Decoding Google’s Ranking System: What the Massive Document Leak Revealed

Oregon, United States, June 27, 2024 The papers show how Google Search ranks things based on clicks, links, text, entities, Chrome data, and more. A bunch of leaked Google papers have given us a look inside Google Search that weve never had before. They also showed us some of the most important things Google uses to rank content.

The SEO community is shocked by a huge leak of Google Searchs private ranking documentation. The leak, which showed more than 14,000 possible ranking factors, gives us a look inside Googles closely guarded search results system that has never been seen before.

There was a leak of a Google API document that was given to Rand Fishkin of SparkToro. Fishkin then asked Michael King of iPullRank to help get the word out about the story.

The files that were leaked came from a Google API document change called yoshi-code-bot /elixer-google-api. This does not mean that someone hacked or leaked information.

Most SEOs fall into one of three groups:

All of what Google tells SEOs is true, and we should take their words as gospel. I call them the Google Cheerleaders.

• You cant believe a word that Google says. (They seem like bad SEOs to me.)
• As Per Aaron Rains SEO, Google isnt always right, so you have to check everything to see if you can find it. This is the camp I belong to, and Ill call it Bill Slawski rationalism because he was the one who persuaded me to believe this way.
• After this leak, I think a lot of people will change their minds.
• You can find all the files here, but you should know that there are more than 14,000 different ranking signals or features. It would take you all day (or night, in my case) to go through them all.
• I read the whole thing and boiled it down to a 40-page PDF, which Im now turning into a Search Engine Land outline.

In addition to telling you what I think, Im also telling you the names of the rating factors so you can search the database on your own. I want everyone to come to their own conclusions.

Important things from the Google Search leak

• PageRank (now defunct) has been changed by nearest seed. The algorithm is called pageRank_NS, and it has to do with reading documents.
• Google talks about seven different kinds of PageRank. The well-known ToolBarPageRank is one of them.
• Google uses a certain way to find news, YMYL, personal blogs (small blogs), ecommerce, and video sites that are good for business. Its not clear why Google is looking for personal blogs in particular.
• It looks like navBoost, NSR, and chardScores are the most important parts of Googles algorithm.
• A site-wide authority metric and a few site-wide authority cues, such as traffic from Chrome browsers, are used by Google.
• In its score system, Google looks at page embeddings, site embeddings, site focus, and site radius.
• Google tracks bad clicks, good clicks, clicks, clicks that last the longest, and impressions across the whole page.

Why does Google only screen personal blogs and other small sites?

Why did Google say many times in public that they dont have a way to measure name or site authority?

• What was Googles lie about how they used click data? Why does Google have seven different kinds of PageRank?
• Even though I dont know the answers, Im sure that people in the SEO community would love to know them.

Things that stand out: Favorite finds

Page Quality (PQ) is something that Google has. Googles use of an LLM to figure out article pages effort is one of the most interesting parts of this measurement. This number sounds like it would help Google figure out if a page is easy to copy.

This means that tools, pictures, movies, unique information, and a lot of information are some of the best ways to get high scores on effort tests. By chance, these things have also been shown to make people happy.

Subject boundaries and subject power look like they are real

Topical authority is an idea that came from Googles study into patents. If youve read the patents, youll see that this leak backs up a lot of what SEOs have learned from patents.

We can see from the algo leak that siteFocusScore, siteRadius, siteEmbeddings, and pageEmbeddings are used to rank sites.

• The siteFocusScore shows how focused a site is on a certain subject.
• The siteRadius tool checks how far page embeddings are from the site embedding. Plain English: Google gives your website a subject identity, and each page is judged by that identity.

SiteEmbeddings are embeddings of sites or pages that are reduced

Contact Info:
Name: Aaron Rains

Company: Aaron Rains SEO

Website: https://rainsaaronseo.com/

Email: [email protected]

You May Also Be Interested In

• Southgobi Announces Issue Of Payment-In-Kind Interest Shares Under Convertible Debenture
• AidKit Wins Gold and "Best of Category" for GovTech Innovation at 2026 Globee Awards for Technology
• Antimony Resources Corp. (ATMY) (ATMYF) (K8J0) Appoints Former Director and Professor of the United States Military Academy as Advisory to the Board of Directors