Backlink Analysis Tools: How to Tell Real Authority From Inflated Metrics

Link Building · Buyer’s Guide

Domain Rating 74. Domain Authority 68. Trust Flow 41. None of these numbers come from Google, all of them can be manufactured, and an entire industry depends on you not knowing that.

Updated August 2026 · 15 min read · Prices verified against official pricing pages

The first time somebody sold me a link, they sent a screenshot. Domain Rating 71, a healthy-looking backlink graph, a niche that vaguely matched mine, and a price of $180. I paid it. The link went live, and precisely nothing happened — no ranking movement, no referral traffic, nothing. Six months later the site was gone entirely, which at least explained the outcome retroactively.

What I’d been shown was a metric. What I’d assumed I was buying was authority. The gap between those two things is where most link-building money disappears, and closing it is mostly a matter of knowing how the numbers are made.

Here’s the thing that should reframe everything below: Google does not publish a link authority metric. PageRank scores stopped being visible in 2013 and the toolbar was retired in 2016. Every number you have seen since — DR, DA, Authority Score, Trust Flow, Domain Trust — is a vendor’s model, built from that vendor’s own crawl of the web, calibrated to correlate with rankings. They are educated guesses sold as measurements. Useful ones, often. But guesses.

What Each Metric Actually Measures And what it deliberately ignores

Domain Rating (Ahrefs)

A 0–100 logarithmic score derived almost purely from the quantity and DR of unique domains linking to a site, with the linking site’s outbound link count factored in. It is, deliberately, a link-graph metric and nothing else. It does not know about traffic, content quality, spam, or relevance. This makes it clean and predictable — and easy to game, because if the input is “unique referring domains, weighted,” then anyone who can acquire referring domains can move the number.

The logarithmic scale matters and is constantly misread. Going from DR 20 to DR 30 is a modest amount of work. Going from DR 70 to DR 80 is an enormous amount. When someone tells you their site went from DR 30 to DR 50 in three months, that is either a remarkable PR campaign or, far more likely, a purchase.

Domain Authority (Moz)

The original, and the one link sellers still quote most often, mostly through inertia. DA is a machine-learning prediction of ranking ability, trained against actual search results. Because it’s a trained model rather than a formula, it’s rebalanced periodically, and scores can shift across an entire index without any site doing anything. Moz’s index is meaningfully smaller than Ahrefs’ or Majestic’s, which means DA is likelier to miss links, which means it’s likelier to understate genuinely good sites and to be fooled by cheap ones.

Authority Score (Semrush)

A composite — link signals, estimated organic traffic, and a spam factor blended together. That blend makes it harder to inflate with links alone, which is a real advantage, and simultaneously harder to interpret, because a low score could mean weak links or thin traffic and the number won’t tell you which.

Trust Flow and Citation Flow (Majestic)

The most interesting pair, and the most underused. Majestic splits the concept deliberately in two. Citation Flow measures how much link volume points at a site. Trust Flow measures proximity to a manually curated set of trusted seed sites — universities, governments, major news organisations. The ratio between them is the signal.

A site with Citation Flow 55 and Trust Flow 12 has a great many links from places nobody trusts. That’s a link farm profile, and it’s visible in ten seconds. A healthy site tends to run those two numbers within reasonable distance of each other. Majestic also offers Topical Trust Flow, which scores authority within about 800 subject categories — so you can distinguish a site trusted in Health from one trusted in Gambling, which no single-number metric can do.

“Citation Flow 55, Trust Flow 12. In one ratio, you have learned more about a domain than any single 0–100 score will ever tell you.”

How the Numbers Get Faked The methods, so you can spot them

You don’t need to know how to do any of this. You need to recognise it when it’s being sold to you.

Expired domain resurrection. Someone buys a domain that once belonged to a genuine organisation — a defunct local newspaper, a closed nonprofit, a university department page — and rebuilds a generic blog on it. All the historical links remain. DR stays high for a year or two while the actual entity that earned those links no longer exists. This is the most common single trick in the link-selling market, and the tell is a site whose content bears no relationship to the topics of its backlink profile.

Redirect stacking. Buy several expired domains, 301 them all into one site, and the target inherits a link profile it never earned. Ahrefs shows redirected links in Site Explorer if you know to look; most buyers don’t look.

Sitewide footer and sidebar links. One deal produces links from ten thousand pages. Metrics that count pages rather than domains inflate wildly. Every serious tool counts referring domains for this reason, but the raw backlink count is still what appears in sales screenshots.

Reciprocal networks. Fifty sites in a group agree to link to each other. Each accumulates forty-nine referring domains from “real” sites. DR rises across the whole cluster. Google has been algorithmically identifying this pattern for a very long time.

Metric-farming links. Some sellers explicitly buy cheap links to their own domain purely to raise the score they then sell against. The site exists to have a number.

What to Look At Instead Six signals worth more than any score

1. Organic traffic to the linking page

Not the domain. The specific page your link would sit on. A link from a page that gets 400 organic visits a month is worth more than one buried on a page nobody has visited since 2021, regardless of domain-level metrics. This single check filters out most bad deals.

2. The referring domain growth curve

Every backlink tool plots referring domains over time. Look at the shape. Steady, gradual accumulation is what happens to sites that people cite. A flat line followed by a vertical cliff is what happens when someone buys a package. That chart is the single most honest artefact in any backlink tool, and it takes three seconds to read.

3. Anchor text distribution

A natural profile is dominated by brand names, bare URLs, and unhelpful phrases like “click here” and “this article.” When 30% of anchors are exact commercial keywords, somebody bought them. Check the profile of any site offering to link to you — if their own anchors look purchased, they’re a link seller regardless of what they call themselves.

4. Topical relevance

A link from a mid-tier site in your exact subject beats a link from a high-authority site with no connection to it. Majestic’s Topical Trust Flow makes this measurable rather than intuitive, which is a good reason to keep a cheap Majestic subscription alongside whatever else you use.

5. Placement and context

Editorially placed in the body of a relevant paragraph is the gold standard. Author bio, footer, sidebar, and “sponsored resources” blocks descend from there. Majestic’s Link Context feature shows you the surrounding paragraph, which tells you immediately whether the link is part of an argument or part of a list.

6. Is the linking page even indexed?

Astonishingly often, no. Search the exact URL in Google. If it doesn’t appear, Google either hasn’t crawled it or has chosen not to index it, and a link on an unindexed page transfers nothing. I’ve seen paid link campaigns where a third of the delivered placements failed this test.

The Tools, Priced

A note on index size, because it’s the argument vendors have most publicly. Ahrefs claims the largest live index of currently active links and refreshes fastest. Majestic maintains a Historic Index running to trillions of URLs, which is why it wins for archaeology — tracing what a domain’s profile looked like five years ago, before someone rebuilt it. These are different products solving different problems, and the fact that they disagree about any given domain is not evidence that one is wrong.

The stack I’d actually recommend for someone doing serious link work on a limited budget: Ahrefs Starter at $29 for the live view and the growth curves, plus Majestic Lite at $49.99 for Trust Flow ratios and topical scoring. Roughly $79 a month, and between them they’ll catch every fraud pattern described above. That’s less than a single mid-tier purchased link, which should tell you something about relative value.

The Disavow Question Short answer: almost certainly not

Every backlink tool has a “toxic links” report, every one of them flags dozens of domains, and every one of them offers to generate a disavow file. Resist.

The disavow tool was built in 2012, in response to Penguin, when bad links could actively demote a site. Since Penguin 4.0 in 2016, Google’s approach shifted from demoting sites for spam links to simply ignoring them. SpamBrain, Google’s machine-learning spam system, now neutralises manipulative link patterns algorithmically at scale. Google’s own documentation is explicit that most sites will never need the tool.

Two data points that ought to settle the argument. Bing removed its equivalent disavow tool in September 2023, with Microsoft stating plainly that its spam detection had matured to the point where user input was no longer needed. And Google’s Gary Illyes has said publicly that he would remove Google’s version if it were his call, because it hurts more than it helps most sites that use it.

John Mueller has been consistent on this for years, including through 2026. Asked about a site receiving a steady stream of spam redirect links, his position remains that ignoring it is fine, and that it is in a spammer’s interest to encourage site owners to waste time on defensive work. Which is a genuinely elegant point: if disavowing is largely futile, then a spam attack that provokes forty hours of disavow work has succeeded even though the links did nothing.

The arithmetic against pre-emptive disavowing is straightforward. Links that are genuinely worthless are already discounted, so disavowing them changes nothing. Links that carry small positive signals get those signals stripped. The expected value is negative. I’ve seen a site disavow a list of tool-flagged “toxic” domains, lose rankings three weeks later, remove the file, and recover.

Two situations still justify a disavow file: a manual action in Search Console explicitly citing unnatural inbound links, and a documented negative-SEO campaign you can’t resolve at the source and which correlates with an actual sustained ranking drop. Outside those, close the tab. And check Search Console’s own link report before any third-party toxicity score — Google’s list is the only one that reflects what Google is counting.

Reading Your Own Profile Honestly

Most of this article has been about vetting other people’s domains. Turn the same lens on your own once a quarter and it takes twenty minutes.

Pull your referring domain chart and look for spikes you can’t account for. If a cluster of links appeared in a week and you didn’t do anything that week, find out what it was — occasionally it’s a scraper network, occasionally it’s a piece of content that quietly got picked up, and knowing which is worth the effort. Then check your anchor text distribution against the natural pattern: brand terms and bare URLs dominant, commercial phrases a small minority. If your own profile is anchor-heavy on money keywords, that’s a legacy of whatever you or a previous agency did, and the remedy is diluting it with brand-anchored links rather than removing anything.

Finally, list the ten pages on your site with the most referring domains, then check whether those are the pages you actually want to rank. On a surprising number of sites they aren’t — the links have accumulated on an old resource post while the commercially important pages have almost none. Internal linking from the link-rich pages to the revenue pages is the cheapest link building available, costs nothing, and takes an afternoon.

A Vetting Rubric You Can Actually Use

When someone offers you a link, or when you’re deciding where to pitch, run this. It takes about two minutes per domain once you’ve done it a few times.

Automatic rejection if: the site has high domain metrics and negligible organic traffic; the referring domain chart shows a vertical spike; the content is generic multi-topic filler with no clear owner; the anchor profile is heavy on exact-match commercial phrases; there’s a “write for us” page advertising prices; or the linking page isn’t indexed.

Proceed with interest if: the site ranks for terms in its own niche; the traffic curve is gradual; there’s a named author with a traceable professional identity; the topical trust flow matches your subject; and the page your link would sit on already receives search traffic of its own.

Ignore entirely: the DR, DA or AS figure in the seller’s screenshot. It’s the one piece of information in the pitch that was chosen by the person trying to sell you something.

I still use Ahrefs daily and I’d renew Majestic without hesitating. Tools aren’t the problem. Treating their output as a verdict rather than as evidence is. The number is a starting point for a five-minute investigation, and the investigation is where the actual judgment lives — which is, unfortunately for everyone selling shortcuts, the part that can’t be bought.

Pricing verified August 2026 against Ahrefs, Semrush and Majestic’s official pricing pages; Google’s disavow guidance quoted from current Search Console documentation. This article contains no affiliate links — every link points to the vendor’s own site or a free Google property.