Referring domains vs backlinks: which number matters

A backlink is one link to your site. A referring domain is one website that links to you, however many times. Referring domains are the better number to track, because the first link from a new site says more than the fiftieth from the same one.

In short

  • A backlink is a single link from another site, and a referring domain is a unique website with at least one backlink to you, so 100 links from one site count as one referring domain.
  • Ahrefs' January 2025 study of one million keywords found referring domains correlate with rankings at 0.255, against 0.248 for raw backlinks.
  • The same study put Domain Rating at 0.131, a weaker correlation than a plain count of referring domains.
  • A single sitewide footer link can add thousands of backlinks and only one referring domain, which is why backlink totals mislead.
  • Every tool crawls its own copy of the web, so referring domain counts differ between Ahrefs, Semrush, Majestic and Google Search Console.

A backlink is one link from another website to yours. A referring domain is one website that links to you, no matter how many times. If a blog links to you from 100 different pages, you have 100 backlinks and one referring domain. The number that matters more is referring domains. Each new site that links is a separate decision by a separate publisher, and it is that decision a search engine is trying to count. The hundredth link from the same site adds little to what the first one already said. Most of the confusion beginners have with link reports comes from reading the larger of these two numbers as the important one, when it is usually the smaller that deserves the attention.

The two definitions

A “domain” here means a website as a whole, the name you would type to reach it, as opposed to any single page on it. Tools report links at several levels, and the same blog linking to you looks very different depending on which level you read.

Term What it counts Also called Example
Backlink Every individual link pointing to your site Inbound link, inlink One blog links to you in 100 posts: 100 backlinks
Referring domain Each unique website with at least one backlink to you Linking root domain The same blog: 1 referring domain
Referring page Each unique page that links to you Linking page The same blog: 100 referring pages
Referring IP or subnet Each unique server address or address block that hosts linking sites C-class Used by tools to spot networks of sites hosted together

A single backlink lives on a page, and a page lives on a domain. Tools report all three levels. The domain level is the one most link builders plan and report by, because it comes closest to counting independent endorsements. The fourth row works at a different level again. An IP address is the numerical address of the server a site runs on, and a subnet is a block of neighbouring addresses, so this count asks how many separate hosting locations your links come from. It becomes useful later in this lesson, when the question is whether a group of domains is really a group of independent publishers.

Why referring domains say more

The data points this way. Ahrefs studied one million US keywords and published the results in January 2025. The number of referring domains correlated with rankings at 0.255. The raw number of backlinks came in at 0.248. A correlation measures how closely two things move together, and these two figures say that pages with more referring domains tend to rank a little higher, and that the relationship is very slightly tighter for domains than for links. The gap is small, but it points the same way as the theory.

The theory is that link analysis treats a link as an endorsement. Fifty endorsements from the same person are worth less than endorsements from fifty people, for the plain reason that the first group tells you about one person’s opinion and the second tells you about fifty. The documentation that leaked from Google in May 2024 includes a field that records several links from the same source to the same target, according to analyses of the leak. Analysts read that as support for diminishing returns from repeated links on one domain. The field is documented. Its use and weight are not confirmed, so this is supporting evidence and should be described that way.

Practice gives a third reason. Backlink counts are easy to inflate and referring domains are harder. One site adding you to its footer can add thousands of backlinks overnight, and nothing about your reputation has changed. A number that can jump by thousands because one person edited one template is a poor measure of progress. One caution applies to all of this, which is that correlation is not cause. Pages that rank well are seen by more people and attract more links, so some of the relationship runs the other way, from ranking to links.

When backlinks far outnumber referring domains, one of four things is usually the reason. The most common is a sitewide link. A link in a footer, sidebar or template appears on every page of the site, so a site with several thousand pages produces several thousand backlinks from one editing decision. Sitewide and footer links explains when that is a problem. Pagination and archives are the second cause. One article can appear on tag pages, category pages and date archives, and each copy counts as a backlink in a tool, even though the publisher linked to you once. Syndication is the third. A press release or article copied across a network creates many links with little independent value, since every copy repeats the same decision. The fourth is automated spam. Scrapers and comment bots can create large numbers of links that Google is designed to ignore.

A ratio of backlinks to referring domains that jumps suddenly is worth a look. Open the tool’s report, sort by links per domain and see which sites are responsible. In most cases the explanation is harmless and takes a minute to find, such as a partner who has added your logo to a template. What you are checking is whether the jump reflects something you would want to know about, such as a sitewide link with a keyword anchor that you did not ask for.

Diminishing returns do not mean zero, and three cases are worth having. The first is a link to a different page. A second link from the same site to a different page of yours brings value and anchor text to a page that had none from that source. The second is a link that sends readers. A publication that covers you often keeps sending visitors and keeps your name in front of its audience, which is worth having whatever it does for rankings. The third is a fresh editorial mention. Repeated coverage in the same outlet is what a real relationship with a journalist looks like, and it is the normal result of good press work. What adds almost nothing is the same link repeated across every page of one site.

Not all referring domains are equal

Counting domains is better than counting links, and it is still only counting. One link from a publication your customers read can outweigh fifty from sites nobody visits. Three refinements make the number more honest, and the first is to look at followed referring domains. Tools can filter out domains whose links all carry nofollow, sponsored or ugc, which are labels a site can add to a link to tell search engines it does not vouch for it, was paid for it, or did not write it. In the Ahrefs study, followed referring domains correlated at 0.250 and all referring domains at 0.255. The two are too close to separate, so report both. Dropping labelled domains from a report would throw away links that may well count, and hiding the split would overstate what is certain.

The second refinement is the quality of the domain. The same study found Domain Rating correlated with rankings at only 0.131, weaker than a plain count of referring domains. Domain Rating is Ahrefs’ score for the strength of a site’s link profile, and the finding is a useful warning for anyone tempted to buy links by score. A score on its own is a poor filter. Domain Authority vs Domain Rating explains what those scores measure. Real organic traffic and topical fit say more, and how to vet a publisher gives the checks.

The third refinement is independence. Fifty domains owned by one operator and hosted together are closer to one source than fifty. The whole case for counting domains rests on each one being a separate decision by a separate publisher, and that case collapses when one person controls them all. That is why tools report referring IPs and subnets, and it is one of the footprints that exposes a private blog network, which is a set of sites built or bought by one operator in order to link to chosen targets. A report showing many referring domains on very few subnets is a reason to look closer before counting them as separate endorsements.

Why every tool shows a different number

Google does not publish a full list of your links. The Links report in Search Console shows a sample. Ahrefs, Semrush, Majestic and Moz each run their own crawler and keep their own index, so each sees a different slice of the web. A crawler is software that visits pages and records the links on them, and no crawler reaches everything, which is why one tool can report a link that another has never seen. Tools also differ in how they count: whether subdomains are separate domains, how long a lost link stays in the total, and whether redirects are included. A subdomain is a section of a site with its own prefix in the address, and a tool that counts each one separately will report more referring domains than a tool that groups them under the main site.

The sensible response is to pick one tool and compare like with like over time. The absolute figure in any tool is an estimate, but the direction of travel within one tool is reliable enough to manage by. Our desk review of Ahrefs and the wider backlink tools comparison cover the differences. Counts also fall. An Ahrefs study dated February 2024 found that about two thirds of links to sites over the previous nine years were dead. A referring domain total should be read as live domains now, not as every domain that ever linked, and a site that stops earning links will see the number drift downwards as pages are deleted and sites close.

Which number to use for what

Different jobs call for different counts, so it helps to decide what you are trying to find out before opening a report. For reporting link building progress, use new referring domains per month, with followed domains shown separately. For sizing the gap to a competitor, use referring domains to the page that ranks, compared with yours. That comparison is made page against page because the page is what competes in a search, and a competitor with a huge site may have only a handful of domains linking to the page you are up against. It is also the comparison that sets budgets. How many backlinks you need shows how to turn it into a target.

The raw backlink count still has uses, mostly diagnostic ones. For checking for a spam pattern, look at backlinks per referring domain, and the anchors on the largest sources. For checking for a network, look at referring IPs and subnets. For finding lost value, look at lost referring domains, and backlinks pointing to broken pages, since both show links you once had and could recover. Read how to track backlinks to set up monitoring. Definitions of referring domain, linking root domain and related terms are in the glossary.

Common questions

What is a referring domain?

It is a website that links to yours at least once. Tools also call it a linking root domain. If a site links to you from 300 pages, that is 300 backlinks and one referring domain.

Which is more important, backlinks or referring domains?

Referring domains. In Ahrefs' January 2025 study they correlated with rankings slightly better than backlink counts, and the model behind link analysis gives less to each extra link from the same site.

Why do I have far more backlinks than referring domains?

Usually because one or more sites link to you from every page, through a footer, sidebar or template. It can also come from one site citing you in many articles.

Is it bad to get many links from the same domain?

Not in itself. Repeated editorial links from a site that covers you often are normal. A sitewide link with a keyword anchor is different, because Google's spam policies name widely distributed footer and template links.

How many referring domains do I need?

There is no fixed number. Compare the page you want to rank with the pages that already rank for that search, and count the referring domains to each.

Vendors to look at

  • BazoomEditor's pick

    Sponsored content and link marketplace, managed service

  • MotherlinkEditor's pick

    Backlink services, guest posts, niche edits, full SEO