The Patent Files on High Quality Backlinks

Every backlink audit I run starts the same way. The client hands me a spreadsheet of their backlinks ranked by domain rating, the high numbers shaded green, the low ones red, and a question about which red links to disavow. I understand why people build that spreadsheet. The whole SEO industry trained them to. But domain rating is a third party score invented by a tool, and Google has never seen it. When I want to know what a backlink is actually worth to search rankings, I do not open the spreadsheet. I open the patents.

The advice about quality backlinks has hardened into the same five bullet points on every SEO blog. Authority, relevance, anchor text, dofollow, editorial placement. None of that is wrong. It is just shallow. Those five traits are the surface of a scoring system Google has been describing in patent filings for two decades, and the 2024 documentation leak proved the machinery underneath is more detailed than any checklist admits. So let me walk through what the link graph actually measures, because once you see it, you stop ranking links by a number that does not exist inside Google.

What high quality backlinks actually are

Quality backlinks are backlinks Google scores as high quality: relevant, authoritative, editorial links. A quality backlink earns its quality from relevance, authority, anchor text, and placement, not from a domain rating. Backlink quality is whether your backlinks are the relevant, authoritative backlinks the patents reward instead of the low quality backlinks Google ignores.

That definition lands close to the standard checklist, and that is the point. The checklist is a list of symptoms. The patents are the diagnosis. A backlink is a vote, and Google has spent twenty years building systems to weigh how much each vote counts. Google still names PageRank as one of its core ranking systems in its own documentation, so the link graph is not a museum piece. It is live, it is scored, and the score has very little to do with the green cells in a backlink tool.

Why a domain rating is not backlink quality

Domain rating and domain authority are reverse engineered guesses. Moz built one, Ahrefs built another, and both estimate a site's strength from their own crawl of the link graph. They are useful as a rough relative proxy. They are not what Google uses, and treating a domain rating as backlink quality is how good sites waste budget chasing high number links that pass almost no value. A high domain rating on the linking site tells you the site is strong overall. It tells you nothing about whether the specific page linking to you is relevant, well placed, or even indexed in a tier where the link counts. Two backlinks from the same high authority domain can carry wildly different value depending on which page they sit on and what surrounds them.

What Google's reasonable surfer patent says about backlink quality

The single most useful document for understanding backlink quality is the reasonable surfer patent. Google's user behavior ranking patent describes a model where every link on a page is assigned a probability that a real person clicks it, and the link passes value in proportion to that probability. Not every link on a page passes the same authority. A link buried in a footer beside forty other links is worth a fraction of a link sitting in the first paragraph of the body content.

This is the idea that should replace the domain rating reflex. A backlink is not a flat unit of value. It is a weighted signal, and the weight comes from features Google can read directly on the linking page.

How link placement on the page changes its value

Bill Slawski, who spent years reading these filings, documented how the reasonable surfer model weights links by features like font size, position on the page, and how topically related the linking page is to the target. A link high in the body content, inside the main editorial text, carries more modeled click probability than the same link stranded in a sidebar, a footer, or a block of boilerplate that repeats across the whole site.

I stopped caring about sitewide footer links years ago because of this. I audited a manufacturer once with eight thousand backlinks, almost all of them a single template link in a partner's footer, repeated across every page that partner published. Eight thousand links, one editorial decision, a modeled click probability near zero. The backlink tool showed a number that made the owner feel rich. The reasonable surfer model showed a link nobody would ever click. The search rankings sided with the model, not the tool.

Why anchor text raises or lowers a link's value

Anchor text feeds the same model. Google reads the words in and around the anchor text to understand what the linked page is about, so descriptive, relevant anchor text raises a link's modeled value and generic anchor text like "click here" lowers it. The anchor text is also where a profile of otherwise decent backlinks quietly signals manipulation, which is the part most people get wrong and the part Google's spam systems are built to catch.

Why authority and relevance decide a backlink's quality

Authority and relevance are not two separate checkboxes. They work together. PageRank flows through the link graph from page to page, and relevance gates how much of that flow actually helps you rank for a query. A backlink from a high authority site on an unrelated topic still passes some value. A backlink from a relevant page in your niche passes value that compounds with the topical signals already on your own page.

The old picture of authority spreading outward from a set of trusted seed sites is still the cleanest way to think about it. A page two links from a source Google deeply trusts inherits more authority than a page twenty links away through a chain of thin sites. That distance from trust is the real thing a domain authority score is trying to approximate, and it is why a single relevant link from an authoritative site beats a hundred links from sites Google barely trusts.

Relevance is the other half of the link graph

Relevance is the half of backlink quality that the domain rating crowd keeps underrating. Google does not just count that a strong page linked to you. It reads whether that page, and the site around it, covers content related to yours. A relevant backlink from a mid authority page in your industry often moves rankings more than an irrelevant link from a giant site, because the topical match tells search engines the link is a genuine editorial reference rather than a placement. When I map a client's best performing pages against their backlinks, the links that correlate with movement are almost always the relevant, in content ones, not the high domain rating links from unrelated sites.

Why the content around a backlink matters

A backlink does not arrive in a vacuum. Google reads the content of the linking page and the pages around it, and a link sitting inside relevant, well written content on a site that publishes useful pages is worth more than the same link dropped into thin, scraped content. This is the practical reason high quality content earns high quality backlinks. People link to pages worth citing, and search engines can tell an editorial reference from a manufactured placement. When I grade a backlink, I read the linking page the way a reader would. If the content is real and the site looks like it serves an audience, the link probably carries value. If the page exists only to host links, it does not, no matter what its domain rating says.

Domain authority, site authority, and what Google actually computes

Here is where the 2024 leak earned its reputation. The leaked internal documentation, Mike King's analysis, contains an attribute called siteAuthority, a site level quality score Google computes and stores per document, despite years of public statements implying no sitewide authority score existed. So Google does compute something like domain authority. It just is not the third party number on your dashboard, and you cannot see it.

It just is not the third party number on your dashboard, and you cannot see it.

The backlinks you earn feed signals like that one, and the pace matters. Ahrefs found that top ranking pages earn new referring domains 5–14.5% per month. That number reframes the quality versus quantity argument. Quality backlinks are not a one time trophy. They are something the strongest pages accumulate continuously from relevant, authoritative sites, which is the kind of durable authority that real backlinks are built on.

Anchor text and the link spam signals Google tracks

Anchor text is where most backlink profiles get themselves in trouble. The reasonable surfer model already told us descriptive anchor text raises a link's value. The spam systems tell us what happens when you push it too far. If most of your backlinks use the exact keyword you want to rank for as the anchor text, the profile looks engineered, because real editorial links almost never behave that way. Natural links use your brand, your bare URL, the article title, or a plain phrase that fits the sentence. A healthy profile shows a natural spread of anchor text, and a profile where every anchor is a commercial keyword is the easiest low quality pattern for Google to spot.

The leak put names on signals SEOs had only theorized about. The documentation includes a feature called phraseAnchorSpamDays, which lets Google measure the velocity of spam anchor text, the spikes of identical commercial anchors that arrive in a tight window. There is droppedLocalAnchorCount, which suggests Google simply does not count some anchors at all. These map straight onto Penguin, the 2012 update that started demoting sites for unnatural anchor text patterns before it folded into the core algorithm. Google's link spam policies now describe an automated system, SpamBrain, that neutralizes manipulative links by ignoring them.

The links Google counts as low quality backlinks

That shift from penalty to silence is the part most people miss. For years the fear was a manual penalty. Today the common outcome is that Google identifies a low quality or manipulative link and gives it zero weight. You built it, you paid for it, and it does nothing. I see this in audit after audit where a client bought a package of backlinks a year ago and cannot understand why search traffic never moved. The links were not penalized. They were disregarded, the same way Google handles links from a private blog network or a link farm. There is nothing to disavow, because the algorithm already filed those backlinks under noise. Run the linking website through the same questions and the verdict is obvious. Thin content, no real audience, every outbound link for sale. That is a low quality website, and the low quality backlinks it hands out inherit that judgment. If you want links that survive that filter, the link building strategies that algorithms actually respect start from earning the link on a relevant, high quality page, not placing it.

Dofollow, nofollow, and the links that pass no value

The dofollow versus nofollow split is real but oversimplified. A dofollow link is eligible to pass PageRank. A nofollow link carries a rel attribute that, since 2019, Google treats as a hint rather than a strict command, so it may or may not count. Both kinds send referral traffic and both belong in a natural profile, which is why chasing a pure dofollow profile is its own unnatural pattern that the spam systems can read.

What the dofollow debate ignores is that plenty of dofollow links pass nothing anyway. The leak documented a metric called sourceType that ties a link's value to where the linking page sits in Google's index. Google stratifies its index into tiers, and a dofollow backlink from a rarely crawled page in a low tier carries less value than a link from a page in the freshest, most important tier. The two links look identical in your backlink tool. Inside Google, one is currency and the other is close to worthless. That gap between counting backlinks and understanding them is the same one I covered in the backlink obituary everyone keeps writing too early.

How to judge high quality backlinks the way Google does

Quality over quantity is true, but it gets repeated so often that people forget the quantity half still matters. You do not need one perfect backlink. You need a steady supply of genuinely good ones, which is what that Ahrefs growth data shows the winners building. The pages that win keep building high quality backlinks to high quality content on their own website month after month, and let the search authority compound naturally. The right reading is not fewer links. It is more links that each clear the bar the reasonable surfer model sets.

So here is what I actually do with a backlink profile, and it is not the green and red spreadsheet. I read each meaningful backlink against the features Google scores. Is the link inside the body content where a reader might click it, or stranded in a footer. Is the linking page topically relevant to mine, or a generic roundup. Does the anchor text read like something a human editor wrote, or like a target keyword someone pasted. Is the linking page itself the kind of page Google indexes and trusts, or a thin page in a forgotten tier. Those four questions match the patents. A domain rating does not. Score your backlinks that way and the quality backlinks separate themselves from the noise, whatever authority number a tool assigns them.

The five bullet checklists were never lying to you. They described the shadow on the wall instead of the object casting it. Authority, relevance, anchor text, placement, and follow status are real because the reasonable surfer model, PageRank, and the spam systems make them real. If you want to build quality backlinks that move search rankings, earn them the slow way on relevant, high quality content, treat HARO for editorial mentions rather than manufacture anchors, and bring in help with link building when the math says you are behind. Google has told us, in patent after patent, exactly what it measures in a backlink. The least we can do is read the files.

By Michael McDougald