Brand naming — the founder's playbook.

What brand naming actually is

Brand naming is not branding. Branding is the entire system — the logo, the palette, the voice, the way the product talks to the customer. Brand naming is the single, narrow, load-bearing decision underneath all of it: the choice of word. One word. Sometimes two. Almost never three. The word that goes on the homepage, on the receipt, on the trademark filing, on every podcast and every search result for the rest of the company's life.

A naming decision is unusual among founder decisions in three ways. It is made once and rarely revisited. It is highly asymmetric — a wrong name costs forty thousand dollars and a quarter of momentum to undo; a right name costs nothing extra to keep. And it is governed by five constraints that almost never appear in the same room: trademark registries, domain markets, foreign-language semantics, phonetic perception, and acoustic resilience. A name that passes one of those five gates and fails the other four is the typical case, not the exception.

In our verifier's corpus of about 1.4 million observed brand-name candidates, the median name fails on two of the five axes. The mean founder checks one axis before launch. The arithmetic is hostile, and the cost of that arithmetic is what keeps showing up in our inbox as USPTO Office Actions, handle squatters, and the quiet rebrand notices that appear in the third paragraph of a Series A blog post.

The good news: the work of choosing a name well is bounded. A small set of frameworks and a small set of receipts close the gap. The rest of this page is a map to that work — and to the seven articles that unpack the moves in detail.

The five questions every naming process answers

Across every credible naming methodology — Lexicon's in San Francisco, Operative Words' in Brooklyn, Eat My Words' in Sausalito, Catchword's in Oakland — the same five questions sit at the top of the workflow. They are the questions the worksheet asks before a single candidate is written down. They are also, in our experience, the questions founders skip.

Positioning. What is this product, in five words a stranger would understand? "Issue tracker for software teams." "Payments infrastructure for the internet." "AI pair-programmer." The naming brief lives downstream of the positioning brief. A name briefed against fuzzy positioning produces fuzzy names — and the fuzz compounds when the brief moves from the founder's head to a brainstorm with eight other people.

Audience. Who buys, and what does their vocabulary feel like? Names targeted at infrastructure engineers should sound different from names targeted at high-school students should sound different from names targeted at private-wealth clients. "Linear" lands with one cohort and not another; "Loops" lands with a different cohort. Cohort-fit is not a vibe — it is a checkable property, and the cohort canon for B2B SaaS is the subject of our SaaS patterns piece.

Emotion. What feeling does the name need to evoke at first encounter — competence, warmth, danger, calm, speed, gravity? The feeling shapes the phonemes. Words built on hard plosives (k, t, p) read fast and sharp. Words built on liquids and nasals (l, m, n, r) read warm and slow. The Sapir / Köhler / Maurer / Ćwiek line of phonosemantic research, which we summarize in Sound Symbolism for Brand Naming, quantifies what a sensitive copywriter has always known.

Linguistic style. Real word, coined word, compound, foreign borrowing, mythology, deliberate misspelling? Each move opens different doors and closes different ones. A real English word maximizes meaning and minimizes domain availability. A coined word maximizes distinctiveness and minimizes first-encounter comprehension. The move is a deliberate trade — not a default.

Constraints. What must be true before a candidate goes on the shortlist —.com required or aftermarket acceptable, USPTO clearance required in which Nice classes, social handles acceptable as @brand or acceptable as @brandhq, foreign-market launch in years one through three. Constraints are the rebrand-prevention layer. Constraints that aren't written down get forgotten in week 14, which is when the Office Action arrives.

The four canonical frameworks

The encoded craft of the world's best naming agencies — Lexicon's phoneme-association framework, Operative Words' creative-move taxonomy, Eat My Words' SMILE/SCRATCH rubric, and Catchword's clarity model — informs the methodology Etymolt scores against. None of those frameworks is a secret. They have been described, in print and on conference stages, by their authors. Few founders have read them, because the founder's naming process is typically a Friday night in week 14 of a launch, not a graduate seminar.

Lexicon (David Placek's shop, the firm behind Pentium, BlackBerry, Subaru, Swiffer, Sonos, and Dasani) treats a name as a sequence of phonemes that prime an association. The work is to find phoneme sequences whose acoustic shape already sounds like the brand promise — the K of speed, the soft labials of comfort, the long vowels of luxury. The framework is the most rigorous published account of why a word feels the way it feels.

Operative Words (Brooklyn) publishes the cleanest taxonomy of creative moves. Their categories — Metonym, Gerund, Verb, Compound, Mythology, Found-object, Translation, Minimalism, Misspelling — are the vocabulary we use when we tell a founder "your shortlist is heavy on Metonym and light on Verb; that's why every candidate feels static." The taxonomy is the diagnostic.

Eat My Words (Alexandra Watkins, Sausalito; author of Hello, My Name Is Awesome) gave the industry the SMILE acronym for what a great name does (Suggestive, Meaningful, Imagery, Legs, Emotional) and the SCRATCH acronym for what a bad name suffers (Spelling, Copycat, Restrictive, Annoying, Tame, Curse, Hard-to-pronounce). It is the most memorable mnemonic in the field and the closest thing to a shared rubric.

Catchword (Oakland) anchors on the clarity model — names that telegraph the category quickly with the right amount of strangeness. Their published work emphasizes the balance between distinctiveness and comprehensibility, and the cost of leaning too far in either direction.

Each framework illuminates a different surface of the same shape. We unpack all four, with worked examples, in The Best Brand Naming Frameworks.

The seven articles in this cluster

Each article below answers a single, narrow question. The cluster is meant to be read in any order; the pillar is the map.

• Process

  How to Come Up With a Brand Name — the 7-Step Process

  Positioning, audience, emotion, linguistic style, constraints, brainstorm, verify. The full process behind Linear, Stripe, Notion, Cursor. The three founder mistakes to avoid.

  9 min read →

• Criteria

  What Makes a Good Brand Name — the 6 Criteria

  Distinctive, pronounceable, memorable, available, meaningful, cohort-fit. Evidence drawn from the 1.4M-name corpus. Anti-examples that fail each criterion.

  8 min read →

• Frameworks

  The Best Brand Naming Frameworks (Lexicon, SMILE/SCRATCH, Operative)

  Lexicon's phoneme-association framework, Operative Words' creative-move taxonomy, Eat My Words' SMILE/SCRATCH rubric, Catchword's clarity model. Each mapped to a real brand decision.

  10 min read →

• Comparison

  AI Brand Name Generators Compared (Namelix, Looka, Squadhelp, Etymolt)

  Namelix, Looka, Squadhelp, Brandbucket and Etymolt's verification layer. What generators do well, what they miss, and how the generation/verification stack actually splits.

  7 min read →

• Patterns

  SaaS Brand Naming Patterns (Linear, Vercel, Resend, Cursor)

  The 32-name YC/B2B SaaS canon analyzed by creative-move: Metonym, Gerund, Verb, Compound, Mythology, Minimalism, Found-object. What's working in 2025-2026.

  9 min read →

• Patterns

  D2C Brand Naming Patterns (Allbirds, Casper, Warby Parker, Glossier)

  What D2C naming gets right that B2B forgets. The cultural-meaning weight in consumer brands, the comfort of human-sounding names, and what differs from SaaS.

  7 min read →

• Mistakes

  The 7 Brand Naming Mistakes Founders Make (and How to Avoid Them)

  Descriptive trap, period-marking, cohort-mismatch, sound-symbolism backfire,.com obsession, handle blindness, late verification. The founder's Office Action story belongs here.

  8 min read →

How Etymolt fits in

The market splits naming work into two halves. Generators — Namelix, Looka, Squadhelp, Brandbucket, the LLMs — produce candidates. Verifiers attest them. Etymolt is the verifier. We do not invent names; we run any candidate (from a generator, from a brainstorm, from your shower) through five axes and return a verdict you can show an attorney, an investor, or a co-founder.

The five axes are trademark (USPTO TESS, TTAB, Madrid, with collision scoring and §2(d) refusal probability), domain & handle (registrar APIs, GitHub, X, Instagram, the canonical 12), cultural meaning across 20 markets, sound symbolism against the Sapir / Köhler / Maurer / Ćwiek corpus, and pronunciation resilience via a TTS→Whisper round-trip across 12 accents. Each axis returns a score; the scores compose into a single Clearance Confidence Score with a verdict — PROCEED, STRATEGIC, or ABANDON.

Every flag traces to a record number. A USPTO collision cites the live serial; a cultural flag cites the lexicographic source; a pronunciation flag returns the audio file of the misheard variant. The output is a receipt, not a verdict with vibes attached. The full methodology is published at /methodology and recalibrated weekly.

The legal posture is the Bureau Model: we surface signals; we are not a law firm. Attorneys partner with us — we save their associates time on prep, they keep the opinion letter and the regulated work. Founders use us at the candidate-screening stage; attorneys use the dossier as their preflight. The product is attested research, not legal advice.

The arithmetic of a name

A name is checked thousands of times over the life of a company — every customer first-encounter, every podcast mention, every legal filing, every cap-table line item, every Hacker News thread. The cost of a bad name is paid in fractional decrements over each of those encounters. The cost compounds because brand recall compounds.

In the corpus, names that score above 85 on the Clearance Confidence Score show a measurable bump in first-touch attention that persists across the company's first three years. Names that score under 50 show a measurable drag at the same touchpoints. The bump and the drag are small in any single encounter — but the company has a million encounters in its first year, and the integral of the difference is real.

The framing we use internally: a name is a perpetuity. The annual coupon is small; the present value over the company's life is large; the cost of changing the instrument is one-time and severe. The right naming process treats the name as a ten-year bond, not a Friday-night sketch.