What metrics do you use to verify domain quality?
Every domain in the directory is checked against the same set of signals before it's eligible to list. Nothing is estimated from a registration record, and nothing is taken on trust from a previous owner. This page covers each metric you'll see on a listing, what it means, and where its limits are.
Revised Score
The headline number: a 0–100 grade that lets you compare two listings without weighing five separate metrics against each other. It's a percentile blend of link-graph centrality, referring domains, and archive history, computed across the whole directory, so a 70 means the same thing on every listing rather than just relative to the domains beside it.
Scores are graded into five bands, which is what the color of the circle on a listing tells you:
- Emerging: under 30
- Fair: 30 to 49
- Good: 50 to 64
- Strong: 65 to 79
- Elite: 80 and up
There's deliberately no red band. A domain that would grade badly doesn't get listed at all, so the bottom of the scale is "modest history", not "bad history".
Referring domains
The count of distinct domains linking to it. That's a better signal than raw link volume, since a thousand links from one site is one relationship rather than a thousand. Listings show this as a band: under 10, 10–50, 50–100, 100–250, or 250 and above.
Read it alongside the score rather than on its own. A domain in the 50–100 band whose links come from well-cited sources will usually outrank one in the 250+ band whose links are all from the same low-quality neighborhood.
Notable linking sources
Where a count can't tell you much, a name can. When a domain has links from sources with genuine editorial weight (Wikipedia, Hacker News, university departments, Reddit, GitHub, news organizations, government and standards bodies), those are named individually on the listing.
This is usually the most informative thing on the page. You can also browse the directory by linking source if a specific kind of citation is what you're after.
Archive history
A registration date proves someone paid for a name. It doesn't prove a site existed. We use the Wayback Machine's snapshot record instead: how many times the site was captured, and the window it was actively being crawled in, shown as a snapshot count band and an active-window range like 2005–2009.
Thin archive coverage isn't automatically disqualifying, but it does mean less evidence, and the score reflects that.
Spam screening
Every candidate is scored for spam risk before listing, and anything in the high band is dropped outright rather than listed with a warning. What survives is labeled either low risk or moderate. Moderate is worth a closer look rather than an automatic no, and it appears on the listing precisely so you can make that call yourself.
Category and keyword themes
Where we can establish what a site was actually about, the listing carries a category and a set of keyword themes. This isn't available for every domain, since a site with a thin or ambiguous archive record often can't be classified honestly. Treat it as a bonus signal when it's there rather than something to filter the whole directory on.
Why ranges instead of exact numbers
Revised Score appears as a band (say 65–75) rather than a single integer, and referring domains and snapshots appear as buckets. Two reasons:
The first is fairness. Exact figures across thousands of listings would let anyone reconstruct the underlying dataset and re-identify specific domains without revealing anything, which would give away the one thing every listing holds back.
The second is honesty about precision. These metrics come from third-party crawls that shift week to week. A band is an accurate description of what we know; a single decimal-precise number would imply a confidence the underlying data doesn't support.
The exact figures are the basis for the band you see, and the band never flatters the domain. It's centered on the real value, not stretched upward from it.
What gets a domain excluded
Not everything we look at gets listed. A candidate is dropped when it has:
- Spam history: a high spam score, or any archived snapshot flagged as spam
- Parked-only history: every archived snapshot is a parking page, which is no history worth claiming
- Trademark risk: the name contains a major brand, so whoever registered it would inherit a dispute rather than an asset
- No story at all: no measurable link or authority signal to sell on in the first place
These are hard exclusions applied before anything reaches the directory, which is why you won't see a listing that fails them at all.
Checking the numbers yourself
You don't have to take ours on faith. The free tools run the same public checks we do: the backlink checker for referring domains, and the domain report for a full profile of any domain you already own or are considering elsewhere.
For how these roll into one number, see what Revised Score is; for what each plan opens up, see what you get on each plan.