During the Google SEO office hours session, Duy Nguyen of Google’s Search Quality team answered the question of what links and trust have to do with links on spammy sites.
It was interesting to hear that a Google employee said they were protecting anchor text signals. It’s not generally discussed.
Building trust with Google is an important consideration for many publishers and SEOs.
The idea is that “trust” helps your site index and rank properly.
It’s also known to have no “trust” metric, which can confuse some of the search community.
How can an algorithm be trusted if it’s not measuring something?
Googlers don’t really answer that question, but there are patents and research papers that give us ideas.
Google doesn’t trust links from spam sites
A person who submitted a question to SEO Office Hours asked:
“If a domain is penalized, will links sent from that domain be affected?”
Google employee Duy Nguyen replied:
“By ‘penalty’, I assume you mean that the domain was downgraded by a spam algorithm or by manual action.
In general, yes, I don’t trust links from sites that are known to be spam.
This helps maintain anchor signal quality. ”
trust and link
Googlers talk about trust, but they’re clearly talking about what their algorithms trust and don’t trust.
In this case, it’s not that we don’t count links on spam sites, but that we specifically don’t count anchor text signals.
The SEO community talks about “building trust”, but in this case, what really matters is not building spam.
How does Google determine if a site is spam?
Not all sites are penalized or subject to manual action. Some sites aren’t indexed, but that’s the job of Google’s spam brain. Spam Brain is an AI platform that analyzes web pages at various points in time, starting at crawl time.
The Spam Brain Platform works like this:
- indexing gatekeeper
The spam brain blocks sites with content discovered through Search Console and sitemaps as they are crawled. - track indexed spam
Spam Brain also detects spam that is indexed at the time a site is considered for ranking.
How the Spam Brain platform works is to train an AI based on Google’s knowledge of spam.
Here’s what Google has to say about how the spam brain works:
“Over the past year, our deep knowledge of spam combined with AI has allowed us to build our own anti-spam AI that is very effective at catching both known and emerging spam trends.”
I don’t know whatknowledge of spamGoogle is talking about it, and there are various patents and research papers about it.
For those interested in learning more about this topic, consider reading an article I wrote on the concept of link distance ranking algorithms, a method for ranking links.
We also published a comprehensive article on multiple research papers describing link-related algorithms that may explain what the Penguin Algorithm is.
Many of the patents and research papers are from within the last decade, but none have been published by search engines or university researchers since then.
The significance of these patents and research papers is that they may be incorporated into Google’s algorithms in other ways, such as training or spam-brain-like AI.
The patent discussed in the Link Distance Ranking article describes a method for assigning page ranking scores based on the distance between a set of authoritative “seed sites” and the pages they link to. Seed sites are like a starting point for calculating which sites are good and which are not (that is, spam).
Intuitively, the farther a site is from a seed site, the more likely it is to be considered spam. This part of judging spamminess by link distance is explained in the research paper cited in the previously referenced Penguin article.
Patent, (Use Web Link Graph Distance to Create Page Rankings), which explains:
“The system then assigns link lengths based on the properties of the link and the page it connects to.
The system then calculates the shortest distance from the set of seed pages to each page in the set of pages based on the length of the links between the pages.
The system then determines a ranking score for each page in the set of pages based on the calculated shortest distance. ”
Reduced link graph
The same patent also mentions what is known as a contracted link graph.
But there’s more than one patent that discusses the shrinking link graph. Collapsible link graphs have also been explored outside of Google.
A link graph is like a map of the Internet created by mapping links.
The reduced link graph removes low quality links and related sites.
What remains is the so-called shrinking link graph.
The following is an excerpt from the Google patent cited above.
“Reduced Link Graph”
Note that the links participating in the k shortest paths from the seed to the page form a subgraph containing all the links “flow” ranked from the seed.
Although this subgraph contains far fewer links than the original link graph, the k shortest paths from the seed to each page in this subgraph are the same length as the paths in the original graph. Become.
…Furthermore, the rank flow to each page can be traced back through this subgraph path to the nearest k seeds. ”
Google doesn’t trust links from sites that have been penalized
In some ways, it’s no surprise that Google doesn’t trust links from penalized websites.
However, you may not know if your site has been penalized by the spam brain or marked as spam.
Before attempting to get a link from a site, we recommend researching whether the site is untrustworthy.
In my opinion, third-party metrics should not be used for business decisions like this, as they hide the calculations used to generate the scores.
If your site already links to potentially spammy sites that have inbound links from potentially paid links such as PBNs (private blog networks), then the site is probably a spam site.
Featured Image by Shutterstock/Krakenimages.com
Check out our SEO office hours:
