Referring domains: the count that matters more than the rating
A domain with ten thousand backlinks from twelve sites has twelve votes. Referring domains, the count of distinct sites that link, is the number that tracks ranking power; backlink totals are inflated by sitewide footers and by design.
What is a referring domain?
One distinct registrable domain that links to the target at least once, however many times. A sitewide footer link from one site is one referring domain and perhaps ten thousand backlinks. Ranking signals behave much more like the first count than the second, which is why every gap analysis and every target profile here is built on referring domains.
How do we use it on inventory?
Independently measured referring domains sit beside each seller’s claimed figure as referring_domains. It is one of the fields the subdomain test uses (a subdomain needs its own referrers to pass) and one the gap map is made of. For a target page, the profile call returns the referring domains, their own ranks in bands, first-seen dates, nofollow share and which of them are buyable in the catalogue.
Does the quality of the referrers matter?
More than the count. The profile reports referrers by their own rank band, so a page with 300 referring domains that are all rank-0 sites reads differently from one with 60 in the top bands. The same logic runs over the whole web graph for publishers: how many outside domains link in, against how many the publisher links out. A publisher linking out to five times as many domains as link in is selling more than it earns; that is the fifth quality test.
What is a reasonable referring-domain target?
The count the ranking pages have, discounted for what is not buyable. Run the gap map and read gap_total against carried; the second number is the plan.
Profile the page first, then plan:
“Profile https://yoursite.com/page: referring domains by rank band, anchor mix, and how many of its competitors' referrers are buyable.”
Methods: research methodology.