How to Name a Startup

Why naming a startup is not naming a business

A restaurant, a plumbing firm, and a consultancy can all be named well by asking what the business does and who it serves. A startup cannot, because both answers are provisional. You are naming something that intends to change shape, sell across borders early, and be spoken aloud in rooms where nobody has seen it written. Five pressures follow, and they are why a general business-naming checklist leaves founders with the wrong shortlist.

• It has to survive a pivotMost businesses end up doing roughly what they set out to do. A startup is still looking for its model, and the search moves it. Samsung began in 1938 as a Daegu trading company dealing in dried fish, produce and noodles, and because the word means “three stars” in Korean it needed no adjustment when semiconductors arrived. Stripe spent its first months as /dev/payments. Shopify was Snowdevil, a snowboard shop, until the software underneath turned out to be the business.
• It is global from day oneA corner shop’s name only has to work on its street. A startup launching on the web is read the same week in markets you have not planned for, by people who guess the pronunciation and type what they hear. A word that collides with slang somewhere you sell is cheap to drop while it is one of a hundred candidates and expensive once invoices carry it. A screen reports no flags in the languages checked; nothing reports that a name is safe.
• It gets decided on a deadline, not by consensusNaming expands to fill whatever time you give it, and a startup rarely has that time. Blue Ribbon Sports became Nike in 1971 because a factory needed a name to stamp on the boxes: Phil Knight preferred a different candidate and accepted Nike anyway. Stripe’s founders, by Patrick Collison’s telling, set a date in December 2010 and agreed to keep the name they had if nothing better arrived. Neither name was reached by everyone agreeing.
• It is said aloud far more than it is readDemo days, investor intros, a podcast host reading it cold, a voicemail to somebody’s procurement team. For its first years a startup name travels mostly through air, and every time a listener asks for the spelling or needs the joke explained, the name charges rent.
• You are naming the company, not the first productWhat you launch with is rarely what you sell in three years, and a name that fits today’s product will argue with the second one. By then it is on contracts, in a store listing, and in your customers’ mouths. Decide up front whether the word sits over everything you will ship or only over the first thing.

Three of those pivots are worth the full read: why it is called Samsung, why it is called Stripe, and why it is called Shopify. The deadline that produced Nike has its own breakdown. If the company-or-product question is still open, settle it first with naming a product vs. naming a company, and once you know your markets, the screen to run is the cross-language safety checklist.

The startup naming process, in six steps

Each step is short and hands the depth to the guide that owns it. What this section owns is the running order, which is the part founders get wrong: every step that generates comes before every step that judges, and judging comes before spending money.

1. Write the brief, with the decision date in itOne page before any candidate exists: what the product does, who buys it, the markets you sell into, three adjectives the name should carry, what it must never be mistaken for, and the date you will decide. It is what co-founders point at instead of at each other’s taste. Budget an hour.
2. Generate far wider than feels reasonableAim for roughly a hundred candidates from at least four construction methods, and judge none of them while you are generating. The first dozen you write are the ones your category already uses, and they are worth writing to clear them out. The survivors arrive once the obvious patterns are exhausted.
3. Cut by style before you cut by tasteDecide where you stand between describing what you do and standing apart from it, then cut everything on the wrong side of that line. Descriptive names are found by people searching the category and are hard to own; abstract names are ownable and cost you a stretch of explaining. A shortlist holding both is what makes the final choice feel impossible.
4. Score the shortlist on facts, not impressionsEight to twelve survivors, scored on the same criteria, on the same day, written down. Domain position, memorability, and how the word behaves in the mouth and in the languages you sell into are measurable rather than arguable. Impressions drift toward whichever candidate somebody said aloud most recently. Numbers hold still.
5. Break the tie with a rule you set in advanceTwo finalists usually means both are fine and the team is tired. Agree the tiebreak before you arrive at it: the wider domain gap wins, or the market you actually sell into wins. A rule agreed in advance costs ten minutes; one invented in the meeting costs a fortnight.
6. Secure it in one sittingThe hour after the decision is the only one when everybody is aligned and nothing has leaked. Take the domain, the handles, the company registration and any code namespaces in that hour, and get a proper trademark search done before you spend on packaging or ads. Choosing a name does not make it yours.

Every step after the first has a guide of its own. How to brainstorm brand names covers step two, with seven construction methods and a session format that stops a group converging too early, and there are fifty worked examples if you would rather start from a list. Descriptive vs. abstract brand names is the style axis in step three, how to test a brand name is the evaluation battery for step four, and how to choose between two business names is the tiebreak ladder for step five. Step six, the registrations and the order to do them in, is how to secure a brand name.

Building a developer tool, an AI product, or anything with a package namespace? The sub-sector conventions and the namespaces a free .com will not warn you about are in how to name a tech startup, and the extension question is in tech startup domain names.

The startup name checklist, before you commit

This is the last gate, run once on the name you have chosen rather than on all hundred candidates. Every item is a fact you can confirm in an afternoon, which is the point: by this stage nobody should be relying on how a word feels. A line that fails is a cost you found in time.

• It survives one hearingSay it once to somebody who has never seen it written and ask them to type it. Corrections are not taste, they are customers landing somewhere that is not yours.
• It survives a bad connectionSay it into a phone in a noisy room, the way you will at a conference. Soft consonant contrasts (f against s, m against n) blur exactly where it matters most.
• The languages you sell into came back quietScreen it against the markets you actually enter, not the aspirational ones. The honest result is no flags in the languages checked, and that is what you write down.
• The domain position is decided, not hoped forKnow which extension you launch on and what you do if the .com moves. Any lookup is a snapshot worth confirming at a registrar, because registries change by the minute.
• A knock-out trademark search came back quietSearch the public registers for your classes, and get a professional opinion before you spend on anything printed. A free domain is not legal clearance.
• The handles and namespaces existThe social handles you will actually use, plus the package name and organisation handle for a technical product. First-come surfaces, quoted forever in other people’s code.
• It still fits if the product changesRead the name beside a product you might ship in three years. If it needs explaining in that sentence, you have named the launch rather than the company.
• It is not simply the category wordIf the name is the thing you sell with a suffix attached, people searching for you will find competitors instead, and you will spend years buying back your own traffic.

Two of those lines have guides behind them: trademark basics for startup naming explains what a knock-out search can and cannot tell you, and is your brand name too generic gives you four tests for the last item. Neither is legal advice.

What should I name my startup? Four questions that narrow it

Nobody can hand you the word, and any page claiming otherwise is selling a generator. What can be handed over is the set of questions that turns a hundred candidates into a dozen with no argument about taste. Answer these in writing, in the brief, before you generate anything.

• What should a stranger picture?Not what you want them to know, what you want them to see. A name carrying a picture (a mountain, a tool, a bird) travels further than one carrying an adjective, because pictures survive repetition.
• How far from the category do you want to stand?Close in, you are legible on day one and interchangeable by year three. Far out, you own the word and pay for the explanation. Pick an end deliberately and write it into the brief.
• Who says it, and in which language?List the people who will pronounce it most: buyers, your support team, investors, a journalist reading it off a release. If a language on that list is not one you speak, the screen is not optional.
• What does it cost you when you pivot?Assume the product changes once. A name with no meaning welded to it costs nothing when that happens; one describing today’s feature costs a rename, and renames cost traffic and recognition.

All four rule candidates out rather than in, which is why they work where “which name do we like best” does not. For a sense of what the answers look like when they went well, the patterns behind the names people keep citing are unpacked in what the best startup names have in common.

What a score settles, and what it does not

Step four is the one a machine can take, because everything it measures is a property of the word rather than an opinion about it. Namoly scores a name on four weighted categories: domain availability (0.30), memorability and distinctiveness (0.25), phonetics and readability (0.25), and cross-language safety (0.20). The overall is the weighted mean, and the bands are 80–100 Strong, 60–79 Good, 40–59 Fair, 0–39 Weak.

The analysis is deterministic, so the same name returns the same scores for the same reasons every time, which is what makes ten candidates comparable. Every category carries a justification and a confidence signal, so a thin result never presents itself as a confident verdict.

What it does not settle: whether anybody else holds a legal right to the word, whether the handles are free, whether your package name is taken, and whether the name is right for the company you intend to build. The first three are manual checks in step six. The last is the founder’s call, because a score ranks a shortlist but cannot hold a position.

Running an analysis is free and so is re-running one. Read the justifications and not only the numbers: the reason a name scored 62 is more useful than the 62.

Frequently asked questions

How do I name my startup?

Run six steps in order: write a one-page brief that includes the date you will decide, generate around a hundred candidates without judging them, cut by style in one pass, score eight to twelve survivors on identical criteria, break the tie with a rule agreed in advance, then secure the domain, handles and registrations in one sitting. The order matters more than the creativity.

What should I name my startup?

Nobody can hand you the word, but four questions narrow it fast: what a stranger should picture when they hear it, how far from your category you want to stand, who has to pronounce it and in which language, and what the name costs you if the product changes. Answer those in the brief and most candidates rule themselves out.

How long should naming a startup take?

About a week of deliberate work: an hour on the brief, an evening generating, one pass to cut, a day to score, then a decision. Projects that run for months are missing a stopping rule rather than a good name, so put the decision date in the brief on day one.

Should a startup name describe what the company does?

It depends which cost you would rather pay. A descriptive name is understood immediately and is hard to own, hard to trademark, and awkward after a pivot. An abstract name is ownable and easier to protect, and costs you a period of explaining what you sell. What breaks a shortlist is comparing both as if they were the same bet.

Does the .com matter when naming a startup?

Less than it did, and more than founders hope. Plenty of funded companies launch on another extension, and the exact-match .com still absorbs traffic you would otherwise buy. Decide the position before you fall in love with a candidate: which extension you launch on, and what you do if the .com is not obtainable.

What does Namoly check, and what does it not?

Namoly scores a name on four deterministic checks: domain availability, memorability and distinctiveness, phonetics and readability, and cross-language safety. It does not check trademarks, social handles or package registries, and it never calls a name safe: a language screen reports no flags in the languages checked. Running an analysis is free, and so is re-running one.