Fresh doc dump from DOJ v Google: courtroom slides reconfirmed the trifecta running Search: PageRank, Navboost, RankEmbed, even the mysterious Q*. Turns out it still runs on 100+ handbuilt signals, blended with ml.
www.searchengineworld.com/fresh-doc-du...
#SEO #Google
www.searchengineworld.com/fresh-doc-du...
#SEO #Google
Fresh Doc Dump: PageRank, Navboost, and RankEmbed: DOJ Papers Outline Google’s Stack
Google’s own court-filed slide deck just spilled more tea on Navboost, and the details punch a few holes in years of SEO guesswork. The newly released trove goes further, mapping more than 100 raw si...
www.searchengineworld.com
May 12, 2025 at 5:05 PM
Fresh doc dump from DOJ v Google: courtroom slides reconfirmed the trifecta running Search: PageRank, Navboost, RankEmbed, even the mysterious Q*. Turns out it still runs on 100+ handbuilt signals, blended with ml.
www.searchengineworld.com/fresh-doc-du...
#SEO #Google
www.searchengineworld.com/fresh-doc-du...
#SEO #Google
Fresh Doc Dump: PageRank, Navboost, and RankEmbed: DOJ Papers Outline Google’s Stack - www.searchengineworld.com/fresh-doc-du...
Fresh Doc Dump: PageRank, Navboost, and RankEmbed: DOJ Papers Outline Google’s Stack
Google’s own court-filed slide deck just spilled more tea on Navboost, and the details punch a few holes in years of SEO guesswork. The newly released trove goes further, mapping more than 100 raw si...
www.searchengineworld.com
May 13, 2025 at 9:19 AM
Fresh Doc Dump: PageRank, Navboost, and RankEmbed: DOJ Papers Outline Google’s Stack - www.searchengineworld.com/fresh-doc-du...
Great episode from Marie Haynes about the importance of user interaction signals, happy vs. unhappy users, and more. She covers Glue, Navboost, RankEmbed, etc.
I've been saying "Hell hath no fury like a user scorned" for a long time (medieval Panda days). Don't be on the wrong side of Navboost.
I've been saying "Hell hath no fury like a user scorned" for a long time (medieval Panda days). Don't be on the wrong side of Navboost.
What if user satisfaction is all that matters for ranking on Google?
YouTube video by Marie Haynes
www.youtube.com
January 26, 2026 at 4:03 PM
Great episode from Marie Haynes about the importance of user interaction signals, happy vs. unhappy users, and more. She covers Glue, Navboost, RankEmbed, etc.
I've been saying "Hell hath no fury like a user scorned" for a long time (medieval Panda days). Don't be on the wrong side of Navboost.
I've been saying "Hell hath no fury like a user scorned" for a long time (medieval Panda days). Don't be on the wrong side of Navboost.
Court rules Google must share data with competitors. Looks like they'll be forced to share:
1) Basic Search Index: URLs/first seen/last crawled/spam score/etc.
2) "Glue": uses click data to build SERPs
3) RankEmbed: deep learning to understand semantic meaning
www.courtlistener.com/docket/18552...
1) Basic Search Index: URLs/first seen/last crawled/spam score/etc.
2) "Glue": uses click data to build SERPs
3) RankEmbed: deep learning to understand semantic meaning
www.courtlistener.com/docket/18552...
September 2, 2025 at 9:12 PM
Court rules Google must share data with competitors. Looks like they'll be forced to share:
1) Basic Search Index: URLs/first seen/last crawled/spam score/etc.
2) "Glue": uses click data to build SERPs
3) RankEmbed: deep learning to understand semantic meaning
www.courtlistener.com/docket/18552...
1) Basic Search Index: URLs/first seen/last crawled/spam score/etc.
2) "Glue": uses click data to build SERPs
3) RankEmbed: deep learning to understand semantic meaning
www.courtlistener.com/docket/18552...
Google's RankEmbed system forced to share secrets in antitrust remedy #Google #RankEmbed #Antitrust #DeepLearning #SearchData
Google's RankEmbed system forced to share secrets in antitrust remedy
Deep learning model using 70 days of search data and human rater scores must be disclosed to qualified competitors under September 2 court ruling
ppc.land
September 4, 2025 at 6:01 AM
Google's RankEmbed system forced to share secrets in antitrust remedy #Google #RankEmbed #Antitrust #DeepLearning #SearchData
If you haven't read it yet, @mariehaynes.bsky.social w/a great rundown of what we learned from the latest trial doc:
"RankEmbed BERT is a ranking model that uses 70 days of search logs & the scores of quality raters who follow a detailed set of guidelines to assess a page's expertise, authority..."
"RankEmbed BERT is a ranking model that uses 70 days of search logs & the scores of quality raters who follow a detailed set of guidelines to assess a page's expertise, authority..."
What Google’s Trial Docs Reveal About Clicks, Links and Other Ranking Signals - Marie Haynes
The DOJ vs Google trial has finally wrapped up. The closing document is full of interesting insight into how Google’s algorithms work. While the final verdict of the trial is a separate story, the doc...
www.mariehaynes.com
September 7, 2025 at 1:34 PM
If you haven't read it yet, @mariehaynes.bsky.social w/a great rundown of what we learned from the latest trial doc:
"RankEmbed BERT is a ranking model that uses 70 days of search logs & the scores of quality raters who follow a detailed set of guidelines to assess a page's expertise, authority..."
"RankEmbed BERT is a ranking model that uses 70 days of search logs & the scores of quality raters who follow a detailed set of guidelines to assess a page's expertise, authority..."
Google's RankEmbed system forced to share secrets in antitrust remedy #Google #RankEmbed #Antitrust #DeepLearning #SearchData
Google's RankEmbed system forced to share secrets in antitrust remedy
Deep learning model using 70 days of search data and human rater scores must be disclosed to qualified competitors under September 2 court ruling
ppc.land
September 4, 2025 at 6:01 AM
Google's RankEmbed system forced to share secrets in antitrust remedy #Google #RankEmbed #Antitrust #DeepLearning #SearchData
There was some real gold in Liz Reid's declaration in Google's appeal of the DOJ vs Google judgment.
There are some nuggets here about user data as it is used to train GLUE (a form of Navboost) and RankEmbed BERT.
Wrote it all up in a new blog post:
www.mariehaynes.com/user-data-in...
There are some nuggets here about user data as it is used to train GLUE (a form of Navboost) and RankEmbed BERT.
Wrote it all up in a new blog post:
www.mariehaynes.com/user-data-in...
User Data is Important in Google's Ranking Systems. What We Learned from Liz Reid's Appeal Declaration. - Marie Haynes
I found some interesting things from the latest document in the DOJ vs Google trial. Google has appealed the ruling that says they need to give proprietary information to competitors. Key takeaways: G...
www.mariehaynes.com
January 22, 2026 at 6:55 PM
There was some real gold in Liz Reid's declaration in Google's appeal of the DOJ vs Google judgment.
There are some nuggets here about user data as it is used to train GLUE (a form of Navboost) and RankEmbed BERT.
Wrote it all up in a new blog post:
www.mariehaynes.com/user-data-in...
There are some nuggets here about user data as it is used to train GLUE (a form of Navboost) and RankEmbed BERT.
Wrote it all up in a new blog post:
www.mariehaynes.com/user-data-in...
“Dot Product is key for Signal Fusion (RankEmbed): Google's RankEmbed system uses dot product (not just cosine similarity) for query-document matching, allowing it to fuse semantic similarity with the magnitude of embeddings, which can encode ‘hand-crafted’ signals like PageRank and quality”
🧵 2/5
🧵 2/5
July 22, 2025 at 7:32 PM
“Dot Product is key for Signal Fusion (RankEmbed): Google's RankEmbed system uses dot product (not just cosine similarity) for query-document matching, allowing it to fuse semantic similarity with the magnitude of embeddings, which can encode ‘hand-crafted’ signals like PageRank and quality”
🧵 2/5
🧵 2/5
“The data underlying RankEmbed models is a combination of click-and-query data and scoring of web pages by human raters.”
www.hobo-web.co.uk
September 7, 2025 at 1:08 PM
“The data underlying RankEmbed models is a combination of click-and-query data and scoring of web pages by human raters.”
Simply put, cosine similarity tells you if terms or content are semantically related.
Dot product, used by Google’s RankEmbed, includes extra signals beyond semantics, such as authority or relevance, captured through vector magnitude.
🧵 4/5
Dot product, used by Google’s RankEmbed, includes extra signals beyond semantics, such as authority or relevance, captured through vector magnitude.
🧵 4/5
July 22, 2025 at 7:32 PM
Simply put, cosine similarity tells you if terms or content are semantically related.
Dot product, used by Google’s RankEmbed, includes extra signals beyond semantics, such as authority or relevance, captured through vector magnitude.
🧵 4/5
Dot product, used by Google’s RankEmbed, includes extra signals beyond semantics, such as authority or relevance, captured through vector magnitude.
🧵 4/5
DOJ odhaluje: Google hodnotí obsah:
Stránka = dokument, signály = faktory hodnocení (RankEmbed s LLM).
Rodina signálů: 100+ raw signálů, které tvoří top-level signály.
Kvalita: Měřeno odkazy a dwell time.
Hybridní přístup (tradiční + AI) pro lepší vyhledávání.
searchengineland.com/google-abc-r...
Stránka = dokument, signály = faktory hodnocení (RankEmbed s LLM).
Rodina signálů: 100+ raw signálů, které tvoří top-level signály.
Kvalita: Měřeno odkazy a dwell time.
Hybridní přístup (tradiční + AI) pro lepší vyhledávání.
searchengineland.com/google-abc-r...
June 19, 2025 at 7:25 AM
DOJ odhaluje: Google hodnotí obsah:
Stránka = dokument, signály = faktory hodnocení (RankEmbed s LLM).
Rodina signálů: 100+ raw signálů, které tvoří top-level signály.
Kvalita: Měřeno odkazy a dwell time.
Hybridní přístup (tradiční + AI) pro lepší vyhledávání.
searchengineland.com/google-abc-r...
Stránka = dokument, signály = faktory hodnocení (RankEmbed s LLM).
Rodina signálů: 100+ raw signálů, které tvoří top-level signály.
Kvalita: Měřeno odkazy a dwell time.
Hybridní přístup (tradiční + AI) pro lepší vyhledávání.
searchengineland.com/google-abc-r...
👩⚖️ @btabke.bsky.social reports Google court docs show it’s rethinking its search stack with LLMs, per the DOJ trial 👇
www.searchengineworld.com/fresh-doc-du...
www.searchengineworld.com/fresh-doc-du...
Fresh Doc Dump: PageRank, Navboost, and RankEmbed: DOJ Papers Outline Google’s Stack
Google’s own court-filed slide deck just spilled more tea on Navboost, and the details punch a few holes in years of SEO guesswork. The newly released trove goes further, mapping more than 100 raw…
www.searchengineworld.com
May 16, 2025 at 4:02 PM
👩⚖️ @btabke.bsky.social reports Google court docs show it’s rethinking its search stack with LLMs, per the DOJ trial 👇
www.searchengineworld.com/fresh-doc-du...
www.searchengineworld.com/fresh-doc-du...
"Newly released documents from the DOJ’s antitrust case reveal that Google’s core ranking system still relies on numerous hand-crafted signals, not just black-box AI."
I doubt if they would make it 100% AI-driven, for it could be too risky for reversal when warranted.
I doubt if they would make it 100% AI-driven, for it could be too risky for reversal when warranted.
May 13, 2025 at 5:02 AM
"Newly released documents from the DOJ’s antitrust case reveal that Google’s core ranking system still relies on numerous hand-crafted signals, not just black-box AI."
I doubt if they would make it 100% AI-driven, for it could be too risky for reversal when warranted.
I doubt if they would make it 100% AI-driven, for it could be too risky for reversal when warranted.
Organizing content for AI search: A 3-level framework
Google's FastSearch and RankEmbed reveal why clarity drives AI visibility. Here's how to strengthen the signals that count.
#google #seo
Google's FastSearch and RankEmbed reveal why clarity drives AI visibility. Here's how to strengthen the signals that count.
#google #seo
Organizing content for AI search: A 3-level framework
Google's FastSearch and RankEmbed reveal why clarity drives AI visibility. Here's how to strengthen the signals that count.
searchengineland.com
October 1, 2025 at 8:38 PM
The ABCs of Google ranking signals: What top search engineers revealed
New DOJ exhibits reveal insights into how Google Search ranks content, Navboost, RankEmbed, and LLMs reshaping the future of search.
#google #seo
New DOJ exhibits reveal insights into how Google Search ranks content, Navboost, RankEmbed, and LLMs reshaping the future of search.
#google #seo
The ABCs of Google ranking signals: What top search engineers revealed
New DOJ exhibits reveal insights into how Google Search ranks content, Navboost, RankEmbed, and LLMs reshaping the future of search.
searchengineland.com
May 14, 2025 at 12:41 AM
Full story 👇
www.searchengineworld.com/fresh-doc-du...
www.searchengineworld.com/fresh-doc-du...
Fresh Doc Dump: PageRank, Navboost, and RankEmbed: DOJ Papers Outline Google’s Stack
Google’s own court-filed slide deck just spilled more tea on Navboost, and the details punch a few holes in years of SEO guesswork. The newly released trove goes further, mapping more than 100 raw…
www.searchengineworld.com
May 14, 2025 at 5:02 PM
Full story 👇
www.searchengineworld.com/fresh-doc-du...
www.searchengineworld.com/fresh-doc-du...
3️⃣ RankEmbed, an LLM-trained signal, enhances query understanding but struggles with long-tail searches, complementing traditional signals.
May 14, 2025 at 5:02 PM
3️⃣ RankEmbed, an LLM-trained signal, enhances query understanding but struggles with long-tail searches, complementing traditional signals.
The ABCs of Google ranking signals: What top search engineers revealed
New DOJ exhibits reveal insights into how Google Search ranks content, Navboost, RankEmbed, and LLMs reshaping the future of search. #LLMS #SEO #Google zurl.co/o631P
New DOJ exhibits reveal insights into how Google Search ranks content, Navboost, RankEmbed, and LLMs reshaping the future of search. #LLMS #SEO #Google zurl.co/o631P
The ABCs of Google ranking signals: What top search engineers revealed
New DOJ exhibits reveal insights into how Google Search ranks content, Navboost, RankEmbed, and LLMs reshaping the future of search.
searchengineland.com
May 27, 2025 at 3:00 PM
The ABCs of Google ranking signals: What top search engineers revealed
New DOJ exhibits reveal insights into how Google Search ranks content, Navboost, RankEmbed, and LLMs reshaping the future of search. #LLMS #SEO #Google zurl.co/o631P
New DOJ exhibits reveal insights into how Google Search ranks content, Navboost, RankEmbed, and LLMs reshaping the future of search. #LLMS #SEO #Google zurl.co/o631P
4️⃣ A lof of familiar models within SEO (RankBrain, DeepRank, RankEmbed BERT) are being absorbed into newer, less-intensive systems (or their functions are)
5️⃣ Core IR fundamentals are still at the heart of search, so go follow Dawn Anderson at least!
🧵 4/6
5️⃣ Core IR fundamentals are still at the heart of search, so go follow Dawn Anderson at least!
🧵 4/6
Deconstructing Google's AI Search: Insights from Antitrust Files & Expert Analysis on MAGIT, Query Fan-Out, and the Reality of Internal Systems
Curious how Google's AI Mode really works, or what powers AI Overviews? This deep dive into the Google Antitrust trial documents reveals: Google's "Blueprint" Revealed: For the first time, a detailed…
www.linkedin.com
July 8, 2025 at 6:27 PM
4️⃣ A lof of familiar models within SEO (RankBrain, DeepRank, RankEmbed BERT) are being absorbed into newer, less-intensive systems (or their functions are)
5️⃣ Core IR fundamentals are still at the heart of search, so go follow Dawn Anderson at least!
🧵 4/6
5️⃣ Core IR fundamentals are still at the heart of search, so go follow Dawn Anderson at least!
🧵 4/6
⚡RankEmbed
Este modelo combina consultas y documentos en un “espacio de embedding”, y calcula distancias semánticas.
Rápido y eficaz, aunque brilla más con consultas comunes que con específicas.
Este modelo combina consultas y documentos en un “espacio de embedding”, y calcula distancias semánticas.
Rápido y eficaz, aunque brilla más con consultas comunes que con específicas.
May 28, 2025 at 7:02 PM
⚡RankEmbed
Este modelo combina consultas y documentos en un “espacio de embedding”, y calcula distancias semánticas.
Rápido y eficaz, aunque brilla más con consultas comunes que con específicas.
Este modelo combina consultas y documentos en un “espacio de embedding”, y calcula distancias semánticas.
Rápido y eficaz, aunque brilla más con consultas comunes que con específicas.
🤖 Señales de ranking básicas:
Google usa más de 100 señales raw (individuales) y señales top-level (combinaciones). Entre las más conocidas están:
✨PageRank
✨Navboost
🌟Su señal estrella entrenada con IA, RankEmbed.
Google usa más de 100 señales raw (individuales) y señales top-level (combinaciones). Entre las más conocidas están:
✨PageRank
✨Navboost
🌟Su señal estrella entrenada con IA, RankEmbed.
May 28, 2025 at 7:02 PM
🤖 Señales de ranking básicas:
Google usa más de 100 señales raw (individuales) y señales top-level (combinaciones). Entre las más conocidas están:
✨PageRank
✨Navboost
🌟Su señal estrella entrenada con IA, RankEmbed.
Google usa más de 100 señales raw (individuales) y señales top-level (combinaciones). Entre las más conocidas están:
✨PageRank
✨Navboost
🌟Su señal estrella entrenada con IA, RankEmbed.