How a name gets judged, and on what evidence
Scopeweb grades names. This is the page that says how, so the grade can be argued with instead of believed. Every component below reports its score, its weight and its basis.
Our own name, and the lowest-scoring one we hold. Here is the whole calculation.
Length
weight 0.18Character count of the label. Shorter is easier to type, say and remember.
Syllables
weight 0.12Estimated by vowel-group heuristic, not a phonetic dictionary. Treat it as approximate.
Pronounceability
weight 0.20Vowel ratio, consonant-cluster length and repeated letters. The radio test: can someone type it after hearing it once?
Dictionary presence
weight 0.16Checked against a published 3–7 letter English word list. Real words are memorable; invented names are more ownable. This scores recognisability, not desirability — neither result is better.
Brand collision
weight 0.20Levenshtein distance to the nearest entry in a published 68-brand list. Confusability only. This is not a trademark search.
Hyphens and digits
weight 0.08Both survive typing but not speech. They have to be spelled out aloud every time.
Ending perception
weight 0.06A stated judgement from a published table, not a measurement. Argue with the table — that is why it is published.
What this number is not
It is not a prediction of search ranking. Exact-match-domain value has been largely dead since Google's 2012 EMD update, and nothing in this calculation forecasts how a name will perform in results. Anyone selling you a domain on ranking grounds is selling you 2011.
It is not legal clearance. Brand collision is string similarity to a published list. A name can score 100 there and still infringe a live mark.
What we cannot measure
These are absent from every score. They are reported as unavailable rather than estimated, and they are never quietly folded into the total.
Zone rarity
How contested a root is across an entire zone — how many registered domains already contain it. Needs zone-file access; a CZDS application is pending.
Trademark collision
Live marks in USPTO or EUIPO matching the label in relevant classes. No trademark database is integrated. If you are naming a company, this is the check we cannot do for you and you should not skip.
Connotation
What a word makes people feel. The example above is the case in point: “creeper” means someone who watches without consent, and the score of 84 cannot see that. A number will not catch a name that is fine to say and wrong to mean.
Severe properties cap the score
Most components are averaged. A few are not. A name one edit away from a major brand cannot score well no matter how short or sayable it is — that property sets a ceiling instead of being diluted by six good numbers around it.
When a cap applies, the score says so by name. A capped 40 and an averaged 40 mean different things and are not presented alike.
Why publish the weights
A score you cannot inspect is a score you have to trust, and there is no reason to extend that trust to a company that profits when you buy. Publishing the components means you can disagree with a specific number instead of the whole verdict — decide that four syllables does not matter for your product, discount that row, and reach your own conclusion.
That is the intended use. The number is an input to your decision, not a replacement for it.