What Can We Learn About SEO from Google's Leaked Documents?

[paulcarl]

Politics aside, Project Veritas released a new video on a Google whistleblower and they discuss some topics like blacklisted domains from search results, flagged search queries, and some other SEO-related stuff. Check it out here: https://www.projectveritas.com/2019...oes-public-says-burden-lifted-off-of-my-soul/

From that URL there is also a trove of leaked documents from Google, one that I'm currently looking at is their "Realtime Boost" slides which details how that process works (pic related).

Here's a quote from some guy's resume at Google as a part of the "Clicks Team"

I'm the manager for logs-based ranking projects. The team's efforts are currently split
among four areas:

• Navboost. This is already one of Google's strongest ranking signals. Current work is on automation in building new navboost data; recency and anti-grandfathering; and building an improved model of user behavior, taking into account impressions and display bias.
• Generalizing navboost: extending the system to help with (query, document) pairs for which we have no direct evidence.
• Mergeserver and navseek: These are really two sides of the same coin. The former is an experimental framework for comparing user behavior in the presence of changes to ranking or UI; the latter, a system for exploring new results for existing queries.

If anyone else is checking out these documents, maybe we can share some of the good SEO-related findings in this thread? So far I haven't found anything juicy, just some interesting tidbits.

 

[bernard]

Looks interesting indeed. I assume Navboost is an (AI?) prediction model that takes a query or even half a query and makes some judgement as to what previous searches preferred or didn't prefer: Did they need to refine the search after clicking position 1,2,3 or did they do something else etc. Just from the above, it does seem to indicate User Experience above all, which is nothing new to this forum.

 

[paulcarl]

It is interesting but after skimming through the files for about an hour I didn't really find anything useful for SEO. It's definitely possible that I missed stuff.