Google TrustRank is the idea that search engines can estimate how trustworthy a page is by starting from a small set of hand-vetted “seed” sites and watching how trust flows outward through links. Pages close to clean, reputable seeds inherit more trust; pages that get most of their links from spam inherit very little. The term comes from a 2004 Stanford and Yahoo paper. Google has never confirmed a literal “TrustRank score.” The underlying logic (trust as a property of the link graph) is alive in how modern search and AI Overviews treat authority.
Google TrustRank
Google TrustRank is a link-analysis concept that estimates a page’s trustworthiness by propagating trust scores outward from a small, manually or algorithmically curated set of reputable seed sites, demoting pages that draw most of their links from spam.
Where TrustRank Actually Came From
Most “TrustRank” advice online quietly confuses two different things, so separate them up front.
- The academic paper. “Combating Web Spam with TrustRank” (Gyöngyi, Garcia-Molina and Pedersen, 2004) proposed a concrete algorithm: pick a seed set of human-evaluated good pages, then run a PageRank-style propagation that flows trust through links, attenuating with each hop. It was a research method for fighting spam, and nothing in it describes a Google ranking factor.
- The marketing folklore. Over two decades, “TrustRank” mutated into a catch-all SEO buzzword for “how much Google trusts your domain.” Google has never published a metric by that name, and John Mueller has said for years that Google does not use third-party scores such as Domain Authority. Google’s internal API documentation that leaked in 2024 did reference a “siteAuthority” attribute, but it says nothing about how that value is computed or how much it weighs.
The honest practitioner position: TrustRank is a useful mental model for how trust propagates through links and brand signals. There is no number to chase. We use it as a lens and keep it off the dashboard.
This matters because third-party tools sell proxies (Majestic’s Trust Flow, Moz’s Domain Authority, Ahrefs’ Domain Rating), and clients assume these are Google’s TrustRank. They are vendor models that correlate with authority, built by companies outside Google. Treat them as directional estimates.
How TrustRank Propagation Works
The mechanism is simpler than the jargon suggests. It’s PageRank with a moral compass.
1. Seed selection
The algorithm starts with a small, curated set of seed sites: domains a human (or a high-confidence classifier) has judged clearly trustworthy. Think major news outlets, government and university sites, recognized brands and standards bodies. The quality of the entire system depends on this seed set being clean.
2. Trust propagation through the link graph
Trust flows outward from seeds along links. A page linked directly from a seed inherits high trust; that trust then passes to pages it links to, and so on. Crucially, trust decays with distance: each hop applies a damping factor, so a page five clicks from any seed receives far less inherited trust than a direct neighbor. This is why a single link from a genuinely authoritative domain can outweigh a hundred links from the web’s sludge.
3. The inverse: spam demotion
The complementary techniques (Anti-TrustRank and spam-mass estimation, both published in 2006) run the same logic from known-bad pages. Pages that receive most of their score from spammy neighborhoods get flagged. In practice a clean profile means earning good links and also staying structurally clear of link farms and unnatural schemes that drag your neighborhood down.
| Signal | Increases inherited trust | Decreases inherited trust |
|---|---|---|
| Link source | Editorial link from a high-trust domain | Link from a spam network or PBN |
| Link type | In-content, followed | Paid, UGC, or nofollowed at scale |
| Placement | Body, contextually relevant | Footer or sidebar boilerplate, site-wide |
| Anchor profile | Natural, branded, varied | Exact-match commercial spam |
| Topical fit | Source and target share a subject | Off-topic link buys |
| Distance from seed | One or two hops | Many hops, deep in the graph |
TrustRank vs. PageRank vs. E-E-A-T
These three get blended into mush. They operate at different layers and answer different questions.
| Concept | What it measures | Layer | Confirmed by Google? |
|---|---|---|---|
| PageRank | Link-based importance (how much link equity flows in) | Link graph | Yes: the original algorithm, still one signal among many |
| TrustRank | Link-based trustworthiness (proximity to clean seeds) | Link graph | No literal score; the idea lives on in spam systems |
| E-E-A-T | Experience, Expertise, Authoritativeness, Trustworthiness | Quality raters and systems | Yes: central to E-E-A-T guidance |
PageRank asks “is this page popular?” TrustRank asks “is this page’s popularity coming from a clean part of the web?” E-E-A-T is broader still: it’s the human-evaluated quality framework that Google’s systems try to approximate, and it pulls in author identity, accuracy and reputation well beyond links. The three reinforce each other, and none of them can stand in for another.
Does TrustRank Still Matter in the AI Overviews Era?
Yes, arguably more, because the consequences of getting cited (or skipped) are higher now.
When AI Overviews and other generative answers pull from the web, they lean on sources that look authoritative and uncontroversial. Link-trust logic of the kind TrustRank formalized helps keep a low-trust, spammy page out of a generated answer. A domain in a clean neighborhood with editorial links from recognized sources is a safer pick for an AI to cite. A backlink profile that reads like a 2013 link-buying spree makes you a liability the system would rather avoid.
For YMYL topics (health, finance, legal), the trust bar is highest, and a weak trust profile can quietly cap your ceiling no matter how good the content is.
How We Build Real Trust
Here’s the practitioner playbook. The goal is to become the kind of site trust naturally flows to, because there is no score to manipulate.
Earn editorial links. A handful of in-content links from genuinely relevant, reputable domains beats hundreds of cheap ones. Original data, linkable assets and real PR outreach are how you get backlinks that propagate trust. Buying links invites the spam-mass demotion you’re trying to avoid.
Audit the neighborhood you’re in. Pull your backlink profile, flag toxic or paid links, and disavow only as a documented last resort. Proximity to spam is itself a negative signal; clean adjacency matters.
Keep your link types honest. Use nofollow, sponsored and UGC attributes correctly. Passing off paid links as editorial is exactly the unnatural pattern these systems are built to catch.
Make trust legible on the page. Real author bylines with credentials, transparent About and Contact pages, citations to primary sources and current revision dates. This feeds the E-E-A-T layer that complements link trust.
Structure the site so trust flows internally. A clean site structure and pillar-cluster internal linking route inherited trust from your strongest pages to the ones you actually want ranking. Trust you earn at the domain edge is wasted if it can’t reach money pages.
Avoid the obvious penalty triggers. Link schemes, exact-match anchor stuffing and PBNs are the fastest route to a Google penalty that no content fixes.
Our north star: if a skeptical editor at a reputable publication would link to your page on its merits, you’re building trust the durable way.
When a trust problem becomes a business problem
A young site with a few dozen natural links has nothing to clean up. Keep publishing things worth citing and skip the “toxic link audit” someone is trying to sell you.
It starts to cost money in a few recognizable situations. Search Console shows a manual action for unnatural links. A previous agency left years of paid links, PBN placements or exact-match anchors behind. You bought an expired or aftermarket domain and inherited its neighborhood. Or your content is clearly better than the pages outranking you, on a YMYL topic where trust decides who gets shown.
On a fractional SEO engagement, the two founders read the link profile, document what to disavow and why (often very little), and build editorial links that hold up. The Gyfted growth ran on $7k of total backlink spend.
Frequently Asked Questions
Is Google TrustRank a real ranking factor?
There’s no confirmed metric called “TrustRank” inside Google. The term comes from a 2004 Stanford and Yahoo paper on fighting web spam. Google does use link-trust and spam signals that work on a similar principle (judging links by the company they keep), but no public “TrustRank score” exists. Treat it as a model and stop looking for a number.
What’s the difference between TrustRank and PageRank?
PageRank measures link-based importance: how much link equity flows into a page. TrustRank measures link-based trustworthiness: how close a page sits to a vetted set of clean “seed” sites. A page can have high PageRank but low trust if its links come from spammy neighborhoods, which is exactly the gap TrustRank was designed to catch.
How do I improve my site’s TrustRank?
There is no score to game, so the work is becoming trustworthy. Earn editorial links from relevant, reputable domains; audit your backlinks and disavow only clear schemes; publish credentialed authorship and primary-source citations; label paid and UGC links correctly; and route internal trust to key pages through clean site structure. Trust accrues to sites that deserve it.
Are Majestic Trust Flow and Domain Rating the same as TrustRank?
No. Trust Flow (Majestic), Domain Authority (Moz) and Domain Rating (Ahrefs) are third-party models that correlate with authority and trust. Google does not use them, and none of them is Google’s internal metric. They’re useful for directional benchmarking and spotting toxic neighborhoods.
Does TrustRank affect AI Overviews and AI search?
Indirectly, yes. Generative answers favor sources that look authoritative and low-risk. The same link-trust logic that demotes spam in classic ranking makes a clean, well-linked domain a safer citation for AI Overviews and chat answers. A toxic backlink profile reduces your odds of being cited, especially on sensitive YMYL topics.