Module 9: Additional Patent Foundations
149+ patents covering trust, crawling, local, duplicates, voice, neural, and 2023-2024 research.
Overview
This module expands the patent foundation with 13 topic areas that extend the core patent library. These patents cover the full breadth of Google's ranking systems from trust signals to voice search to the latest neural ranking innovations.
Trust & Authority Patents
TrustRank (US7603350B1)
Filing Date: 2006 Grant Date: 2012
Key Innovation: Propagates trust from seed sites to linked pages.
Mechanism:
- Starts from known high-quality seed sites
- Trust flows through links (like PageRank but for trust)
- Sites with low trust or spam links receive negative signals
- Trust decays with link distance from seed sites
Key Insight: Good sites don't link to spam. Spam sites do link to spam. TrustRank exploits this asymmetry.
Social TrustRank (US20060294134A1)
Filing Date: 2006
Key Innovations:
- Extends TrustRank into social networks
- Trust propagation through social connections
- Author reputation in social contexts
- Community-based trust signals
Query Deserves Freshness (QDF)
US8549014B2 - Content Freshness Scoring
Year: 2012
When QDF Activates:
- Breaking news and current events
- Trending topics with sudden search volume spikes
- Queries where recent content significantly outperforms older content
- Seasonal and time-sensitive searches
QDF Signals:
1. News spike detection — sudden query volume increase
2. Content publication velocity — how fast are new pages appearing?
3. Social signal velocity — rapid social engagement
4. Temporal query language — "latest", "recent", "2024", "now"Crawling & Indexing Patents
Caffeine Architecture
Year: 2010
Key Innovations:
- Near-real-time indexing (minutes, not days)
- Distributed crawling at scale
- Priority-based crawl queue
- Incremental index updates without full rebuilds
Practical Implication: Quality fresh content can rank within hours of publication.
URL Priority & Crawl Budget
- Pages with more internal links get crawled more frequently
- Server response time affects crawl rate
- Duplicate and thin content wastes crawl budget
- robots.txt directly controls what gets crawled
Local Search Patents
US8463772B1 - Geographic Radius by Business Category
Year: 2013
Key Finding: Different business categories have different effective search radii.
| Category | Effective Radius |
|---|---|
| Emergency services (ER, locksmith) | Very small (0.5-2 miles) |
| Restaurants | Small (1-5 miles) |
| Home services | Medium (10-25 miles) |
| Specialty services | Large (25-50 miles) |
| Online-capable services | Can be regional |
Duplicate Content Patents
US7734627B1 - Document Fingerprinting
Year: 2010
Methods:
- Document fingerprinting for exact duplicate detection
- SimHash for near-duplicate detection (>70% similarity)
- Phrase-based indexing identifies copied phrase clusters
- DOM tree analysis separates boilerplate from unique content
Thresholds:
95% similarity = exact duplicate
- 70-95% similarity = near duplicate (duplicate risk)
- <70% = distinct content (generally safe)
Query Understanding Patents
Anchor Text in Query Understanding
Patents Referenced: US7260573B1, US7308643B1
How Anchor Text Informs Query Understanding:
- Google uses anchor text to understand what a linked page is about
- Accumulated anchor text serves as crowd-sourced page descriptions
- Anchor text from high-authority pages weighted more heavily
- Anchor diversity signals naturalness
Visual Search Patents
| Patent | Innovation |
|---|---|
| US12051205B1 | Multimodal embeddings (text + image combined) |
| US12210825B2 | Image captioning for search indexing |
| Multiple | Visual similarity search |
Voice Search Patents (Overview)
213+ Voice Patents Across 9 Categories
Voice Search Core:
- Speech-to-text query processing
- Acoustic models + language models
- Conversational context tracking
- Intent extraction from speech
Key SEO Implication: Voice queries are longer, more conversational, and question-based. Content should answer natural language questions directly.
Neural Search Patents
BERT Integration (US10452978B2)
Year: 2019
What BERT Changed:
- Full query context processing (not just keywords)
- Understanding prepositions and connective words
- Resolving ambiguous pronouns in queries
- Semantic matching beyond keyword overlap
Before BERT: "how to get rid of flies" → keyword match to "fly" content After BERT: Understands this is about insects, not air travel
Anchor Text & Link Quality
| Signal | Impact |
|---|---|
| Anchor text quality | High |
| Source page authority | High |
| Anchor relevance to target | High |
| Natural anchor diversity | Required |
| Over-optimization | Negative |
Spam Detection Patents
Link Spam (US7533092B2, US7953763B2)
- Identifies spam farms and link schemes
- Analyzes link patterns for spam signals
- PageRank applied to spam detection
- Pattern recognition for artificial link velocity
Semantic & Q&A Patents
US8594996B2 - Entity Recognition and Disambiguation
Year: 2012-2014
Key Innovations:
- Named Entity Recognition (NER) at scale
- Entity disambiguation using context
- Knowledge base integration for entity resolution
- Multi-word entity identification
Re-Ranking Patents
Post-Query Re-Ranking Signals:
- User click data (NavBoost)
- Pogo-sticking signals
- Long-click vs. short-click patterns
- Return-to-SERP behavior
- Session satisfaction metrics
2023-2024 Patents
| Patent | Year | Innovation |
|---|---|---|
| US12353469B1 | 2024 | LLM output verification / source checking |
| 2024 Hypergraph Patent | 2024 | Multi-relationship entity search |
| US20250124067A1 | 2025 | Pairwise ranking with LLM prompting |
| US20250225400A1 | 2025 | LLM training content control |
YouTube & Social Signal Patents
YouTube Ranking Signals:
- Watch time and audience retention
- CTR on thumbnails
- Channel authority and subscriber quality
- Engagement velocity (likes, comments in first 24h)
Social Signal Disclaimer: Social signals are NOT confirmed direct ranking factors. They correlate with ranking due to content quality, not causation.
Key Patents Referenced
| Patent | Title | Year |
|---|---|---|
| US7603350B1 | TrustRank | 2006-2012 |
| US8549014B2 | Content Freshness Scoring | 2012 |
| US7734627B1 | Document Fingerprinting | 2010 |
| US8594996B2 | Entity Recognition & Disambiguation | 2012-2014 |
| US10452978B2 | BERT / Attention Mechanism | 2019 |
| US7533092B2 | Link-Based Spam Detection | 2004-2009 |
| US8463772B1 | Geographic Radius by Category | 2013 |
Next Steps
- Comprehensive Expansion Module — 2024-2025 patents
- Neural AI Search Module — Deep learning patents
- Master Patent Database — Complete index