How to Name Your Business or Startup (Without Regretting It in Year Two)

Q: What is "How to Name Your Business or Startup (Without Regretting It in Year Two)"?

A good business name passes three tests: it is easy to say out loud and easy to spell from hearing it (the "radio test"), it is short enough to fit in a logo and a domain (under 12 characters), and it does not already mean something embarrassing in the languages your customers speak. Generate a batch of candidates with the in-browser name generator, narrow to 3-5, then check the .com and the trademarks. Do not commit to a name until those three checks pass.

The name is the first thing anyone hears about your business, and it is the last thing they remember. Most of the work in choosing a name is in the rejection: kill 90% of what you come up with, and you are left with names that work. The good news is that the rejection criteria are simple, the checks are free, and the brainstorming takes an afternoon, not a week.

The three tests a business name has to pass

Every good business name passes three tests. Run every candidate through all three before you invest any time in it.

The radio test. Say the name out loud. If you were introducing the business on a podcast, would the listener be able to spell it? "Stripe" passes. "Hooli" passes. "X" passes (after the listener sees it once). "Quibi" passed, mostly. "Zwift" passes. If a listener cannot spell it from hearing it once, the name costs you word-of-mouth — every conversation about the business becomes a spelling lesson.
The logo test. Can the name fit in a 200-pixel square logo? Is it readable at 16px? Can it be set in the same typeface as the rest of your brand? Names under 12 characters handle all three. Names over 16 characters usually fail at least one. The sweet spot is 5-10 characters for tech products, 10-15 for service businesses.
The translation test. Does the name mean something embarrassing in Spanish, French, German, Mandarin, Japanese, or Arabic? "Cum" is the Latin for "with", which is fine in English but is a vulgarity in several Romance languages. "Bum" is harmless in English and rude in German. The quick check is a 5-minute Google Translate pass over the major languages your customers speak.

About 80% of the names you generate will fail one of these three. That is the funnel working.

What makes a name good in the first place

Names that pass the three tests usually have one of these properties. Pick the property that matches the personality you want the business to have, then generate candidates in that style.

Two-word, common + specific. "North Star", "Sweetgreen", "Stripe", "Linear". One common word for tone, one specific word for clarity. The pattern works for almost every business.
Single invented word. "Spotify", "Zapier", "Notion", "Figma". Real words mashed together, or a made-up word that sounds like a word. The risk is the trademark search; the upside is a name no one else has.
Founder name or place name. "Tesla", "Nvidia", "Patagonia", "Hershey". Works when the founder is the brand. The risk is the personal association.
Metaphor or concept. "Amazon", "Apple", "Slack", "Basecamp". The name points at the deeper idea. Powerful but hard to invent without a strong point of view.

For most small businesses and side projects, the first two patterns are the right pick. The third works for personal brands. The fourth is the move for the well-resourced launch.

How to generate a batch of candidates in 10 minutes

The fastest path is a name generator that pulls from curated word lists and combines them in a few composition patterns. The Business Name Generator does this in your browser — pick a style (modern, classic, playful, technical), pick a count, and get a batch of names with the .com suggestion. Generate 30-50 names, then run them through the three tests above. Expect 5-10 to pass.

Why browser, not a SaaS? Because the name brainstorm is the part you do not want to leak. The business you are naming is the business a competitor could pre-empt. A browser-based generator that does not log your candidates keeps the brainstorm private.

How to find an available .com

Once you have a shortlist of 3-5 names, the .com check is the first hard gate. The .com is still the default for most businesses; if you cannot get the .com, the alternative TLD (.io, .co, .app) costs you trust for users who type the .com by reflex.

The check is two steps:

WHOIS lookup on the exact domain (e.g. northstar.com). The WHOIS Lookup tool does this in your browser. If the domain is taken but parked (no real site), the owner may be open to selling. If the domain is in active use, move to the next candidate.
DNS lookup on the candidate domain. If the domain resolves to a real site, it is in active use. If it does not resolve, it is either available or the owner has it parked without pointing it. WHOIS confirms.

Do not pay for a domain search service. WHOIS and DNS are public. The only time a paid broker makes sense is when the perfect .com is squatted by a known domain investor and you have a budget for it.

The trademark check you should do before you commit

WHOIS tells you the domain is yours. It does not tell you the trademark is yours. The trademark check is a separate step, and it is the one most founders skip.

USPTO TESS search (US) or your local trademark office. Free. Search the exact name in the live database. If there is a live registration in the same class as your business, you cannot use the name without infringing.
Google search for the name in quotes. If a company of any size is using the name in your category, you are taking a risk even without a registered trademark. Common-law trademarks exist.
Social media handle check. If a competitor has the @name handle, you will lose brand consistency. Decide whether to live with @somethingelse, buy the handle, or pick a different name.

The trademark check is the one a lawyer does for you. The USPTO search is something you can do in 10 minutes. Do the search before you spend any money on the name.

The final shortlist test

Once you have a name that passes the three tests, has an available .com, and clears the trademark search, you have a candidate. The last test is the one that actually picks the winner:

Imagine the business card, the website, the email signature, the support email ([email protected]), the Slack channel, the pitch slide. Does the name work in every one of those? Names that look great in a logo but read badly in a paragraph (or vice versa) cost you later.
Say the name out loud ten times. If you stumble once, the name is not done yet.
Sleep on it for a week. If you are still happy on day eight, you are probably not going to regret it in year two.

Most founders pick a name in an afternoon and regret it in a year. The afternoon is fine. The day-eight check is the part that catches the bad picks. Sleep on it before you commit.