The Cases Examiners Cite, Part 1: The Confusion Canon

When a trademark examiner refuses your application because it's too close to someone else's, the refusal almost never rests on the examiner's gut. It rests on a short stack of decided cases — a canon — that tells the examiner how to weigh one name against another. Learn those cases and you stop being surprised by refusals. You start predicting them.

This is the first entry in a series on the cases examiners lean on, and it starts with the single most important question in trademark law: when are two marks confusingly similar? The legal shorthand is a "likelihood of confusion" refusal under Section 2(d) of the Lanham Act — the statute that governs federal trademarks. It is the most common reason the USPTO refuses registration. If you understand nothing else before you name a company, understand the cases below.

du Pont: the thirteen-factor scorecard

Start with the case that gave the whole analysis its name. In re E.I. du Pont de Nemours & Co. (C.C.P.A. 1973) laid out thirteen factors an examiner or the Trademark Trial and Appeal Board — the USPTO's in-house court, the TTAB, which hears disputes over registration — weighs when deciding whether confusion is likely. Similarity of the marks. Similarity of the goods. Channels of trade. The strength of the earlier mark. Whether any actual confusion has happened. And on down the list.

Here's the thing newcomers get wrong: the factors are not a checklist where the side with more boxes wins. In practice, examiners call two of them the key considerations — how similar the marks look, sound, and feel, and how related the goods and services are. The rest come off the bench only when the record supports them. And when examiners compare the marks themselves, they lean on a companion case, In re National Data Corp. (Fed. Cir. 1985), which says a mark is judged as a whole but its dominant feature — usually the distinctive word, not the generic one — can carry extra weight.

There's a pleasing footnote here: marks that passed through the du Pont family are still quietly demonstrating the doctrine on today's register. YIELDMASTER (filed 1988, registered 1989), for circuit-board laminating equipment, is still registered to DuPont Electronics. HOLDBAC and LITESSE, both ingredients for food manufacturing, are still registered to DuPont's former nutrition arm, which merged into IFF in 2021. Dissimilar names, each in its own defined lane, and all three still live today. YIELDMASTER went through without a single office action; that is what the two big factors look like when they never fire.

Coach v. Triumph: fame must be proven — and it isn't a magic wand

The strength of the earlier mark is one of the thirteen factors, and Coach Services, Inc. v. Triumph Learning LLC (Fed. Cir. 2012) is the case the Board cites to explain how fame is measured. Coach, the luxury handbag maker, opposed Triumph Learning's applications to register COACH for materials that prepare students for standardized tests. Different goods entirely — leather bags versus workbooks.

Here is the part the folklore gets wrong: Coach lost. The Federal Circuit — the appeals court that reviews TTAB decisions — affirmed that no confusion was likely. A test-prep COACH calls to mind a tutor, not a handbag; the goods are unrelated and move through different channels of trade. The court also rejected Coach's dilution claim, drawing a line most founders miss: fame for confusion purposes is a sliding scale, but fame for dilution — the doctrine that protects household names even without confusion — is a far higher bar, and the owner has to prove it with hard evidence of sales, advertising, and public recognition. Coach's evidence fell short of that higher bar. The register shows the receipt: Triumph's COACH marks registered in October 2012, and the core registration is still live under a successor owner.

So fame cuts both ways. Coach did prove its mark famous for confusion purposes and still lost, because the goods were unrelated. When fame meets related goods, it plays what the Federal Circuit has called a dominant role — in Palm Bay Imports v. Veuve Clicquot (Fed. Cir. 2005), the champagne house's proven fame helped sink VEUVE ROYALE. If you're clearing a name near a well-known brand, assume the brand will spend heavily to prove its strength. But Coach is the canon's reminder that a famous name is not a force field across every category.

Shell Oil and Detroit Athletic: relatedness reaches further than you think

Two cases stand for the principle that trips up first-time filers constantly: goods and services do not have to be identical — or even competitive — to be "related" for confusion purposes. In In re Shell Oil Co. (Fed. Cir. 1993), the question was whether consumers would assume two offerings come from a single source, and the court added a ratchet: the more similar the marks, the less related the goods need to be.

For how that plays out in real life, look at In re Detroit Athletic Co. (Fed. Cir. 2018). A Detroit retailer filed DETROIT ATHLETIC CO. (serial 86625093) in May 2015 for sports-apparel retail services. The examiner refused it over DETROIT ATHLETIC CLUB — a private social club with a registration covering clothing. Retail services on one side, clothing goods on the other; the Federal Circuit affirmed the refusal anyway, because consumers typically notice the first words they see — "Detroit Athletic" — and shirts are sold in exactly the kind of shops the application described. The application went abandoned months later. For a founder, the takeaway is blunt: a clearance search limited to your exact product category is not a clearance search. Confusion analysis reaches into every category a consumer would plausibly connect to yours.

Juice Generation: weak marks can share the sandbox

The most useful case for anyone building a brand out of ordinary words is Juice Generation, Inc. v. GS Enterprises LLC (Fed. Cir. 2015). Juice Generation, the New York juice-bar company, applied in 2012 to register PEACE LOVE AND JUICE for its juice bars. GS Enterprises, owner of a family of PEACE & LOVE marks for restaurants, opposed. The TTAB sided with GS. The Federal Circuit sent it back — and the reason is the whole point.

Juice Generation had put a long list of third-party PEACE-and-LOVE marks into the record, and the court held that evidence was "powerful on its face": when a phrase is used widely by many owners, each individual mark is weak, and consumers learn to tell them apart by small differences. A crowded field shrinks the zone of protection around every name in it. The ending is on the register: within months of the remand the opposition was over, and PEACE LOVE AND JUICE registered in March 2016. On October 2, 2026, that registration was canceled, because nobody filed the ten-year renewal. We mapped how the same doctrine could shape Spalding's fight with MrBeast over BEAST marks.

Juice Generation's broader portfolio shows it playing both sides of the lesson. It fought three years for its common-words name — and it also stocked its shelf with marks that reached registration without an opposition: COLD WARRIOR, THE DEFENDER, WHAT'S YOUR JUICE?, GET UR GREEN ON, ACTIVATED LEMONADE, and the house mark JUICE GENERATION, all still live. That is the strategic math of a crowded field: the more common your words, the weaker everyone's rights in them — including yours.

What to do with all this

These cases are not academic. They are the machinery that decides whether your name survives examination — and whether a competitor's near-copy survives yours. If you're new to how examination works at all, start with what an office action actually is.

Before you commit to a name, run it against the two du Pont factors examiners call the key considerations: does anything on the register look or sound like it, and does that owner sell anything a customer might connect to what you sell? Widen the net past your exact category, because Shell Oil-style relatedness is real — marks collide in more ways than founders expect. If you're circling a famous brand, expect its strength to weigh against you once proven. If your candidate name is built from everyday words, search how crowded that field already is; a Juice Generation crowd means weaker rights for everyone, including you. And one more canon entry worth knowing: under Stone Lion Capital v. Lion Capital (Fed. Cir. 2014), confusion is judged by the least sophisticated buyer your application's wording covers — not the discerning customers you actually serve.

The examiners already know these cases cold. Running your candidate name through a clearance search of the USPTO register, about 14 million applications and registrations, is how you find out what they'll see before they see it; GleanMark's knockout search runs examiner-style passes and scores the closest live marks it finds. A later entry in this series will move from the confusion canon to the descriptiveness cases — the ones that decide whether a name is a brand at all, or just a description a competitor is free to use.

Register statuses as of October 27, 2026. This analysis is based on public USPTO records and published case law and is not legal advice.