Google Search Documents Leak (2024)

The Google Search documents leak was the publication in May 2024 of thousands of pages of internal Google API documentation for its Content Warehouse. The files listed data attributes that Google stores about pages, sites and links, giving SEOs a rare glimpse at the kinds of information its systems can use.

Written by Paul Hoda, SEO consultant since 2005. Last updated 8 November 2026.

What happened

The documents appear to have been accidentally published to a public code repository in early 2024. They were shared with SEO industry figures, and in late May 2024 Rand Fishkin and Mike King published detailed analyses. Google then confirmed to journalists that the documents were authentic, but warned against making assumptions based on out-of-context, outdated or incomplete information.

The material was technical reference documentation: lists of modules and attributes with short descriptions. Reports put the total at around 2,500 modules and over 14,000 attributes.

What it contained

Attributes that drew the most attention included:

  • References to NavBoost and click-related fields, such as counts of good and bad clicks.
  • A field described as site authority, which many read as a site-wide quality measure.
  • Data apparently linked to Chrome browsing.
  • Information about links, anchor text and how trusted a linking page is.
  • Fields about authors, dates and whether a site is a small personal site.
  • “Twiddlers”, which are re-ranking functions that adjust results after the main ranking, and various demotions.

Why it caused a stir

Some attributes seemed to conflict with past public statements. For example, Google representatives had at times played down the role of clicks and denied using anything like domain authority. The leak suggested Google at least stores data related to these ideas.

What it does not prove

It is important to be careful. The documents:

  • Show that an attribute exists, not how much weight it carries, or whether it is used in ranking at all.
  • Include deprecated fields and may not reflect current systems.
  • Do not explain how attributes combine into rankings.

Much of the online commentary went further than the evidence. Claims of having “cracked” the algorithm should be treated as theories.

Sensible lessons for SEO

The leak mostly reinforced long-standing good practice rather than revealing shortcuts:

  1. Satisfy searchers, since user signals such as clicks appear to be recorded.
  2. Build a recognisable brand that people search for and choose.
  3. Earn links from relevant, trusted sites rather than chasing volume; see link building.
  4. Keep content fresh and accurate where dates matter.
  5. Focus on site-wide quality, as some signals appear to apply to whole sites.

It also showed why it helps to compare what Google says with what testing and evidence show, a theme familiar from discussions of ranking factors.

Common misconceptions

  • “The leak is Google’s algorithm.” It is API documentation, not source code or weightings.
  • “Every attribute is a ranking factor.” Many fields may be used for other purposes or not at all.

Key Takeaways

  • Internal Google Search API documents became public in May 2024.
  • Google confirmed they were authentic but warned they lacked context.
  • They mentioned NavBoost, click data, site authority and Chrome data.
  • The documents show attributes exist, not how they are weighted.
  • The main lessons echo good practice: satisfy users, build a brand and earn real links.

Related Terms

Wondering what to fix first on your site? Get a free check of your site’s SEO, or keep learning in the SEO Encyclopedia.