Google search leak: what the leaked ranking factors mean for B2B SEO

In May 2024, more than 2,500 pages of internal Google API documentation leaked, containing thousands of possible ranking factors. For anyone without the time to read it all: the leak mainly proves that click behaviour, Chrome data, backlinks and author authority count, while Google claimed for years that some of them were not used. Good news for B2B: there is no secret trick in there. The message is that you need to build a site that leaves real users satisfied. In this article you will read what the google search leak revealed and what it means concretely for your approach.

Measure it yourself: turn your organic traffic into revenue with our SEO ROI calculator.

What exactly is the Google search leak?

The leak is a collection of 2,596 modules with 14,014 attributes from Google’s internal API documentation. The documents were first shared with Rand Fishkin (founder of SparkToro), who had them analysed by SEO expert Mike King (CEO of iPullRank). The original source turned out to be Erfan Azimi of EA Eagle Digital. Their analyses appeared on 27 May 2024.

The timeline in brief:

• March 2024: the documents accidentally appear in a public code repository (GitHub) and are picked up by an automated service.
• 5 May 2024: Fishkin is handed the documents and starts verifying them.
• 27 May 2024: Fishkin and King publish their findings.
• 29 May 2024: Google confirms the leak is genuine, via Search Engine Land and The Verge among others.

Google responded cautiously and warned against incorrect assumptions based on “out-of-context, outdated or incomplete information”. In other words: the documents are real, but we do not know how much weight each factor gets, and some parts may be outdated. So read the insights as direction, not as law.

Which ranking factors did the leak reveal?

The documents confirm a number of things Google had publicly denied or kept vague. The most important ones:

Chrome user data counts

Google claimed since 2012 that it did not use data from Chrome for rankings. The leak suggests the opposite: attributes such as Chrome_trans_clicks and ChromeInTotal indicate that Google uses click behaviour from Chrome to assess a site’s popularity and quality. The same data would also help determine which pages appear as Sitelinks under a search result.

Click behaviour and NavBoost

A second denied signal: click behaviour in the search results themselves. The leak confirms the existence of NavBoost and Glue, two systems that use clicks to raise or lower rankings. They also came up in Google’s antitrust case with the US Department of Justice. Click modules such as goodClicks, badClicks and lastLongestClicks show up in the documentation.

That last one matters: Google appears to measure whether someone stays on your page after a click, or returns straight to the search results because they did not find what they were looking for. For you that means: if your content does not match search intent, you perform worse, no matter how technically sound your page is.

PageRank is still alive

The patents around PageRank expired in 2019, but the documentation shows the concept is still alive, in line with Google’s own guide to ranking systems. Interesting is PageRank_NS (NS stands for Nearest Seed): links close to a group of authoritative “seed sites” would carry more weight. So backlinks are not dead, they remain a vote of trust.

Link diversity over link volume

The leak stresses that the quality, diversity, freshness and relevance of backlinks weigh more heavily than sheer volume. Google wants to see links from a varied range of authoritative sites and can ignore irrelevant links. Spammy links can harm your site. That fits seamlessly with how solid link building has always been supposed to work.

Authors as entities

Authorship is the only E-E-A-T component that appears in detail in the leak. Google stores authors as text and checks whether an entity on a page is also the author. So there is a form of author measurement in the system. For B2B this is the most underused opportunity, and we will come back to it shortly.

Separate treatment for YMYL and niches

The documentation contains classifiers that assign scores to YMYL content (“Your Money or Your Life”, think health, finance and legal advice). Local sites and product review sites also get their own ranking factors. One term stood out: Twiddlers, functions that can adjust the original ranking and push content down, for instance with dissatisfied users, exact-match domains or bad links.

What does this mean concretely for B2B SEO?

Here we depart from the panicked tone you read elsewhere. Honest advice: the leak changes nothing about the fundamentals. It confirms them. What does change is that you now have it in black and white why certain things pay off. For a Belgian B2B company with a small team, that translates into a few clear choices.

Aim for intent, not traffic. Because click behaviour counts, it hurts you when visitors drop off. In B2B that is doubly true: your niche has small search volumes, so every visitor counts. A page that ranks for a term but does not answer the reader’s question generates badClicks and slips. That aligns with what Google has stressed about helpful content for years. So steer on quality visitors who convert, not on a high session count in your dashboard.

Make your site fast and clear. If Chrome data and click signals carry weight, a slow or confusing site is a direct ranking risk. People clicking back to Google are a negative signal. This is where technical SEO and a good user experience meet.

Invest in author authority. This is THE B2B opportunity from the leak. In B2B a purchase decision revolves around trust, and Google measures authors as entities. So let real experts write, not anonymous content. Concretely:

• Give every author an extensive bio with role, experience and credentials.
• Link to professional profiles (LinkedIn, trade publications).
• Collect all of an author’s articles on their bio page.

Build a varied, honest link profile. No bulk buying of links, but relevant backlinks from authoritative sites in your sector. Diversity and relevance weigh more heavily than volume.

What you should NOT do: use the leak as an excuse for an expensive technical overhaul of everything at once. Many attributes may be outdated, and nobody knows the weights. Anyone selling you a guaranteed recipe based on this leak is selling you an illusion. The gain lies in consistently good SEO strategy, not in a secret button.

Does this change how you measure SEO?

Not really, and that is reassuring. The leak actually underlines that you should steer on the right things. Click behaviour and user satisfaction are not vanity metrics: they are tied to whether your content genuinely helps people. But the final measure for B2B stays the same: leads and revenue, not rankings or visitor counts.

A ranking in position three that produces not a single lead is worth less than a ranking in position eight that brings in a quote request every week. Use the leak insights to improve your content and your technical setup, but judge the result by the pipeline. Anyone who wants to know more about realistic expectations should read how long SEO takes.

Need help turning the insights into results?

The Google leak is interesting, but insights are worth nothing without execution. We are a small Belgian team that moves fast, advises honestly (including when something does NOT pay off) and ties SEO to customers and revenue, not to vanity numbers. No webshop tricks, but a B2B approach aimed at leads. Want to know what these ranking factors mean concretely for your site?

Book your free intake.

Frequently asked questions about the Google search leak

Is the Google leak real or a rumour?

Real. Google confirmed it was genuine on 29 May 2024, via Search Engine Land and The Verge among others. It concerns 2,596 modules with 14,014 attributes from internal API documentation, first analysed by Rand Fishkin and Mike King.

Does Google use click behaviour for rankings?

The leak strongly points that way. Modules such as goodClicks, badClicks and lastLongestClicks and systems such as NavBoost suggest that Google measures whether users are satisfied after a click. Content that matches search intent and keeps people on the page performs better.

Are backlinks still important after the leak?

Yes. The documentation shows that PageRank is still alive and that link diversity, quality and relevance count. The number of links matters less than where they come from. Spammy links can do harm.

What are Twiddlers?

Twiddlers are functions that can adjust the original ranking and push content down, for instance on signals of user dissatisfaction, exact-match domains or poor quality links.

Should I overhaul my SEO approach now?

No. The leak mainly confirms what good SEO always did: build a fast, clear site with strong content from credible authors and an honest link profile. Because the weights are unknown and some data is outdated, a complete about-turn is unwise.