A domain is four assets wearing one name.
Most appraisals hand back a single score, which is the one thing a domain cannot honestly give you. A name can be excellent while its history is worthless. A history can be worth paying for while the name is unsellable. Averaging those into one number destroys the only information you were buying.
Apex Ranks reads four measures and keeps them apart. This page is the method, written out.
Read against itself
The clearest way to show what the four readings do is to point them at this domain. Nothing below is illustrative — the numbers come from running the appraisal on apexranks.com, which was registered days ago and has no past to speak of.
Nine letters, no hyphen, no digits, and a vowel run close enough to natural English that it survives being read aloud over a phone.
The web archive holds no captures. A domain with no past floors at fifteen rather than zero — never having existed is not the same as having failed.
Nothing points here. No referring domains, so nothing to inherit and nothing to clean up.
A debit, not a credit. With no links to sample there is nothing to vouch for the domain, so risk starts at thirty-five. Absence of evidence is scored as uncertainty, not innocence.
Four readings, and they disagree by seventy-one points. That disagreement is the appraisal. Collapsed into a single figure it would read somewhere around forty and tell you nothing you could act on.
The name
Judged as a word, in isolation, before anything is known about what the domain has done. Length carries most of it, because short names are scarce and long ones are typed wrong.
- +14four to eight letters
- +9nine or ten
- +2eleven or twelve
- +5three or fewer, where scarcity outweighs meaning
- −3 / ltrevery letter past twelve, down to a floor of −28
- −22a hyphen, which is lost in every spoken handoff
- −18any digit, for the same reason
- −8the same letter three times running
Whether it survives being said out loud
A separate reading, weighted at just under a third of the name score. Vowels are counted against a target ratio of 0.40 — English words cluster there, and names that stray far in either direction get spelled back to you wrong. No vowel at all costs 45. Five consonants in a row costs 30.
The zone is never folded in
Extension quality is kept as its own figure, because a strong word on a weak extension is a different asset from a weak word on a strong one, and the average hides which you are holding.
- com 100
- ai 87
- org 86
- net 83
- io 82
- app 82
- de 82
- co.uk 82
- com.ua 82
- pl 80
- fr 80
- ua 79
- co 78
- dev 76
- me 74
- travel 70
- shop 65
- store 64
Anything unlisted reads 60 if the extension is three letters or fewer, 52 if it is longer.
The record
What the web archive remembers. Three things matter, and they are not the same thing: how much was captured, how many distinct years show activity, and how far apart the first and last capture sit.
- up to 38volume of captures, on a log scale — the difference between ten and a hundred matters far more than between ten thousand and a hundred thousand
- up to 37years with activity, at 3.7 each
- up to 25span between first and last capture, at 1.4 a year
- 15nothing archived at all
Counting years separately from captures is deliberate. A site crawled heavily across one year and abandoned is a worse inheritance than one crawled lightly across nine, and volume alone cannot tell them apart.
The links
The reading most often inflated, and the one worth least at face value. Reported totals are treated as a claim to be checked, not as a result. Where the budget allows, a sample of the live links is fetched and read, and what comes back moves the score more than the totals do.
- up to 38referring domains, log scaled
- up to 28the provider's own rank for the target
- up to 16share of sampled links still resolving
- up to 23share of those that are editorial rather than filler
- −28reported spam score
What a fetched link is worth
Each sampled link is classified from its URL, anchor and page title. The weights are blunt on purpose — the job is to separate a placement someone made a decision about from one that accumulated.
- 95university and government hosts
- 65blogging and newsletter platforms
- 60everything else editorial
- 40tag, author, profile, forum and comment URLs — real pages, but nobody chose to place them
- 5gambling, pharma, private networks, paid guest posts
What the reading is for
The four numbers exist to answer one question — what kind of asset is this, if any. Six answers, and they are not a ranking. A donor and a brand are bought by different people for different money.
- Premium brandthe name alone carries it, history irrelevant
- Buildgood name, nothing to inherit — pay for the word, not the past
- SEO donorhistory and links worth having, name not worth defending
- Rebuildsomething real was here and could be again
- Hold or flipworth owning, not worth developing
- Passneither half survives inspection
Each appraisal ends on buy, watch or pass, with a ceiling bid attached. The bid is a decision aid and deliberately conservative. It is not a market appraisal and should never be quoted as one.