Why names fail in the wild, not on the slide
Most naming problems do not appear in a presentation. They appear the third time someone tries to tell a colleague about you and cannot remember the spelling. They appear when a prospect types a near-miss into a search bar and lands on a competitor. They appear in a sales introduction where the other person nods politely and writes down the wrong word.
The goal of testing is not to produce a score that decides for you. It is to surface those moments of friction while the name is still inexpensive to change. A handful of structured checks will tell you more than a room full of people saying they like it.
The spoken test: say it once, no screen
Say the name aloud once, without showing it, and ask the listener to write it down. Do not repeat it, do not spell it, do not give context. Then compare what they wrote with what you intended.
Note where they hesitate and what they guess. A hesitation on a vowel pair (“ae”, “ou”) or an unusual consonant cluster is a signal, not a verdict. Repeat the exercise across a dozen or so people who resemble your audience. The pattern matters more than any single miss.
This is the practical idea behind Namegrain’s internal phone-test rubric, which classified 45% of a 2,211-name YC directory sample in its strongest spelling-clarity band. That figure is a classification of one dataset, not a controlled listener experiment—read the methodology notes for its limits. The transferable lesson is simpler: pay attention to how a name behaves when people only hear it.
The spelling and search test
Type the name into a search engine and look at what comes back. Is your intended destination obvious, or is the results page shared with an unrelated company, a product, or a common word? Then type the likely misspellings a listener would produce and see where those lead.
You are not looking for a completely empty results page—that is rare and not required. You are looking for whether a reasonable person, searching the way they actually search, arrives at you. If every plausible misspelling lands somewhere confusing, that is friction you will pay for repeatedly in word-of-mouth.
Distinctiveness and searchability deserve a place in the brief alongside personality. Any candidate should still go through appropriate professional clearance before commitment; a clean search result is not a legal opinion.
Test in the markets you actually sell into
A name that is effortless in one language can be awkward or unintentionally meaningful in another. If you sell into more than one market, run the spoken and spelling checks with native speakers of each, including the accents and dialects that matter to the business.
Ask two questions in each market: can someone write it down after hearing it once, and does it carry any unintended association when spoken or read? The second question is best answered by people who live in the language, not by a translation tool.
Run a small listener panel, not a poll
A poll of friends optimises for politeness. A listener panel optimises for friction. Recruit a small, deliberate group—perhaps ten to twenty people who match your buyer, user or partner profile—and run the same structured exercises with each: spoken recall, written spelling, and a quick “what do you assume this company does?” prompt.
Record answers verbatim. Resist the urge to correct or explain during the session; the misunderstanding is the data. Afterwards, look for repeated patterns rather than outliers. One confused listener is noise; five confused listeners in the same place is a design problem.
Interpret results without over-claiming
These checks measure friction, not destiny. A name with a little friction can succeed with strong positioning and repetition; a frictionless name can fail for entirely other reasons. The research behind this site describes observed choices in selected samples—it does not establish that any naming property causes commercial outcomes.
So treat a failed check as a question to answer, not a veto. Can the friction be designed around with spelling cues, a consistent visual wordmark, or a clear spoken introduction? Or is it structural to the name itself? That distinction is what informs whether you keep, adjust or replace a candidate.
Turn findings into a decision
Bring the results back to the brief you started from. For each candidate, write one line on spoken recall, one on spelling and search, one on cross-market behaviour, and one on the assumptions people made about the category. Then compare candidates against the criteria you set before testing, not against whichever name the room grew fondest of.
If two candidates are close, prefer the one your audience can pass on without help. Word-of-mouth is the cheapest growth a startup has, and it runs entirely on whether a name survives being repeated.
When the testing is done and a direction is clear, the next step is a brief that carries it forward—see how to write a naming brief, or the .com decision if the address is the open question.