How the numbers are measured, and what we leave out

When you scan a domain or open a directory page, you are reading measurements we already hold. No crawl happens on demand, which is why answers are immediate and why two people asking the same question on the same day see the same numbers. The crawler fills the store in the background at polite rates, and pages state when a value was last checked.

Directory traffic: we have no instrument for another site's visitor counts, so we do not publish any. Domain rating: DR is third-party data we do not verify; we hold no DR value for any directory today, so every DR field on the site reads "Unknown". If we ever record one it will carry the catalog it came from and the date we read it, it will stay display only, and it will never enter a rating or a default sort. Product categories are matched from the words a product uses about itself and are labeled as inference, because everything else on the site is something we read off a real page. Community votes and comments are opinion, kept in their own labeled surfaces. And we publish no ranked judgments of directories beyond what the measurements themselves say, because we operate in this market too.

An absent number is information; a guessed one would be a claim we cannot back, on a site whose product is backing its claims. So the site distinguishes the kinds of absence rather than printing one dash for all of them. A directory we have not crawled reads "not verified yet". A directory with no rating reads "not rated". An indexation rate under its publishing floor reads "not measured", beside a count of the checks already on record. A link whose rel we have not read reads "unchecked". A third-party field we never recorded reads "Unknown".