How to Evaluate a Brand Name: 6 Tests Before You Commit

In short: judge a brand name on six things (memorability, pronounceability, spelling, distinctiveness, fit and scalability) plus two checks that are easy to skip: what the name sounds like, and what it means in other languages. Taste is not on the list. Score a shortlist on all six and the argument usually settles itself. Further down, we run two real Namilio shortlists from 15 October 2026 through the checks.

Why 'which one do you like?' is the wrong question

Ask a room to pick between three names and you get three opinions and no decision, because liking a name is a property of the person, not of the name. The names that survive contact with customers are the ones that get repeated correctly, and that is something you can test. It breaks down into six dimensions you can score independently.

The six dimensions

Score each one from 1 to 10 and be strict: 5 is workable, 7 is solid, and a 9 should be rare enough that you suspect your own scoring. The value is not the total; it is seeing where a name is weak. A name that is memorable and distinctive but unspellable has a specific, fixable problem. A name that scores 6 across the board has no problem you can name, which usually means it is forgettable.

The spelling test is the one that matters most

Spelling is the dimension founders are worst placed to judge, because you already know how your name is spelled and cannot hear the ambiguity. A customer who hears your name in conversation and cannot reproduce it searches a plausible spelling instead, may find someone else, and you never learn it happened. Silent letters, doubled consonants and a choice between i and y are the usual offenders. If two reasonable spellings exist, either pick a different name or buy both domains.

Two checks that are easy to skip

What does the name sound like?

Names carry an impression before they carry a meaning. In two studies with invented brand names, Richard Klink found that the sound of a name alone can tell people something about the product, such as its size, speed, strength or weight (Klink, Marketing Letters, 2000). So a shortlist can feel energetic or calm before anyone reads the meaning: compare the hard stops in Kodak, TikTok and Google with the softer, more open sounds of Aesop, Aveda and Serena. Neither is better; they are different promises. A wellness brand built from hard consonants is fighting its own sound.

What does it mean somewhere else?

This is the check with the largest downside and the lowest cost. A name that is clean in English can be vulgar, comic, or simply unfortunate in Spanish, French, German, Portuguese, Italian, Mandarin, Japanese, Korean or Arabic, and if nobody checks, the discovery comes after the logo, the packaging and the domain. Mitsubishi sold its Pajero as the Montero in Spanish-speaking markets because pajero is vulgar slang in Spanish (that case and others are in our study of naming mistakes that forced rebrands). Run any shortlisted name past speakers of your likely markets, or a tool that screens for it, before you commit.

Score your own shortlist. The Namilio generator produces names across 16 styles and checks domain availability on the extensions you pick, up to 22, as they appear. Its brand analysis scores any name on all six dimensions above, adds a sound profile, and runs an AI screen for unintended meanings in nine languages (a first pass, not a substitute for a native speaker). If you want to know what a name is worth as an asset, the domain appraisal tool prices any domain against 8,586 recorded sales.

The order to run the checks in

1. Spelling and pronounceability first. They are objective and they eliminate fastest. There is no point debating the fit of a name nobody can write down.
2. Then domain availability. Instant, and it removes candidates before you get attached to them.
3. Then meaning: fit, distinctiveness, scalability. These need judgement, so spend the judgement on a shorter list.
4. Then the language check on the final three or four. It is cheap per name but not worth running on forty.
5. Trademark last, because a proper clearance search is the slowest and most expensive check and should be run only on names you would actually use. See which business names are easiest to trademark for how the distinctiveness spectrum decides what you can protect.

Running this list backwards (falling for a name, then finding the domain gone, then finding it unspellable) is the expensive order. Twenty-four companies that got it wrong are documented in our study of naming mistakes that forced rebrands.

We Ran Two Briefs Through Namilio and Applied the Checks

On 15 October 2026 we gave Namilio two briefs, with the brandable, evocative and compound styles at the default creativity, and took the first 12 names each one returned. We then looked up every .com in Verisign's registration data service (RDAP), the registry's own record of which .com names exist (IANA lists Verisign as the .com registry). "Available" means nobody had registered it. The score is the engine's own 0 to 100 rating. Run a brief again and you will get different names, because the model samples.

Brief 1: “an organic skincare line”

Two of the 12 had a free .com, LeafLotion and BloomBarrier, both two-word compounds. The one-word coinages (Honew, Dermit, Botan) were all registered, and so was the engine's top-scored name, SkinSage (95). Run in the order above, both free names pass spelling and pronunciation, so the choice comes down to fit and distinctiveness: LeafLotion is close to a product description, while BloomBarrier points at a specific benefit. Honew would have failed the first step anyway: heard aloud, it has no obvious spelling.

Brief 2: “a project management tool for design agencies”

One of the 12 had a free .com: RevisionRail. Five start with Brief or Studio (Briefin, BriefBridge, BriefHarbor, StudioFlow, StudioNest), which is the distinctiveness problem in miniature: in a list of agency tools they would blur together. RevisionRail is spellable and specific, and the fit question the table asks is whether you want a product named after revisions.

The one-word options were gone for a reason: single-word category domains change hands for large sums. In Namilio's dataset of 8,586 publicly recorded sales, skins.com sold for $1,459,450 in 2025, plans.com for $275,000 in 2023 and tasks.com for $130,000 in 2015 (what domains sell for). Across both briefs, 3 of 24 names had the .com free; across all 42 briefs we ran for Namilio's guides that day, 72 of 504 (14%) did.