Backlink Profile Analysis: What to Measure Beyond Domain Rating

Run any domain through a backlink checker and the first thing it returns is a single authority score — Domain Rating on Ahrefs, Domain Authority on Moz, or a marketplace's own version of the same idea. That number is a reasonable five-second filter and a poor stand-in for anything past it. A real backlink profile analysis has to look at what's underneath the score: how the links are distributed, not just how many exist or how strong the linking domains are said to be. Two profiles can carry an identical authority number and represent completely different amounts of risk and completely different amounts of real value, because the score compresses several independent variables into one digit and discards the variance between them. This piece works through what to measure once you get past that first number — the same quality-scoring work the broader backlink audit framework on this site treats as one stage among four (inventory, risk classification, quality scoring, action plan), expanded here into the specific dimensions worth checking one at a time.

Why a Single Authority Score Isn't a Backlink Profile Analysis

Domain Rating and Domain Authority aren't two ways of measuring the same underlying fact about a link profile. They're two vendors' proprietary attempts to compress backlink data into one predictive number, built on two different indexes with two different models. Ahrefs calculates DR from the number of unique referring domains pointing at a site, the authority of those domains, and how many other sites each of them links out to, then plots the result on a logarithmic 0-100 scale, per Ahrefs' own explanation of the metric. Moz's Domain Authority runs a comparable idea through a machine-learning model trained against real search results, also scored 1-100, built specifically to predict ranking ability rather than measure it directly. Moz itself is on record saying DA isn't a ranking factor search engines actually use — a point explained in detail alongside Moz's full calculation methodology here.

Here's the part that should give anyone pause about treating either score as sufficient on its own: Ahrefs' own research team has found that the number of referring domains pointing at a page correlates more closely with that page's organic traffic than the page's Domain Rating does, according to Ahrefs' guidance on building website authority. The input the score is partly built from predicts outcomes better than the score itself.

Backlink marketplaces built their own scoring systems for the same reason review sites build star ratings: buyers want a fast filter before they dig in. bklink's marketplace surfaces this as a Rank rather than a DR or DA, precisely because it isn't derived from either vendor's index and isn't meant to be read as interchangeable with either one. But the naming distinction only matters if the number is also used the way any of these scores should be used — as a triage filter, not a verdict. A backlink profile checker that stops at that single figure is running maybe a sixth of a real analysis.

Consider two sites that both show a DR of 55. Site A earned that score from 400 referring domains spread across trade publications, review sites, and a mix of guest posts. Site B reached the same number from 60 referring domains, roughly half of them link-network-style blogs that also link out to a dozen unrelated affiliate sites apiece. Ahrefs' formula would compute a similar DR for both, because it weighs domain count and linking-domain authority — not what else those domains are doing, or whether the profile as a whole makes editorial sense. Only a wider analysis catches the difference, and it's exactly the kind of difference that matters to a buyer deciding whether either profile is actually worth what its score implies.

Referring-Domain Diversity: The Concentration Risk a High Score Hides

Backlinks and referring domains are not the same measurement, and conflating them is probably the single most common mistake in a link profile analysis. A hundred backlinks from one domain and a hundred backlinks from a hundred domains produce very different signals, even though most checkers report both numbers side by side without flagging the ratio between them. By definition, a referring domain counts once no matter how many separate links or pages on it point at you — a distinction laid out plainly in Ahrefs' own metrics glossary — which means a headline "referring domains" number can't be inflated by one enthusiastic linking site the way a raw backlink count can.

The reason concentration matters is diminishing returns: a second link from a domain that already links to you tells Google the same publisher likes you twice, not that a new independent source has discovered you. SE Ranking's explainer on referring domains puts it plainly: when the same domain links to you again, "the value of that second show of support pales in comparison to that of a new domain 'recommending' your website." Push that logic to its extreme and a profile with 500 backlinks from 12 domains behaves, for ranking purposes, a lot like a profile with a much smaller number of distinct endorsements — most of the volume is repetition, not new evidence.

In practice, a useful first check is the ratio of total backlinks to unique referring domains. A ratio sitting in the low single digits — most referring domains contributing just one or a few links each — tends to reflect organic accumulation. A ratio climbing into double digits, or a profile where a small handful of domains account for a disproportionate share of total links, is worth investigating before you assume the underlying authority score reflects a broad base of independent endorsement. That investigation might turn up something benign — a franchise structure with many subdomains, a widely embedded badge or tool, a syndication partner — or it might turn up a small paid link network dressed up as organic growth.

Concentration risk cuts both ways on value, too. A buyer evaluating a vendor's placement portfolio who only checks average DR across placements can be fooled by a portfolio that's actually a handful of publishers rehosting the same content across several owned domains apiece. The referring-domain count, not the backlink count, is the number that should drive that buying decision — which lines up with Ahrefs' own finding that referring domains are the stronger predictor of organic traffic gains in the first place.

Topical Relevance Distribution: The Profile-Wide View, Not Just Link by Link

Most link-by-link relevance checks ask a narrow question: does this specific page make sense as a source for this specific link? That's necessary, but it isn't the same question as: does the profile as a whole cluster around the site's actual subject matter, or has it drifted?

A SaaS company that spends two years doing outreach inside its own niche, then runs a broader digital-PR campaign that lands coverage in general business and lifestyle press, ends up with a profile whose topical center of gravity has shifted — even though every individual link might be perfectly legitimate. That shift isn't necessarily bad; broad coverage has real value. But an analysis that only checks links one at a time never surfaces the shift, because no single link looks wrong in isolation.

This is really the same argument Ahrefs makes about individual pages extended to the whole profile: a domain-wide authority score says nothing about whether a specific linking page is relevant to your topic, actively indexed, or providing useful context, which is exactly why judging a site on site-wide authority alone is a mistake. Scale that up, and a rising share of unrelated-vertical links — directories, generic "best of" listicles in unrelated categories, forum profiles with a bio-line link — dilutes topical relevance across the whole profile even when no individual link would fail a manual spam review.

The practical fix is to bucket referring domains by subject matter — even coarse buckets, like same-industry, adjacent-industry, and unrelated — and track the proportion over time the same way you'd track anchor text or link velocity. A profile where most referring domains sit in the same or an adjacent category to the business is doing real topical work. A profile trending toward a majority of unrelated-category domains is accumulating authority-shaped weight without the topical signal that made the early links valuable — and once enough of those unrelated links start sharing other red flags, the pattern overlaps with what our toxic backlinks framework treats as worth a closer risk-classification pass rather than something to wave off as folklore.

Anchor Text Distribution: Natural Mix vs. Over-Optimization

Anchor text is one of the clearest fingerprints separating a link profile that grew through editorial pickup from one that was assembled. That's been true since Google's 2012 Penguin update, which specifically targeted sites where a suspiciously high share of backlinks used the exact keyword phrase the site wanted to rank for — a pattern that's cheap to manufacture (specify the anchor text on a batch of purchased or traded links) and correspondingly easy to fingerprint once enough sites had done it.

What's counterintuitive is how weak the upside is even when the pattern isn't caught. Ahrefs studied 384,614 ranking pages across roughly 19,840 keywords and found only a weak relationship between the share of exact-match anchor text pointing at a page and where that page ranked — a Spearman correlation of roughly 0.14 on average, close enough to zero to call it noise. Pages don't reliably rank better for carrying more exact-match anchors; they mostly just take on more downside risk if the pattern looks engineered. Ahrefs' own conclusion from the data was that topical relevance in the linking content matters far more than the literal anchor phrase.

A profile-level anchor text analysis buckets every anchor into a handful of categories — branded (the company or product name), natural phrase (a sentence fragment that happens to include relevant words), generic ("click here," "this article," "read more"), naked URL, and exact or partial match of a target keyword — then tracks what share each bucket holds and how that share moves over time. A profile dominated by branded and natural-phrase anchors, with exact-match sitting as a small, stable minority, reads the way editorially earned links actually read: writers link using a brand name or a natural description because they're describing something, not thinking about someone else's target keyword. A profile where one exact-match phrase accounts for a large and growing share of total anchors is a pattern worth investigating regardless of what the aggregate authority score says — anchor concentration alone isn't proof of manipulation, but it's exactly the sort of profile-level signal a risk-classification pass should flag for a closer look, the same way concentrated referring domains or an unexplained traffic spike would.

The trend line matters as much as the snapshot here. A sudden jump in exact-match share over a single month is a stronger signal than the same exact-match percentage held flat for two years, because the jump usually maps to one batch of links landing at once — which is also where link velocity comes in.

Link Velocity: What a Growth Curve Shows That a Snapshot Can't

A backlink profile analysis run once gives you a photograph. Run monthly and compared over time, it gives you a trend line, and the trend line is where velocity-based risk and opportunity both show up — neither a healthy nor an unhealthy profile looks the same at a single point in time as it does in motion.

Search Engine Land's guide to link velocity draws the distinction the SEO industry has converged on since Penguin: natural growth tends to be gradual and comes from diverse, topically relevant domains, while unnatural growth shows up as a large batch of links appearing in a short window with no diversity behind it. Crucially, the guide is explicit that a spike by itself isn't a penalty trigger: "it's not the spike itself that's the problem; it's the quality and source of the links that matter." A product launch, a piece of research picked up by trade press, or a genuinely viral post can all produce a fast jump in referring domains without looking anything like a manipulated pattern, because the new links come from dozens of independent publishers rather than a handful of coordinated ones.

That context-dependence is exactly why velocity has to be read alongside the other dimensions rather than alone. Ahrefs has found that top-ranking pages tend to keep earning new followed referring domains at a steady clip — roughly 5% to 14.5% per month among top performers, based on an analysis across 10,000 search results pages — which sets a rough expectation for what ongoing, healthy acquisition looks like for a page that's actually performing well. A profile that sits flat for six months and then jumps 200% in a week doesn't match that pattern, and neither does a profile that spikes and then loses most of those same referring domains within a few more weeks — a saw-tooth shape more consistent with a rented link package than an earned one. Neither pattern is automatically toxic on its own, but both are exactly the kind of thing a velocity chart surfaces and a single authority snapshot cannot.

Follow vs. Nofollow Ratio: Why 100% Dofollow Is a Red Flag, Not a Goal

Google gives publishers three ways to qualify an outbound link — nofollow, sponsored, and ugc — and its own documentation is specific about when each applies: sponsored for paid placements and advertising, ugc for comments and forum posts, and nofollow as the general-purpose option for when a site doesn't want to vouch for or have a linked page crawled from it. None of these attributes are inherently negative signals about the page being linked to; they describe the linking site's relationship to the content, which is exactly why Google introduced the more specific sponsored and ugc values as alternatives to a blanket nofollow.

That matters for profile analysis because a real, organically earned backlink profile almost never ends up 100% dofollow. Press coverage often carries sponsored or nofollow tags on outbound links as a matter of editorial policy. Comment sections and forums apply ugc by default, often through the platform's own software rather than a deliberate editorial choice. Directories, social profiles, and plenty of large publishers nofollow outbound links across the board regardless of how they feel about the destination. A profile that's overwhelmingly or exclusively dofollow usually means the links were placed somewhere that let the linker choose the attribute — which includes plenty of legitimate placements, but is also exactly what a paid-link or guest-post network optimized purely for ranking value looks like, since nofollow links carry no direct ranking credit and a network selling links has little reason to include them.

The useful question isn't "what's the magic ratio," because no fixed number exists — it varies by industry and by how much of a site's exposure comes through channels like forums, comments, and social platforms that nofollow by default. The useful question is whether a meaningful nofollow, sponsored, or ugc share exists at all, and whether it moves in a way that's consistent with how the site is actually being mentioned.

Traffic-Weighted Authority: Does Anyone Actually See This Link?

A backlink can sit on a page with a respectable authority score and still be functionally invisible, because authority scores measure a page's position in the link graph, not whether the page gets crawled regularly or seen by an actual visitor. A page can accumulate DR or DA by being linked to from other pages in the same network without ever ranking for anything or drawing organic search traffic of its own — and from the link graph alone, that page can look identical to a genuinely popular one carrying a similar score.

This is the gap traffic-weighted analysis is meant to close. Ahrefs built organic-traffic estimates for both the linking page and the linking domain directly into its backlink checker for this reason, noting it as a feature most other tools didn't offer: visibility into estimated traffic for the specific page containing a link, not just an aggregate score for the domain it sits on. The practical use is straightforward. A link on a page that gets meaningful, ongoing organic traffic is more likely to sit inside content that real users and other publishers actually encounter and cite. A link on a page with a decent authority score but no measurable traffic is more consistent with a page that exists mainly to hold outbound links — a private-blog-network pattern, an abandoned domain kept alive for its link equity, or a page that was never meant to be read, only crawled.

None of this makes a low-traffic linking page worthless. A niche technical page can carry real topical authority on modest traffic, and a legitimate new publication may not have built search visibility yet. But traffic-weighting flips the default assumption: instead of trusting an authority score until something looks wrong, it asks every high-scoring link to show a second, harder-to-fake signal — real visitors — before counting fully toward the profile's value. Between two links with an identical authority score, the one sitting on a page that visibly gets used is doing more work than the one that exists only on paper.

The Six Dimensions at a Glance

None of the following six dimensions replaces Domain Rating, Domain Authority, or a marketplace Rank. They sit alongside the authority score and explain what it's hiding — which is the difference between a backlink profile checker's front-page number and an actual backlink profile analysis.

Where This Fits Into a Full Audit

These six dimensions aren't a replacement for a full backlink audit — they're what the quality-scoring stage of one should actually measure, instead of stopping at a single authority number. The backlink audit framework covered elsewhere on this site walks through the full process this analysis feeds into: build a complete inventory of links from more than one data source, classify each one by risk, score what's left for quality using dimensions like the ones above, then turn the scores into an action plan of links to keep, reclaim, request removal on, disavow, or build around. This piece is the backlink profile analysis you run at that scoring stage.

Run it on a quarterly basis for an actively growing site, and again before any material decision that leans on the strength of a backlink profile — evaluating an acquisition target, deciding whether to keep paying a link-building vendor, or negotiating placement pricing on a marketplace. A single authority score answers "is this roughly in the right range." These six dimensions answer the harder, more useful question: is that number backed by something real, and is it moving in a direction that will hold up.

Related reading

• Backlink Audit: A Practical Framework for Finding Risk, Quality, and Opportunity
• Toxic Backlinks and Link Quality: How to Detect, Prioritize, and Respond