How to Test a Company Name Before You Launch

Figma's founders almost named the company "Hemlock." They liked the sound. It felt distinctive. Then someone pointed out that hemlock is a poison. Specifically, the poison that killed Socrates. The association was fatal. They went with Figma, and the rest is an acquisition worth $20 billion. The point is not that Figma is a better word. The point is that they tested before they committed.

Most founders skip testing entirely. They brainstorm, pick a favorite, register the domain, and move on. Six months later, they discover that customers cannot spell it, that it means something unfortunate in Portuguese, or that they cringe every time they introduce themselves. By then, the switching cost is real.

Five tests. Each takes less than an hour. Together, they will either confirm your choice or save you from a costly mistake.

Test 1: The Phone Test

Call someone who has never seen the name written down. Say: "I am starting a company called [name]. Can you spell that back to me?" Do this with at least five people. Not co-founders, not family. People who have no context.

If three out of five cannot spell it correctly on the first try, the name has a friction problem. Every misspelling is a lost Google search, a mistyped URL, a failed email delivery. Stripe passes this test perfectly. Xobni (inbox backwards) did not, and eventually rebranded. MongoDB passes; Kubernetes does not (which is why everyone calls it "K8s").

The Phone Test also reveals phonetic weaknesses that look fine on paper. "Clarius" and "Clarios" sound identical over the phone. If a competitor's name sounds like yours in a conversation, you will lose referrals to mispronunciation.

Test 2: The Bar Test

Go to a loud environment. A coffee shop, a conference networking session, a bar. Introduce yourself: "I work at [name]." If the other person asks you to repeat it, note that. If they look confused, note that. If they nod and move on naturally, the name is working.

This test measures what acousticians call signal-to-noise ratio. A name needs to cut through ambient noise without shouting. Short names with hard consonants (Stripe, Slack, Snap) perform well. Names with soft consonants and vowel clusters (Aiueio, Flouuu) disappear in noise.

The Bar Test also catches cadence problems. Names that require emphasis on the wrong syllable feel unnatural when spoken casually. "Accenture" works because the stress falls on the second syllable, which is how English speakers naturally handle three-syllable words. A name where the stress is ambiguous ("is it DEsolo or deSO-lo?") will be mispronounced forever.

Test 3: The Google Test

Search the name in quotes on Google. What dominates page one? There are three possible outcomes, and only one is good.

Outcome 1: Empty or nearly empty results. This is the ideal. It means the name is unique enough to own its search results from day one. Invented names like "Figma" or "Calendly" had clean results before launch.

Outcome 2: Dictionary entries and Wikipedia articles. This means the name is a real word. You will be competing with the entire internet's usage of that word. "Notion" pulled this off. "Linear" pulled this off. Both had significant marketing budgets. If you are bootstrapping, owning a common word in search results is expensive.

Outcome 3: A dominant existing entity. This is a deal-breaker. If page one belongs to another company, especially one in an adjacent space, you will spend years and significant money trying to outrank them. Move on.

Also search "[name] + [your category]." If that combination surfaces a competitor, you have a positioning collision that goes beyond SEO.

Test 4: The Foreign Language Test

Google Translate is not enough. It catches direct translations but misses slang, phonetic associations, and cultural connotations. The real test requires native speakers.

Identify your top five markets by revenue potential. For most B2B SaaS companies, that is the US, UK, Germany, France, and either Japan or Australia. Find one native speaker from each market. Ask them three questions: What does this name sound like in your language? Does it remind you of any existing brand or word? Would anything about this name make you uncomfortable?

The Chevy Nova story is often cited here: "no va" means "doesn't go" in Spanish, so the car flopped. In reality, the Nova sold fine in Latin America. Spanish speakers no more parse "Nova" as two words than English speakers hear "notable" as "no table." The myth persists because it sounds plausible, which is exactly the point: phonetic associations do not need to be literal to create doubt. Mitsubishi learned a real lesson with the Pajero, an actual obscenity in Spanish-speaking countries. They renamed it "Montero" for every Spanish market. Subtler traps exist too. A perfectly fine English name might sound like a medical condition in German or a children's cartoon in Japanese.

For a deeper treatment of linguistic screening, read our naming for international markets guide.

Test 5: The CEO Test

Stand in front of a mirror. Say: "I am the CEO of [name]." Say it ten times. Now imagine saying it 10,000 more times over the next five years. In board meetings. On podcasts. To your parents. To skeptical enterprise buyers. To journalists who will mispronounce it on air.

This test is not about vanity. It is about durability. A name you are slightly embarrassed by will erode your confidence in subtle ways. You will hesitate before saying it. You will over-explain it. You will add qualifiers ("it is actually a play on..."). Each of those micro-hesitations costs you credibility.

Jeff Bezos named his company Amazon because he wanted something that started with "A" (for directory listings) and evoked scale. He has said the name thousands of times over 30 years. It still works. The CEO Test is really a test of longevity: not whether the name is clever, but whether it is wearable.

There is a subtlety here. The best names often feel strange at first. "Google" sounded ridiculous in 1998. "Spotify" sounded made-up in 2008 (it was). If a name passes the other four tests but feels slightly odd during the CEO Test, give it a week. Strangeness fades. Structural problems do not.

What to do with the results

Score each finalist across all five tests. A name that fails one test is a risk. A name that fails two is probably dead. A name that passes all five is rare and worth committing to quickly.

The most common mistake after testing is going back to generation. Testing is a filter, not a creative exercise. If none of your finalists survive, the problem is usually upstream: the brief was not specific enough, or the naming territories were too narrow. Go back to the strategic foundations before generating more candidates.

Nomenco builds pronunciation testing and linguistic screening directly into the naming workflow, so candidates are validated before they reach your shortlist. See how the method works.