TrustRank is a link analysis method from a 2004 paper by Zoltán Gyöngyi and Hector Garcia-Molina of Stanford and Jan Pedersen of Yahoo. It starts from a small set of human-reviewed trustworthy seed pages and propagates trust through links, so pages far from the seeds are more likely to be spam. It is not a Google system. Google filed a trademark for the name for an anti-phishing filter, and Matt Cutts said Google has nothing specifically called trust rank. Proximity to reputable sites is still a sensible link audit.
What is TrustRank?
TrustRank is a link analysis method for separating reputable pages from web spam, published in 2004 as Combating Web Spam with TrustRank by Zoltán Gyöngyi and Hector Garcia-Molina of Stanford University and Jan Pedersen of Yahoo, at the VLDB conference.
It is not a Google algorithm. Its authors were at Stanford and Yahoo, and the idea has been widely misattributed to Google in SEO writing since.
We first select a small set of seed pages to be evaluated by an expert. Once we manually identify the reputable seed pages, we use the link structure of the web to discover other pages that are likely to be good.
Zoltán Gyöngyi, Hector Garcia-Molina and Jan Pedersen, Combating Web Spam with TrustRank, VLDB 2004
How does TrustRank work?
The method rests on one observation: good pages rarely link to spam. Human reviewers label a small seed set as trustworthy, then trust flows outwards through links, in the same way PageRank weight flows, but starting only from the seeds. Trust weakens with each hop, so a page many links away from any seed receives little.
The same logic can run in reverse, sometimes called anti-TrustRank: pages close to known spam are more likely to be spam. That is why linking out to spam networks is a risk in its own right.
Does Google use TrustRank?
This is General theory for Google. There is no public evidence that Google uses the Stanford and Yahoo TrustRank algorithm, and the confusion has three separate sources that are often blended together.
First, Google filed a trademark application for the name TrustRank at around the time the paper came out. Matt Cutts, then head of Google's webspam team, explained in a December 2007 comment on his blog that it was for an anti-phishing filter, not a ranking system. Second, Google holds an unrelated patent, US7603350, Search result ranking based on trust, filed in 2006 by Ramanathan Guha, which ranks results using trust in the people or entities who labelled documents; it is not the link-based TrustRank and a patent is not proof of use. Third, Google staff do talk about trust in general terms, which is not the same as a system of that name.
Yahoo wrote a paper with that term. At nearly the same time, we filed for a trademark for an anti-phishing filter.
Matt Cutts, Google, Tons of PubCon interviews on video and audio (comment), Matt Cutts blog
What has Google said about trust as a ranking concept?
In 2011 Cutts said Google did not have anything specifically called trust rank, and described trust as a catch-all term, with PageRank as the best known kind of trust, as reported by Search Engine Roundtable.
Wikipedia's TrustRank article states that the algorithm is part of Google, without a supporting citation. We treat that as unverified.
It's not that we have something specifically called trust rank.
Matt Cutts, Google, reported by Barry Schwartz, Google's Cutts: No Such Thing As Trust Rank, Search Engine Roundtable
What does TrustRank mean for sites?
Drop the idea of a Google TrustRank score. Keep the testable principle, which is consistent with how Google describes link analysis: links from established, reputable sites in your sector are worth more than volume from sites nobody trusts, and linking out to spam associates you with it.
The falsifiable audit: list the ten to twenty most reputable organisations in your sector (regulators, trade bodies, leading publications, universities). If none of them, and no site that they link to, links to you, the site sits a long way from any plausible seed set.
- Earn links from recognised sector bodies and publications, not just more links.
- Remove outbound links to spam, hacked or expired-domain sites.
- Stop describing any third-party trust metric as Google's TrustRank.
Which Laurelin audit checks test for TrustRank-style issues?
The new check, No links from recognised authoritative sources in the sector, tests whether any reputable sector organisation links to the site. It is a proxy for seed-set proximity, not a Google metric.
Related checks:
- Outbound links to spam: the anti-trust direction of the same logic.
- Toxic or spammy backlink profile: links from far outside any trusted neighbourhood.
- Low referring domain authority vs competitors: the strength side of link reputation.
- Citations to low-quality sources: what you link to reflects on you.
- Inauthentic bought mentions or links: bought links do not create real proximity to reputable sites.
What are the key dates for TrustRank?
- 2004: Combating Web Spam with TrustRank published at VLDB 2004 (Stanford and Yahoo) (source)
- 2006-05-09: Google files unrelated patent US7603350, Search result ranking based on trust (source)
- 2007-12-18: Matt Cutts says Google's TrustRank trademark was for an anti-phishing filter (source)
- 2009-10-13: US7603350 granted (source)
- 2011-12-06: Cutts: Google has nothing specifically called trust rank (source)
Frequently asked questions about TrustRank
Is TrustRank a Google algorithm?
No. TrustRank was published in 2004 by researchers at Stanford University and Yahoo. Google has not said it uses it.
Why do people call it Google TrustRank?
Google filed a trademark for the name, which Matt Cutts said was for an anti-phishing filter, and Google holds an unrelated trust-based ranking patent. The two got merged with the Yahoo paper in SEO writing.
Does trust matter for Google rankings?
Google staff use trust as a general term, and Cutts called PageRank the best known type of trust. There is no confirmed Google system named TrustRank.