporates. The graveyard is full of former trademarks that became common nouns: aspirin, escalator, cellophane, thermos. (Courts have walked through the autopsy in cases like Murphy Door Bed Co. v. Interior Sleep Systems, Inc., 874 F.2d 95 (2d Cir. 1989), where "Murphy bed" had slipped into the common vocabulary.) The danger is acute for a dominant brand that becomes synonymous with its product. The classic defensive playbook—use the mark as an adjective ("a XEROX brand copier," not "to xerox"), pair it with a generic noun, and police misuse—exists precisely to fend off genericide. Google famously survived a genericide challenge in Elliott v. Google, Inc., 860 F.3d 1151 (9th Cir. 2017), where the Ninth Circuit held that even widespread use of "google" as a verb did not render the mark generic, because the relevant question is whether the primary significance of the word to the public identifies the search engine as a source—not merely whether people use it as shorthand for searching. For the litigation contours, see Genericness Challenges to Trademark Rights treated alongside the broader Trademark Basics.

And lurking behind all three is the improper-assignment route: an assignment in gross, or an improper assignment of an ITU application, can itself cause loss of rights. The unifying thread is unmistakable. Each loss doctrine punishes a severance of the mark from the goodwill and source-identifying function that justify protecting it in the first place. Use it, control it, sell it whole—or lose it.

Key Takeaways

  • Use creates rights; registration confirms and supercharges them. Common law rights arise from actual use and are geographically limited; federal registration on the Principal Register adds nationwide constructive use, evidentiary presumptions, and a path to incontestability.
  • The application pipeline is deadline-driven. File on the right basis (1(a) use, 1(b) intent-to-use, 44, or 66(a)), survive examination and any Office Actions, clear the 30-day opposition window, and—for ITU applications—file a timely Statement of Use after the Notice of Allowance. Monitor status relentlessly; missed deadlines usually mean abandonment.
  • Two registers, very different rights. The Principal Register carries the full Lanham Act benefits; the Supplemental Register is a foothold for not-yet-distinctive marks and can never become incontestable.
  • The Madrid Protocol is the workhorse of international protection, but watch central attack and remember each designated country examines under its own law.
  • License only with real quality control. A naked license (Dawn Donut, Barcamerica, Eva's Bridal) can abandon the entire mark. Draft the controls and exercise them—but not so heavily that you create an accidental franchise.
  • Sell the mark with its goodwill. An assignment in gross is invalid and can forfeit the mark (Sugar Busters, Marshak v. Green); ITU applications carry a special anti-trafficking restriction.
  • To pledge a mark as collateral, file a UCC-1 (that perfects the interest) and also record with the USPTO to guard against good-faith purchasers.
  • Three classic ways to lose a mark: abandonment through nonuse (three-year presumption; Silverman v. CBS), abandonment through naked licensing, and genericide. Each punishes severing the mark from its goodwill.

Frequently Asked Questions

Do I have to register my trademark to have any rights? No. In the United States, trademark rights arise from use of a distinctive mark in commerce, and unregistered "common law" rights are enforceable. But those rights are limited to the geographic area of use and harder to prove. Federal registration adds nationwide rights, evidentiary presumptions, and other powerful benefits, which is why most serious brands register. See Trademark Rights Under Common Law.

What is the difference between the Principal and Supplemental Registers? The Principal Register is for inherently distinctive marks or marks with acquired distinctiveness, and it confers the full slate of benefits (presumptions of validity and ownership, nationwide constructive use, eligibility for incontestability). The Supplemental Register is for marks merely capable of distinguishing goods or services—typically descriptive marks without secondary meaning yet. It allows the ® symbol and blocks later confusingly similar filers, but carries no evidentiary presumptions and can never become incontestable.

What is an intent-to-use (ITU) application and how is it different from a use-based one? A use-based (Section 1(a)) application is for a mark already in use, supported by a specimen. An ITU (Section 1(b)) application is for a mark you have a bona fide intention to use but have not yet used; it lets you stake an early priority claim. An ITU mark cannot register until you actually use it and file a Statement of Use, which you must do (with possible extensions) within a maximum of three years after the Notice of Allowance.

What is "naked licensing" and why is it so dangerous? A naked license is a trademark license granted without the licensor exercising adequate control over the quality of the licensee's goods or services. Because a trademark guarantees consistent quality from a controlled source, uncontrolled licensing causes the mark to "lose its significance" and can result in abandonment of the entire mark, not just the one license. See Dawn Donut Co. v. Hart's Food Stores, 267 F.2d 358 (2d Cir. 1959); Barcamerica Int'l USA Trust v. Tyfield Importers, 289 F.3d 589 (9th Cir. 2002); and Eva's Bridal Ltd. v. Halanick Enters., 639 F.3d 788 (7th Cir. 2011). Always set, reserve, and actually exercise quality controls.

Can controlling my licensee too much create a problem? Yes—the "accidental franchise." If your arrangement combines a trademark license, a required fee, and significant control over (or assistance with) the licensee's method of operation, it may legally be a franchise even though your contract disclaims it. That triggers franchise-disclosure and relationship laws. Calibrate your quality control to the goods or services, and run a deliberate franchise analysis before launching a licensing program.

What is an "assignment in gross"? It is an attempt to transfer a trademark without the goodwill the mark symbolizes—and it is invalid. A trademark cannot be sold apart from the goodwill and business it represents; doing so can forfeit the mark entirely. A valid assignment transfers the mark together with its goodwill, typically by transferring the associated business or continuing the same line of goods or services. See Sugar Busters LLC v. Brennan, 177 F.3d 258 (5th Cir. 1999); Marshak v. Green, 746 F.2d 927 (2d Cir. 1984).

Can I assign an intent-to-use application before I've used the mark? Generally no. Under 15 U.S.C. § 1060(a)(1), an ITU application cannot be assigned before the Statement of Use is filed, except to a successor to the applicant's ongoing business to which the mark pertains. Improperly assigning an ITU application can invalidate the resulting registration.

How do I use my trademark as loan collateral? Perfect the lender's security interest by filing a UCC-1 financing statement under state Article 9—that is what establishes priority for a federally registered mark. Recording a short-form security agreement with the USPTO does not perfect the interest but does protect against later good-faith purchasers, so the best practice is to do both.

How long does a federal registration last, and how do I keep it alive? A registration can last indefinitely in renewable ten-year terms, but only if you keep using the mark and file the required maintenance documents: a Section 8 Declaration of Use between years 5 and 6, and a combined Section 8/Section 9 renewal between years 9 and 10 and every ten years after. Miss the deadlines and the registration is cancelled permanently. See Maintaining Trademark Registrations.

Related Articles

This article provides general information about U.S. trademark law and is not legal advice. Trademark rules are fact-specific and change over time; consult a qualified attorney about your particular situation before acting.