Word Scoring: How Domain Score Works

A negative score usually means one or two penalty rules fired hard enough to outweigh the positive ones — most often because the root is long (past 12 characters) or contains a q, x, or z (rare, hard-to- pronounce letters). A negative score doesn't mean the domain is unavailable, just that our default rules don't rate it highly as a brandable name. Click "Why this score?" on any result to see exactly which rule cost you points.

Available TLD Weights — each available extension (.com, .io, .net…) adds points, weighted by real-world value (.com highest). Dictionary Word Bonus — a small bonus when the root is a real, recognized word rather than an invented string. Short Root Bonus — roots under 12 characters earn a small bonus per character under that line. Long Root Penalty — roots over 12 characters lose points per character over that line. Rare Letter Penalty — every q, x, or z in the root costs points, once per occurrence. Confusable Bigram Penalty — letter pairs that are easy to misread in lowercase (rn, vv, cl, ii) cost points, once per occurrence. Negative String Penalty — a short, curated list of substrings that read badly when concatenated into a domain, each with its own tunable severity. Pronounceability Bonus / Hard-to-Say Penalty — a 0–100 score for how naturally the letters flow feeds a bonus above the "sounds fine" line and a penalty below it. Word List Breadth — a ladder of bonuses (+1 to +5) for words that turn up across many different kinds of word list.

Some words belong to one world and some belong to many. Apple is a food, a colour, a plant and a famous brand; tachyon is one thing only. We count how many different types of word list contain your root — food, colour, tech, animal, mythology and so on — and award +1 at three types rising to +5 at eight or more. It is deliberately a count of types, not of lists: a reference set imported as three hundred separate lists is still one source, and counting the lists would let bulk imports drown out genuine variety.

Only about 3% of words earn anything here, so when it fires it means something. Two things worth knowing: it measures versatility, not quality — so a brilliant invented name scores 0 on this rule alone, by design, and picks up its points from the pronounceability and length rules instead. And it is recomputed periodically rather than instantly — the calculation looks at every word in every list at once, so a word you add right now can take up to an hour to show up in your scores.

Because it does, in the real world — .com carries the most resale value and the most instant trust with visitors, so it's weighted highest in the default rules. If you disagree (say you're specifically building an .io product), you can change every weight yourself.

Yes, on Settings → Scoring Rules. Every rule shows its current weight and threshold; changing it re-ranks your own searches, Favorites, and My Domains — the site defaults stay the same for everyone else. Customizing rules is a Premium feature; free accounts can still see exactly how the whole system works on that same page, read-only.

Names built from familiar English letter patterns are easier to say, spell, and remember out loud — which matters when someone's telling a friend your domain name at a party. Q, x, and z are the letters English speakers stumble on most. If you're deliberately going for an unusual, alien-sounding brand (think Xerox, Zynga), you can turn this penalty off entirely in Settings → Scoring Rules.

We measure it, not guess it. A name reads easily when vowels and consonants alternate (ba-ku, not bk-pf), when it doesn't run three or more consonants in a row, and when it doesn't have doubled letters. Those signals combine into a 0–100 pronounceability score, which is what the Pronounceability Bonus and Hard-to-Say Penalty rules read from.

Not automatically — it's better by the rules currently active on your account. Score is a ranking tool, not a verdict. That's exactly why every rule is visible and every weight is editable — the goal is a score that matches your judgment, not ours.

Word scoring is under active development, including true "exclude, don't just penalize" filtering, so a bad name can be hidden outright rather than just sunk in the rankings. Have an idea for a rule we're missing? Reach out — that's exactly the kind of feedback that shapes what gets built next.