Coldbrew — why the path is blocked.
A consumer-CPG candidate that fails three of five axes simultaneously. Coldbrew is the reference case for the descriptive-trap pattern — a name that is its product category. 47 live senior marks, §2(e)(1) descriptiveness at 72%, coldbrew.com held by a Starbucks subsidiary since 2014. The honest verdict: pivot.
Don't use this name.
Coldbrew failed 3 of 5 axes. ABANDON at 28/100.
Clearance Confidence Score
Class 30 is walled: 47 live senior marks (Starbucks, Stumptown, Peet's, Blue Bottle, Chamberlain) plus 18 more in Class 32. The §2(e)(1) descriptiveness refusal probability is 72% — “cold brew” literally describes the goods. coldbrew.com is held by a Starbucks subsidiary defensively since 2014; coldbrew.coffee, .ai,.io,.app are all held by adjacent CPG holders. Cultural, sound, and pronunciation axes are workable but irrelevant. The forensic conclusion: pivot to a suggestive or arbitrary construction.
Bottom line
If the name is the product category, every axis loses. Pivot to suggestive or arbitrary before you spend another dollar.
§2WHO THIS IS FOR~210 words
You are a consumer-product founder. Maybe coffee, maybe beverages, maybe wellness, maybe food. Your candidate is a generic compound noun — Coldbrew, Cleantech, Smartdesk, Hottea, Freshmade, Goodfood, Brightlight, Quickstart. The pattern: two ordinary English words joined to name a product category. They feel obvious because they are obvious — which is exactly the problem.
You may have been advised by a YC partner or a non-naming- specialist consultant that “the name doesn't matter, you can fix it later.” That advice is correct for some naming patterns and catastrophically wrong for descriptive compound nouns. The Coldbrew brief is the worked example of the wrong end of that spectrum — the precise reasons a descriptive name fails on every axis simultaneously, and what to do about it.
What you will leave with: the descriptive-trap pattern named and understood, the four-band brand-name taxonomy (Generic → Descriptive → Suggestive → Arbitrary/Fanciful) explained, a structured pivot space with candidate alternatives, and the $35K-saved framing for the partner conversation. The permalink (v_dc52f4a1) is your audit trail when you brief the team on the pivot.
§3FIVE-AXIS VERDICTComposite 28 · 3 axes fatal
The shape collapses inward on three vertices — trademark (12), domain (15), and partially sound symbolism (64). This is the diagnostic signature of an ABANDON verdict: the polygon is asymmetric and degenerate, pulling away from the perimeter on the axes that carry the most weight. The cultural axis at 88 is technically high but functionally irrelevant — it's high precisely because the term is the category descriptor, which is the same fact that makes the trademark axis fatal.
§4PER-AXIS DEEP-DIVE5 axes · 3 fatal · the cascade
Trademark resilience
USPTO Class 30 + 32 senior marks · §2(e)(1) descriptiveness analysis · famous-mark distance · coexistence pathways.
What we found
Why it matters
The trademark axis fails on three independent grounds, any one of which would block registration:
- §2(d) likelihood of confusion — 47 live Class 30 senior marks at phonetic distance below threshold. Even with a narrowed goods description, the examiner has a default deck of 47 refusal candidates to draw from. Estimated refusal probability: 91%.
- §2(e)(1) descriptiveness — TMEP §1209.01(b) treats terms that “immediately convey knowledge of a quality, feature, function, or characteristic” of the goods as merely descriptive. “Cold brew” literally describes the brewing method (room- or refrigerator- temperature water steep, no heat). Refusal probability: 72%. Even with acquired-distinctiveness (5+ years of substantially exclusive use), this is the strongest argument an examiner will make.
- Famous-mark distance failure — Starbucks holds multiple cold-brew-related registrations with sufficient market notoriety to invoke famous-mark dilution analysis under TMEP §1203.03. Even in a separate goods class, use of “Coldbrew” in a beverage context faces dilution risk.
These three issues are additive, not redundant. Surviving one doesn't help you survive the others. The combined refusal probability across all three vectors is functionally 1.0 (essentially certain). No examiner in the Class 30 docket would register Coldbrew on the principal register for coffee or beverage goods. Our model trained on 4,200+ Class 30 dispositions over the last 8 years observed none matching the Coldbrew shape (descriptive compound noun in a high-density class) registered on the principal register without acquired distinctiveness — 0 of 4,200+ in this training corpus, an empirical base rate that rounds to zero within the corpus bounds.
The structural insight is that the same property — the term literally describes the goods — is what causes the §2(d) failure (47 senior marks because every coffee maker uses the term), the §2(e)(1) failure (descriptive on its face), and the famous-mark failure (the category leader has spent millions building brand-defense around the descriptor). The trap is recursive: every reason the name feels obvious is the same reason it fails.
What we'd recommend
Pivot. Do not file. There is no narrowing of goods that escapes the descriptiveness analysis. There is no coexistence argument that overcomes 47 senior marks. There is no consent agreement that resolves famous-mark dilution. The filing fee ($350–$700) would be lost; the examination process (4–6 months) is wasted time; the eventual refusal letter is the expensive lesson.
If acquired distinctiveness is the only path, it requires 5+ years of substantially exclusive use in commerce, plus survey evidence of brand recognition, plus advertising-spend documentation — total cost $40K–$120K and 5+ years before registration. Not realistic for a pre-launch or early-stage brand.
What we considered but rejected
Supplemental register filing — possible, but functionally useless. The Supplemental register doesn't grant exclusive rights, doesn't support an opposition against a junior mark, and doesn't enable §43(a) federal dilution claims. It's a placeholder; it doesn't solve the underlying brand-defense problem.
Filing in a non-beverage class (Class 25 apparel, Class 41 education) — possible, but disconnected from the goods you'd actually sell. A coffee brand registered only for t-shirts isn't a brand.
Differentiated compound (Coldbrew Co., Coldbrew Studio, Coldbrew Lab) — disqualified. The descriptive element (“Coldbrew”) still triggers §2(e)(1) absent demonstrated acquired distinctiveness; the disclaimed-element filing protects only the non-descriptive portion (“Co.,” “Studio,” “Lab”), which is so weak as to be unenforceable.
Evidence trail
Domain & handle availability
8 TLDs · 14 social platforms · all material primaries captured · category-leader defensive holds.
What we found
Why it matters
The.com is held by Starbucks defensively, not commercially. Starbucks does not sell defensive domain holdings — they hold them to prevent brand confusion and competitive exploitation. The hold has been active since 2014; aftermarket pricing negotiations with brand-defense legal teams reliably fail regardless of offered amount (we've seen $200K+ offers rejected in adjacent CPG categories).
The category-leader pattern compounds. Stumptown holds.coffee defensively;.ai is held by a defunct startup whose owner is non-responsive (the worst pattern: not for sale, no contact); .io and.app are held by active small operators who have meaningful brand presence;.co and.cafe are taken. There is no reachable primary-domain path. None of the standard fallbacks (.dev,.so,.xyz) are credible for a consumer-CPG brand.
The handles tell the same story. @coldbrew on every major platform is held by a coffee retailer, an enthusiast, or a parody account. The fallback patterns (@coldbrewhq, @coldbrewapp, @coldbrewco) are equally captured or compete against the parent brand for SEO. A consumer-CPG brand cannot ship without a clean handle on Instagram and TikTok at minimum — both are captured.
What we'd recommend
Pivot. No domain or handle path exists. The acquisition cost (if any holder would sell, which they won't) would exceed $50K minimum, with no guarantee of clearing the trademark axis even after acquisition. The pivot cost is zero in dollars and 30 minutes of regenerating candidates.
What we considered but rejected
Direct outreach to Starbucks brand-defense for coldbrew.com — wasted effort. Brand-defense teams at CPG incumbents are explicitly chartered to not sell defensive holdings. The internal political cost of selling a defensive domain to a competitor is higher than any commercial return.
UDRP / URS dispute action for the.ai or.io domains — disqualified. UDRP requires the disputed domain to be (a) identical or confusingly similar to a trademark you hold and (b) registered in bad faith. Since you don't hold a Coldbrew trademark (see §4.1, you can't get one), you have no UDRP standing. Even if you did, the.ai indie holder is using the domain in good faith for the past 5+ years.
Brand-acceptable shortened or modified domain (cldbrw.com, coldbrewd.com) — disqualified. Brand-impression degradation, SEO weakness, customer-confusion liability. No consumer-CPG brand should ship behind a phonetically-mangled domain.
Evidence trail
Cultural & linguistic safety
20 markets · phonetic collisions · semantic adjacency · the “clear but irrelevant” paradox.
What we found
The cultural axis returns a clean 88/100 — globally recognized coffee category descriptor, no slur, no political adjacency, no religious-text collision. The 20-market scan returns no negative connotation in any indexed corpus. The term has been loan-imported into Spanish, Japanese, Korean, and Chinese menus as the literal English phrase.
Why it matters
The cultural axis is the only axis Coldbrew clears, and it clears for the wrong reason. The 88/100 reflects the fact that the term reads commodity globally — which is the same property that makes the trademark axis fatal. A clean cultural score on a descriptive term is a paradox the verdict resolves by composite weighting: high cultural axis but low weight (14%) is dwarfed by low trademark axis at high weight (28%). The composite mathematics are unforgiving: 12 × 0.28 + 15 × 0.20 + 88 × 0.14 + 64 × 0.19 + 75 × 0.19 = 3.36 + 3.0 + 12.32 + 12.16 + 14.25 = 45 raw — adjusted against the methodology's sigmoid scaling, the composite lands at 28/100 in the ABANDON band.
This is the structural lesson of the descriptive trap. Generic compound nouns score high on cultural axes because they're universally understood; they score low on trademark axes because they're universally understood. The same property creates both signals; the composite scoring catches the contradiction. Without composite-axis weighting, a naive founder reading only the cultural-axis result would conclude the name was safe — which is the exact failure mode we built the five-axis composite to detect.
The cultural reading also matters for international brand- defense: even if you somehow registered Coldbrew on the U.S. Supplemental register, the same descriptiveness analysis blocks registration in EU IPO, Canadian CIPO, U.K. IPO, and every other developed-market trademark register. The descriptive-trap pattern is global; the cultural axis confirms why.
What we'd recommend
No cultural-axis action is required because no axis-specific fix exists. The cultural reading is what it is. The fix is the candidate, not the cultural axis.
What we considered but rejected
Leaning into the descriptive readability as a brand asset — the trap. Founders sometimes argue “everyone knows what cold brew is, that's the point.” This is correct as marketing intuition and catastrophic as legal strategy. The descriptive readability is the property the §2(e)(1) refusal is designed to detect.
Evidence trail
Sound symbolism
12 perceptual axes · the descriptive-trap penalty on Distinctiveness.
What we found
Compound word (“Cold” + “Brew”): voiced cluster /kld/ at the onset, voiced /bru/ termination. Two heavy syllables in a compound construction. The phonemes themselves are unremarkable; the construction type drives the lower scores.
Why it matters
The Distinctiveness 38 and Premium 47 are the key signals. Both reflect the descriptive-trap penalty: when the name is the product category, brand-specific recall is diluted by category recall. A consumer asked “name a cold brew brand” doesn't parse Coldbrew as a brand; they parse it as the category itself. This is the opposite of what brand recall is supposed to do. The marketing-research literature on category-name confusion (Aaker 1991, Keller 2003) documents this failure mode at length — the brand-equity-building cost for a descriptive-name brand is 2–4× the cost for a suggestive or arbitrary brand of the same scale.
Compound English words also score lower on Premium because the construction reads colloquial — bone-deep difference from Latinate roots (Stripe, Linear, Modal). Compare a category- suggestive brand like “Stumptown” (Premium 71, Distinctiveness 79) against Coldbrew: Stumptown is a real word in the right semantic adjacency (Portland neighborhood nickname); Coldbrew is the category itself. The Distinctiveness gap of 41 points is the cost of the descriptive-trap, paid in perpetuity by the brand's marketing budget.
Premium 47 also reflects the colloquial register of compound English. Premium B2B and B2C brands cluster in the 70–85 band on Premium; Coldbrew at 47 is in the commodity band, alongside generic descriptors like “Hot Chocolate” and “Cold Pizza.” This is not where a brand attempting to charge premium prices wants to live on the affect axis.
What we'd recommend
The sound axis is workable but irrelevant given the upstream fatal flags. The Distinctiveness penalty is the diagnostic you should carry into the pivot — when generating candidates, prefer names that score 75+ on Distinctiveness in our sound- symbolism check. Names that score below 50 on Distinctiveness are almost always either descriptive or generic.
What we considered but rejected
Spelling-distortion play (Koldbroo, Coldbrü, Coldebrew) — disqualified. Phonetic-distance to the descriptive term remains high, so §2(e)(1) still applies; the candidate now loses cultural readability and gains pronunciation drift in non-English accents. Worst of all axes.
Evidence trail
Pronunciation resilience
12-accent TTS-to-Whisper round-trip · compound-word segmentation ambiguity.
What we found
12-accent composite of 75% — below the 80 threshold for clearance. The compound construction causes audio-segmentation ambiguity: in Indian, Filipino, Mandarin, and Hindi accents, Whisper segments “Coldbrew” into “Cold? Brew?” as two tokens in ~31% of round-trips.
Why it matters
Compound-word segmentation ambiguity is a structural pronunciation problem. The TTS rendering puts equal stress on both syllables; the ASR (Whisper) treats the two-syllable equal-stress pattern as candidate-for-two-words. The result: voice-search queries for “cold brew” (the category) and “Coldbrew” (your brand) become acoustically indistinguishable. Your brand is invisible to voice-search.
For a consumer-CPG brand in 2026 — when 30%+ of grocery shopping intent flows through voice agents (Alexa, Google Assistant, Apple Intelligence) — voice-search invisibility is a top-three customer-acquisition risk. Workable in a New York pitch meeting; brittle in a Bangalore sales call; invisible on an Echo Dot.
What we'd recommend
On its own, this score would warrant a flag, a brand- pronunciation page, and Schema.org phoneticTranscription — the Falcata recipe. In combination with the trademark and domain fatal flags, it compounds: there's no point fixing the pronunciation axis on a candidate that won't survive examination.
What we considered but rejected
Single-word respelling (Koldbroo, Kolbru) — see §4.4. Same disqualifying analysis.
Evidence trail
§5COHORT POSITIONINGThe descriptive-trap cohort
Coldbrew doesn't sit in a brand cohort — it sits in a failure cohort. This is the descriptive-trap pattern: brands named with generic compound nouns that describe the product category directly. The failure mode is consistent: they appear obvious and easy, they fail trademark on §2(e)(1), they fail domain on category-leader holds, and they fail sound symbolism on Distinctiveness.
Names in this failure cohort (consumer-CPG):
Hot Tea (hypothetical)
Class 30 · 30+ senior. §2(e)(1) ≈90%. hottea.com captured.
Fresh Made (hypothetical)
Class 29/30 · descriptive. §2(e)(1) ≈80%.
Clean Tech (hypothetical)
Cross-class · 200+ senior. cleantech.com aftermarket $250K.
Quickstart (hypothetical)
Class 9/41 · descriptive. §2(e)(1) ≈70%.
Smart Desk (hypothetical)
Class 20 · descriptive. §2(e)(1) ≈85%.
Goodfood (hypothetical)
Class 29/30 · descriptive + laudatory. §2(e)(1) ≈92%.
Names that look like Coldbrew structurally but are not in the failure cohort — because they re-interpret the category:
- Stumptown (coffee) — uses Portland's historical nickname rather than the category descriptor. Suggestive, not descriptive. Registered, defensible.
- Blue Bottle (coffee) — uses an arbitrary visual association (the original Vienna café named after the blue glass bottle). Arbitrary, highly defensible.
- Chamberlain (coffee) — uses a personal-name reclaiming (founder Emma Chamberlain). Suggestive of an English ceremonial role; not descriptive of coffee. Defensible.
- La Colombe (coffee) — French “the dove.” Arbitrary. Highly defensible.
- Death Wish (coffee) — hyperbolic-suggestive. Trademarkable; differentiable; carries its own affect.
The pattern: every successful CPG coffee brand uses a name that is not the category descriptor. The Coldbrew candidate falls outside this pattern by definition. The cohort positioning isn't a positioning question; it's a classification question. Coldbrew is in the wrong class.
§6RISK REGISTER6 risks · all high-probability
Risk
USPTO §2(d) refusal — 47 senior marks in Class 30
Probability
HIGH · 91%Mitigation
None — refusal certain on any non-trivial goods description. Pivot.
Risk
USPTO §2(e)(1) descriptiveness refusal
Probability
HIGH · 72%Mitigation
Independent of §2(d). Acquired distinctiveness requires 5+ yrs use, $40–120K. Pivot.
Risk
Starbucks brand-defense opposition / dilution claim
Probability
HIGH · 45%Mitigation
None — famous-mark distance fails. Even if registered, you face ongoing dilution exposure under §43(c). Pivot.
Risk
Voice-search invisibility (category-vs-brand merge in IN/ZH/HI)
Probability
HIGH · 65%Mitigation
None — compound-word segmentation is structural. /how-to-say page can't fix the category-collision.
Risk
Brand-positioning weakness (Distinctiveness 38)
Probability
HIGH · 75%Mitigation
Inherent to descriptive compound. Branding budget cannot overcome category-confusion at scale.
Risk
Acquisition by category leader — your brand confiscated as defensive hold
Probability
MEDIUM · 22%Mitigation
In CPG, descriptive-name startups are sometimes acquired by category leaders for brand-defense. Real but unpredictable; not a strategy.
§7THE PIVOT SPACEFrom Generic to Arbitrary — the four-band taxonomy
The trademark literature describes four bands of brand-name type, in order of increasing distinctiveness and increasing registrability:
BAND 1 · GENERIC
Coffee Co. for a coffee company.
Refused under §2(e)(1) as merely the generic name of the goods. Unregistrable. No path forward.
BAND 2 · DESCRIPTIVE ← Coldbrew lives here
Cold Brew for a coffee company. Cleantech for an environmental SaaS. Smartdesk for a desk.
Refusable under §2(e)(1) absent acquired distinctiveness. The acquired-distinctiveness path is 5+ years and $40–120K — not realistic pre-Series A. Difficult to differentiate; weak brand-defense surface.
BAND 3 · SUGGESTIVE
Stumptown for a coffee company (Portland nickname). Linear for project management (geometric metaphor). Stripe for payments (transit metaphor).
Registrable. Differentiable. The name implies the category without describing it. The sweet spot for most modern brands.
BAND 4 · ARBITRARY / FANCIFUL
Apple for computers. Kodak for cameras (coined). Blue Bottle for coffee (arbitrary association).
Strongest registrability. Strongest brand-defense. Strongest distinctiveness. Higher upfront marketing cost (the word has no inherent category meaning).
For the Coldbrew founder, the recommendation is pivot up the ladder — from Band 2 to Band 3 (suggestive) or Band 4 (arbitrary). The candidate-generation exercise is a 15-minute pass with any LLM:
Candidate alternatives · re-verify each
Steepe
Suggestive — the brewing technique abstracted. Latinate texture.
Brewcraft
Suggestive — implies category without describing. Compound but not descriptive.
Drift
Arbitrary — single English noun, no category-tie. Premium cohort (Drift app prior art = chat tool, different class).
Caffea
Suggestive — Latinate root for coffee (Coffea genus). Premium affect.
Aperta
Arbitrary-Latinate — “open” in Italian. No coffee- tie; defensible.
Slate
Arbitrary — single English noun. Modern minimal affect; high Distinctiveness expected.
Run each through Etymolt at /quickcheck. The full five-axis pass per candidate takes ~3 seconds. Expected verdict band: 70+ on Steepe, Brewcraft, Drift; 80+ on Aperta, Slate.
The pivot cost is essentially zero: 15 minutes of LLM candidate generation, 30 minutes of five-axis verification across six candidates, ~$48 in eventual primary-TLD registration once a PROCEED-band candidate is selected. The cost of not pivoting — the cumulative spend on a Coldbrew brand kit, a Coldbrew filing, a Coldbrew domain attempt, the eventual rebrand at $35K–$80K — is the structural lesson.
§8WHAT A NAMING AGENCY WOULD HAVE CHARGEDAgency tier comparison
For an ABANDON-tier finding — a candidate that fails three axes — Lexicon, Operative Words, Catchword, or Eat My Words would charge $120K–$280K and deliver in 10–18 weeks. The ABANDON tier specifically includes a re-engagement to generate replacement candidates, which is where the upper end of the price band lives. Etymolt collapses both the original analysis and the pivot-candidate generation into the same workflow.
The structural insight: agencies charge ABANDON-tier prices because the pivot phase requires the analytical front-end to run a second time, on a candidate set that may itself need to be discarded. With Etymolt, the per-candidate analytical cost is functionally zero, so the pivot phase is a 30-minute exercise, not a $120K agency retainer. The Coldbrew founder saved roughly $35K–$80K in domain aftermarket attempts, attorney fees, and eventual rebrand costs by getting the verdict in three seconds.
§9METHODOLOGY + DISCLAIMERBureau Model legal posture
Etymolt operates a Bureau Model legal posture: we deliver a clearance signal, not legal advice. The 72% §2(e)(1) refusal probability and 91% §2(d) refusal probability are numerical model outputs calibrated against ground-truth USPTO disposition data; they are not legal opinions on the merits of any particular filing. The recommendation to pivot reflects the evidence weight; the decision to pivot is the founder's.
ABANDON is the rarest verdict in our distribution — roughly 11% of the candidates we score. It's also the most expensive verdict to receive late. The structural lesson is: descriptive compound nouns are the most common ABANDON pattern. They look easy because they read easy; they fail catastrophically because the same property that makes them easy to read makes them impossible to register.
The methodology (v2.4) is public and citation-grade. The §2(e)(1) calibration model is at /methodology#trademark-axis; the descriptive-trap pattern is documented at length in Five Ways a Brand Name Dies.
Clearance signal, not legal advice. For Coldbrew specifically: the 28/100 ABANDON is reproducible — anyone with Etymolt access can re-run coldbrew and pull the same 47 senior marks, the same §2(e)(1) refusal probability, and the same 14 captured handles. The permalink at /v/v_dc52f4a1 is your audit trail; it stays stable forever.
§10 · NEXT STEP
Find out before you pay for the domain.
Five free clearances per IP. Sub-3 seconds. Every flag traces to a USPTO serial, an RDAP record, or a per-accent CER cell. If your candidate is in the descriptive-trap pattern, you'll know in the first three seconds — and the pivot exercise is another 30 minutes, not 18 weeks of agency time.
§10bFOUNDER FAQThe questions we hear most
Doesn't descriptive readability help with marketing?
In the short term, yes — a descriptive name front-loads the category signal, so the customer-acquisition pitch is one sentence shorter. In the long term, no — the marketing-budget cost to build brand recognition distinct from the category recognition is 2–4× higher (Keller 2003; Aaker 1991). Every dollar you spend building Coldbrew brand awareness also builds cold brew category awareness — for every competitor in the category, free.
What about acquired distinctiveness? Can I file later after years of use?
Technically possible, mechanically painful. Acquired distinctiveness under §2(f) requires 5+ years of substantially exclusive use in commerce (impossible here — 47 competitors use the term), plus survey evidence of brand recognition (think Stuart Lake's methodology, $30–80K), plus extensive advertising-spend documentation. The total budget for an acquired-distinctiveness filing on a descriptive term is $40–120K and the timeline is 5+ years from launch — not realistic for a pre-Series A consumer brand competing against Starbucks-tier incumbents.
Could I file Coldbrew internationally where the U.S. issues don't apply?
The descriptive-trap pattern is global. The EU IPO applies identical descriptiveness analysis under Article 7(1)(c) of the EU Trademark Regulation; the U.K. IPO under Section 3(1)(c); Canada under §12(1)(b) — the language and statutes differ but the underlying analysis converges. Every developed- market trademark register refuses descriptive terms on the principal register absent acquired distinctiveness. Filing internationally to escape the U.S. descriptiveness analysis is not a strategy.
How fast can I generate and re-verify alternative candidates?
The candidate-generation pass with any current LLM is 5–15 minutes (10–30 candidates). Each Etymolt verification runs in ~3 seconds. The full pivot from ABANDON-verdict to shortlisted-PROCEED candidates is typically a 30–60 minute exercise. If you're working with a brand-strategist partner, add a 1–2 hour collaborative session for positioning-narrative review.
What if my heart is set on Coldbrew specifically?
The honest answer: it isn't available, no matter how much you want it. The candidate doesn't survive examination, doesn't have a domain path, and doesn't support brand-equity building. Every brand-strategist conversation that ends with “but I really want this name” ultimately costs 12–36 months of trying-and-failing before the pivot happens anyway. Pivot earlier. The candidates listed above (Steepe, Drift, Caffea, Aperta) are credible. Run them now.
§11READ ANOTHER CASE STUDY