People also ask
In practice
The signals work as a sequence, not a scorecard. DR first, because it is cheap: a threshold (40 in this programme) removes the weakest tier in seconds — a filter, never the verdict, because it measures one input and is gameable. Traffic second, because audiences pay the bills: the tool’s estimate verified against the site — trend, sources, top pages, geography — separating real publications from metric shells.
Relevance third, because it is the value: a publisher your buyers read, a page about your category, a paragraph about your product — read at all three levels, because a strong site’s generic “resources” page is not relevance. Content standards last, because they decide durability: is the site’s writing original, are articles maintained, are commercial links labelled, would the article carrying your link still exist in a year — the judgement calls that no metric automates.
The order is the protection: each layer is cheaper than the failure it prevents, and skipping ahead — buying on DR, or worse, on price — is how budgets end up on sites that were never publications. The same four signals reverse into the report: every placement carries its evidence (floor passed, traffic verified, relevance confirmed, content seen), which is what makes a quality claim auditable rather than adjectival. And the AI layer reads them too: engines citing comparison pages inherit the page’s authority, audience and standards — the quality signals are now citation signals.
See also: DR 40+ as an entry filter, Site-level vs page-level qualification, Publisher vetting checklist.
Related service: Guest posts.
