866 F.3d 1315 (Fed. Cir. 2017)
How will.i.am's company lost a §2(d) appeal over the mark I AM — why identical word marks that are neither suggestive nor descriptive make the first DuPont factor weigh heavily, and why a personal-name 'association' restriction couldn't save the application. The holding, the test, how to cite it, and the live USPTO records.
The 30-second version
In 2010, William Adams — the Black Eyed Peas frontman known as will.i.am — filed to register a two-word mark across cosmetics, sunglasses, and jewelry: I AM (i.am.symbolic, llc, the entity that carried the cases through appeal, is his successor-in-interest). The words were not suggestive or descriptive of those goods (arbitrary, in trademark terms) and were spelled in plain standard characters. He already held an I AM registration for clothing in Class 25 — but that registration did not extend to cosmetics, eyewear, or jewelry. The problem was that three other owners had already registered I AM first in those classes: a perfume seller in Class 3, a German eyewear-and-jewelry company in Classes 9 and 14, and an individual with a Class 14 bracelet on the Supplemental Register.
The applicant's answer was clever. It amended every application to add the phrase 'all associated with William Adams, professionally known as will.i.am.' The theory: I AM was not really being registered as a bare mark, but as a piece of the will.i.am identity, tied to a famous person the prior registrants had nothing to do with. The Board refused all three anyway, and the company appealed to the Federal Circuit.
The court affirmed. Its reasoning is why first-years still read the case: you compare the applied-for words to the cited marks — I AM against I AM — not the applicant's separate house mark or the celebrity behind it. Identical word marks that are neither suggestive nor descriptive of the goods make the first DuPont factor weigh heavily. And a restriction that names a person, without limiting the actual goods, channels, or buyers, changes none of that. The affirmance also rested partly on harmless error: the marks' identity was strong enough that any gap in analyzing the cited marks' claimed weakness could not have changed the outcome.
William Adams (with i.am.symbolic, llc as successor-in-interest) files intent-to-use applications for standard-character I AM: Serial 85044494 (Class 3 cosmetics), 85044495 (Class 9 sunglasses), and 85044496 (Class 14 jewelry/watches). Adams already held a prior I AM registration for clothing in Class 25, which did not reach these classes.
Each application is amended to recite that the goods are 'all associated with William Adams, professionally known as will.i.am' — the applicant's attempt to distinguish the cited I AM registrations.
The examining attorney refuses under Section 2(d), citing the Siegel/Danica perfume registration (Class 3), the Beeline eyewear-and-jewelry registration (Classes 9 and 14), and the Finch bracelet registration (Class 14).
In Symbolic I, II, and III, the Board affirms each refusal: the marks are at least legally identical, the goods and channels overlap, and the association restriction imposes no meaningful limitation. (Exact Board mailing dates not independently verified; the Federal Circuit decided in 2017.)
A panel of Prost, C.J., Lourie and Schall, JJ. (opinion by Lourie) affirms, finding no reversible error in the likelihood-of-confusion conclusion and treating any gap in weighing the cited marks' alleged weakness as harmless.
All three I AM applications go abandoned after the ex parte appeal (USPTO status: 'Abandoned — After Ex Parte Appeal'); the precedent on the first DuPont factor remains.
Put yourself in the applicant's chair, because for once the trap is the applicant's, not the examiner's. The instinct is that a famous person should be able to register a phrase tied to their public identity — will.i.am is built from 'I am,' and consumers who see I AM on his branded cosmetics surely think of him, not a perfume company they have never heard of. So you bolt a restriction onto the goods: 'associated with William Adams, professionally known as will.i.am,' and argue the marks now point to different sources. The flaw is that Section 2(d) compares the marks themselves, on the four corners of the application, against the cited registrations — and the words applied for are still just I AM. As the court put it, the will.i.am restriction does not impose a meaningful limitation for the confusion analysis: it does not restrict the trade channels or classes of purchasers, does not alter the nature of the goods, and does not represent that the goods would be marketed in any particular, limited way; the record showed Adams is known as will.i.am, not as I AM. Identical word marks that are neither suggestive nor descriptive of the goods, on overlapping goods, are exactly the fact pattern that makes the first factor weigh heavily, and a precatory name-association clause does not move it.
Prong 1 — the marks
Similarity of the marks — compared in their entireties as to appearance, sound, connotation, and commercial impression. The test is overall commercial impression, not a side-by-side comparison; legally identical word marks weigh heavily toward confusion.
Prong 2 — the goods
Relatedness of the goods or services and the trade channels — judged on the goods as identified in the application and the registration, not on real-world use the applicant points to outside the four corners.
The 'two prongs' are a teaching shorthand: §2(d) is formally the 13-factor DuPont balancing test, and any single factor can control. This case is a first-factor (marks) case. Standard-character marks claim no particular font or style, so I AM in standard characters is legally identical to the cited I AM marks however they are displayed (note the older cited marks issued as pre-2003 'typed drawings,' which the USPTO treats as the equivalent of standard characters, so the legal-identity point holds). The key move is what the comparison excludes: you compare the applied-for words to the registrant's mark, not the applicant's separate will.i.am house mark and not the celebrity's fame — and a restriction tying goods to a named person, without limiting the goods, channels, or purchasers, adds no meaningful distinction. Procedurally, the case is also read as a reminder that the Board must consider every DuPont factor for which there is record evidence (here the applicant pressed the number-and-nature-of-similar-marks angle); the Federal Circuit affirmed only because, with legally identical marks that are neither suggestive nor descriptive, any failure to address the cited marks' claimed weakness was harmless error — which is why practitioners (e.g., Finnegan's INCONTESTABLE blog) cite it for the 'consider all factors' point as much as for the marks prong. The 'weighs heavily' rule the opinion applies is itself quoted from In re Majestic Distilling (word marks identical but neither suggestive nor descriptive of the goods); the commercial-impression test traces to Coach Servs. v. Triumph Learning, resting on the original DuPont opinion.
These are the two factors that usually decide §2(d), but the governing standard is the full 13-factor DuPont likelihood-of-confusion test — any factor can control on the right record.
Legally identical word marks
The applied-for mark and all three cited marks were the words I AM in standard-character (or, for the older marks, pre-2003 typed-drawing) form. That form claims no particular display, so the Board treated them as at least legally identical — and I AM is neither suggestive nor descriptive of cosmetics, sunglasses, or jewelry (arbitrary, in trademark terms), which is the category that makes the first factor weigh heavily.
The restriction limited nothing real
The 'associated with William Adams, professionally known as will.i.am' clause did not restrict the goods, the trade channels, or the classes of purchasers. The court held it 'does not impose a meaningful limitation' for the likelihood-of-confusion analysis — it does not alter the nature of the goods or show they would be marketed in any particular, limited way; the record showed Adams is publicly known as will.i.am, not as I AM.
Overlapping goods, channels, and ordinary buyers
The applicant's and registrants' goods were identical or closely related and moved through the same trade channels to the same ordinary consumers, with no limitation in the identifications — so the second, third, and fourth factors also favored confusion.
Coexistence evidence fell short
The applicant pointed to third-party I AM registrations to argue a weakened, crowded field. The court distinguished Juice Generation (which involved 26 third-party PEACE & LOVE marks) and Jack Wolfskin (16 paw-print designs): a meaningful crowded-field showing for identical word marks needs evidence the third-party use is 'ubiquitous' or 'considerable,' which the handful of I AM registrations here did not approach. And because the marks were legally identical, any error in not fully weighing the cited marks' alleged weakness was harmless.
“when word marks are identical but neither suggestive nor descriptive of the goods associated with them, the first DuPont factor weighs heavily against the applicant.”
The rule to memorize
You compare the marks as applied for, in their entireties, on overall commercial impression — not against the applicant's separate house mark or a celebrity's fame. When the word marks are identical and are neither suggestive nor descriptive of the goods, the first DuPont factor weighs heavily toward confusion, and you cannot defeat that by tacking a personal-name or 'associated with' restriction onto the goods unless the restriction actually limits the goods, channels of trade, or classes of purchasers — the i.am.symbolic restriction did none of these and 'does not impose a meaningful limitation.' Shorthand: a precatory name-association clause is not a marks-prong argument.
The proper comparison is between the marks as identified in the application and the cited registration, in their entireties, not between the cited mark and any separate house mark or persona Applicant associates with the goods. See In re i.am.symbolic, llc, 866 F.3d 1315, 1319 (Fed. Cir. 2017) (quoting In re Majestic Distilling Co., 315 F.3d 1311, 1315 (Fed. Cir. 2003)). Where, as there, the word marks are legally identical and are neither suggestive nor descriptive of the goods, the first DuPont factor weighs heavily, and a restriction reciting that the goods are 'associated with' a particular person does not impose a meaningful limitation unless it restricts the goods, the channels of trade, or the classes of purchasers. Here, by contrast, [Applicant's mark and the cited mark differ in [appearance/sound/connotation] / Applicant's amended identification genuinely restricts the goods and channels so that the marketplaces do not overlap], which materially distinguishes this record from i.am.symbolic.
This is usually the EXAMINER's cite, not yours — so use it defensively. If your marks really are identical, do not pretend a name-association tagline solves it; either show a real, enforceable restriction on the goods/channels that the i.am.symbolic restriction lacked, or move your fight to the marks differing in connotation or to a genuine crowded field with 'considerable' third-party use. Quote the 'weighs heavily' line back accurately — including its internal '(quoting Majestic Distilling)' attribution — and then explain why your facts are outside it.
Of 603 applications that cited this case against a §2(d) likelihood of confusion refusal, 190 have reached a final outcome — 32.1% registered, below the 37.6% §2(d) baseline (N=10,097).
When a usage rate is shown for this case, read it as selection bias, not as the case's persuasive value. Applicants reach for i.am.symbolic-style framing — house-mark or personal-name distinctions — precisely when the words are already identical and there is no clean marks-prong difference left to argue. Those are among the hardest §2(d) refusals to overcome, so any below-baseline registration rate in a given corpus reflects the difficulty of the fights the case shows up in, not the citation 'losing' the case. The citation marks the presence of an identical-word-mark problem; the outcome rides on whether a real restriction or a genuine crowded field exists.
These rates are correlational, not causal. Citing a case does not “win” it — a response that cites a precedent often faced a harder, more contested refusal to begin with. Every rate is shown against its refusal-ground baseline and the number of terminal (finally-decided) applications it is based on; pending applications are excluded. This is a descriptive picture of how the profession argues, not legal advice.
I.AM.SYMBOLIC, LLC · Serial 85044494
Applicant's Class 3 application — cosmetics and personal-care goods 'associated with William Adams, professionally known as will.i.am.' Standard characters, intent-to-use. Refused under §2(d) over the Siegel/Danica perfume registration; abandoned after ex parte appeal.
I.AM.SYMBOLIC, LLC · Serial 85044495
Applicant's Class 9 application — the refusal on appeal concerned sunglasses and sunglass cases. Standard characters, intent-to-use. Refused under §2(d) over the Beeline registration; abandoned after ex parte appeal.
I.AM.SYMBOLIC, LLC · Serial 85044496
Applicant's Class 14 application — jewelry, watches, and bracelets. Standard characters, intent-to-use. Refused under §2(d) over the Beeline and Finch registrations; abandoned after ex parte appeal.
DANICA AROMATICS, INC. · Serial 75023790
Cited registration (Reg. 2045626) for perfume, Class 3 — the basis for the Class 3 refusal. The Federal Circuit opinion refers to this as the 'Siegel registration' (after registrant Danica Siegel); the registration was later assigned to Danica Aromatics, Inc.
BEELINE GMBH · Serial 76508415
Cited registration (Reg. 3188447) for sunglasses (Class 9) and jewelry, fashion jewelry, bracelets, and watches (Class 14) — the basis for the Class 9 and Class 14 refusals.
Justin Finch · Serial 77758654
Cited registration (Reg. 3935952) for a silicone stretchable wrist band in the nature of a bracelet, Class 14, on the Supplemental Register — co-cited with Beeline for Class 14. It was a valid citation in 2017 and was later cancelled under Section 8.
Live USPTO records in GleanMark — click through for full prosecution history.
The complete TTAB record behind this decision — every filing and exhibit, the full timeline, and a one-click AI summary.
Open proceeding 85044494A human footnote: the applicant here was the company behind will.i.am, who has filed a long string of I AM-family marks over the years — including an I AM registration for clothing (Class 25) that predates this dispute and survives today; this was the branch that did not take root. The three I AM applications went abandoned after the appeal, and the cited registrations stayed on the register — except the Finch bracelet registration (Reg. 3935952) that helped bar Class 14, which was later cancelled under Section 8. The artist kept building; this particular two-word claim simply belonged, on the register, to the people who got there first.
What did In re i.am.symbolic decide?
The Federal Circuit affirmed Section 2(d) refusals of will.i.am's standard-character mark I AM over three prior I AM registrations. Because the word marks were legally identical and neither suggestive nor descriptive of the goods, the first DuPont factor weighed heavily toward confusion, and a restriction reciting that the goods were 'associated with William Adams, professionally known as will.i.am' added no meaningful limitation. The affirmance also rested partly on harmless error — the marks' identity was strong enough that any gap in weighing the cited marks' alleged weakness could not have changed the result.
Can you add a 'associated with [name]' restriction to overcome a §2(d) refusal?
Not by itself. In re i.am.symbolic holds that a restriction tying goods to a particular person 'does not impose a meaningful limitation' and does not distinguish identical marks unless it actually limits the goods, trade channels, or classes of purchasers. A precatory name-association clause that leaves the goods and channels unchanged does not move the marks comparison.
Why are marks compared 'in their entireties' under the first DuPont factor?
The first DuPont factor asks whether the marks are similar in appearance, sound, connotation, and overall commercial impression, judged as wholes rather than in a side-by-side dissection. As i.am.symbolic explains, you compare the applied-for mark to the cited registration — not the applicant's separate house mark or a celebrity's fame — and identical word marks that are neither suggestive nor descriptive of the goods weigh heavily toward a likelihood of confusion.
Does i.am.symbolic say anything about how the Board must weigh the DuPont factors?
Yes — it is widely cited for the procedural point that the Board must consider every DuPont factor for which there is record evidence. Here the applicant pressed the number-and-nature-of-similar-marks factor, but the Federal Circuit affirmed because, with legally identical marks that are neither suggestive nor descriptive of the goods, any failure to fully address the cited marks' claimed weakness was harmless error.
In re E.I. du Pont de Nemours & Co.
The §2(d) factors · 26.6% registered
Juice Generation, Inc. v. GS Enterprises LLC
Crowded field of weak marks · 22.6% registered
Palm Bay Imports, Inc. v. Veuve Clicquot Ponsardin
Foreign equivalents & fame · 41.7% registered
Jack Wolfskin Ausrüstung für Draußen GmbH v. New Millennium Sports, S.L.U.
Crowded field; weak design elements · 26.7% registered
Coach Services, Inc. v. Triumph Learning LLC
Fame as a §2(d) factor · 33% registered
In re Gyulay
Mere descriptiveness · 53.6% registered
In re Abcor Development Corp.
Descriptiveness standard · 61.7% registered
In re Viterra Inc.
Similarity; standard-character display · 34.2% registered
In re Thor Tech, Inc.
Goods-relatedness coexistence · 21.4% registered
In re Shell Oil Co.
Relatedness of goods and services · 23.1% registered
In re National Data Corp.
Dominant feature; disclaimed matter · 41.1% registered
In re Detroit Athletic Co.
First-word dominance; similarity · 40.6% registered
Stone Lion Capital Partners, L.P. v. Lion Capital LLP
Least-sophisticated purchaser · 42.9% registered
In re Bayer Aktiengesellschaft
Descriptiveness; internet evidence · 54.2% registered
GleanMark turns the same prosecution data behind this analysis into clearance, coexistence evidence, and a structured Office Action response draft — in minutes.
Start a Free Trial