5 Tests Every Startup Name Should Pass Before You Register the Domain
Most startup names are chosen by elimination. Founders generate options, cross off the ones that feel wrong, and register whatever is left with a .com. The process is fast, cheap, and produces names that are mostly fine -- which is exactly the problem. A name that is "mostly fine" is not doing any work for you.
Before you commit to a name, five tests will tell you whether it is merely inoffensive or genuinely effective. Most shortlisted names fail at least two of them.
Why founders pick the wrong names
The dominant selection criteria for most startup names are availability (the .com is free), aesthetics (it looks good in Helvetica), and the absence of obvious problems (nothing offensive, nothing too generic). These are all necessary conditions, but none of them are sufficient.
What founders rarely test is whether the name does anything. Does it carry a signal about the product's character? Does it survive being said out loud in a business context? Does it work in the markets where you plan to grow? Does it sit in the rare territory between familiar and surprising?
The five tests below measure exactly this.
Every name carries a phonetic signature -- a set of acoustic properties that prime subconscious associations before the listener has processed a single word of meaning. Hard stops (K, T, P) project precision and authority. Fricatives (F, S, SH) project speed and elegance. Liquids (L, R) project smoothness and flow. Nasals (M, N) project warmth and reliability. Open vowels (A, O) project confidence and energy.
The test: do your name's phoneme properties match your product's character and category? A payments infrastructure company with a soft, warm name (imagine a hypothetical "Mellow" or "Nurra") is fighting against its own phonetics every time the name is used. A wellness app with a name built on hard stops and fricatives ("Striktly", "Cepsis") creates a similar mismatch.
Stripe passes this test with exceptional clarity. The /str/ onset is structural and precise; the /aɪ/ diphthong is tight and decisive. It sounds like financial infrastructure before you know what it does. Notion passes for a different reason: the /n/ onset and the open /oʊ/ vowel carry warmth and approachability -- exactly right for a productivity tool that needs to feel accessible rather than imposing.
A name has to survive at least three different sentence contexts in which founders and customers will use it. If it sounds wrong in any of them, it will create friction every time that context occurs.
Say your candidate name out loud in each of these sentences:
The name that works in all three is usually the right one. Names that fail the first context tend to be too cute or too casual for business use. Names that fail the third tend to be too formal or too hard to say quickly. Names that fail the second tend to be too category-descriptive -- they sound like a feature, not a company.
Slack passes all three. "We are Slack" works. "Slack just raised $160 million" works (this actually happened). "Join Slack today" works. The name sounds like a real entity in every context despite -- or because of -- the fact that "slack" is a common English word with an unexpected register in a business context.
If you have any international ambitions -- and most software products do -- your name will be encountered by speakers of languages other than English. Some phonemes that are unremarkable in English are rare, absent, or carry unintended associations in other languages.
A name built primarily on the American /r/ -- "Ruvr", "Corrix", "Ferrix" -- will be consistently mispronounced in East Asian markets and will look different in phonetic transcription in every market. This does not necessarily disqualify the name, but it should be a conscious decision rather than an oversight discovered after launch.
Zoom passes this test well. /z/, /uː/, /m/ are all phonemes that exist in or can be approximated cleanly across most major language families. The name sounds similar in Mandarin, Spanish, German, and English.
The most durable brand names occupy a specific territory: unexpected but immediately believable. They do not describe the product. They do not feel invented for the sake of being novel. They feel like something that could have existed for a long time but was never applied in quite this way.
This is what naming strategist David Placek calls the tension zone. A name with no tension -- one that tells you exactly what the product does -- is forgettable because it blends into the category. "Taskflow", "Clearpath", "BrightSpace" all fail this test. A name with too much tension -- arbitrary coinages with no phonemic anchor -- is also forgettable because it provides no handle for memory to grip. "Zoplux", "Forvix", and "Greptia" fail for this reason.
The ideal is a word or root from an adjacent domain, applied with confident precision to your category. Stripe is a visual mark applied to payments. Slack is the amount of rope that is not taut, applied to communication. Vercel is a compression of "versal" (universal) applied to deployment infrastructure. Each of these carries semantic tension -- the referent is not obvious, but once you see the connection, it feels inevitable.
To test your candidate: if someone who knows nothing about your company heard just the name, would they ask "what does it do?" without being able to guess? If the answer is yes AND they could easily imagine it as a company name, you are in the tension zone. If they could guess what it does from the name, you have a descriptive name. If they cannot imagine it as a company name at all, you have a coinage problem.
This is the most practical test and the one founders typically run first -- which is backwards. There is no point running trademark and domain searches on a name until it passes tests one through four. A name that fails on phonetic fit, context, cross-language risk, or tension is not worth the $15 to register the domain.
For domain: the .com is still the default signal of legitimacy for most B2B software products. A product with a .io, .co, or .app domain is not disqualified, but it starts with a trust deficit in enterprise sales contexts. If the .com is taken, check whether it is actively used, parked for sale, or dormant. Dormant .coms are often acquirable for $2,000–$5,000.
For trademark: search the USPTO database for your name in the relevant international classes (class 42 for software-as-a-service is the most common). You are looking for exact matches and close phonetic matches -- a name that sounds similar to a registered trademark in your category is a liability even if the text is different. This search should be done before you make any public use of the name, not after you have printed business cards.
The practical sequence: run tests one through four first. For every name that passes all four, run the domain and trademark check. You will waste less time and find better names.
How many names to test
The practical answer is: more than you think. Most founders begin with five to fifteen candidate names. The problem is that when you generate names yourself, you unconsciously filter toward names that feel familiar -- words you already know, phoneme patterns you are comfortable with, names that resemble brands you admire. This produces a pool that is already biased toward the safety of familiarity.
A broad generation pass -- 300 candidates or more -- before any evaluation ensures that the evaluation is selecting from a full distribution of possibilities rather than a pre-filtered subset. The names that score highest in a large pool are almost never the ones you would have come up with in a whiteboard session.
The phonetic fit test, cross-language test, context test, and tension test are all scored automatically by Voxa's analysis engine. Run any name through the free analysis to see its scores across 14 dimensions -- or get a full 300+ candidate proposal calibrated to your brief.
Analyze your name freeA note on testing order
Run the tests in the order listed. The phonetic fit test eliminates the most names quickly and requires no research. The context test is free and can be done in sixty seconds. The cross-language test requires knowing your target markets but no tools. The tension test requires honest assessment but again no research. Only names that pass all four warrant the time investment of domain and trademark searches.
Most shortlists of five names will have one or two that pass all five tests. Start with the phonetic fit test and the context test -- they will cut your list by half within five minutes.
The goal is not to find the perfect name. It is to find a name that does not fight your positioning, survives the contexts in which it will actually be used, does not create problems in your target markets, and carries enough distinctiveness to be remembered. That is a high bar, and it eliminates most of the names founders are considering at any given time. Which is precisely why it is worth applying before you commit.
Related reading