Module 13: User Behavior, Clicks, and Personalization Patents

111 Patents

This module covers Google's comprehensive patent portfolio for user behavior signals, including click-through rate, dwell time, engagement signals, personalization, and user satisfaction prediction. These patents reveal how Google uses user interactions to refine search rankings.

Overview

User behavior is one of the "Three Pillars" of Google ranking (confirmed in DOJ antitrust case). This module covers:

• Click-Through Rate & Click Data - How clicks influence ranking
• User Engagement Signals - Shares, comments, and interactions
• Dwell Time & Session Data - Time spent on content
• Personalized Search - Tailoring results to users
• User Intent Modeling - Understanding what users want
• Behavioral Signals for Ranking - How behavior affects positions
• A/B Testing - How Google tests ranking changes
• User Satisfaction Prediction - Measuring search quality

---

Part 1: Click-Through Rate & Click Data (20 Patents)

CTR is a fundamental ranking signal in Google's systems.

Patent	Title	Key Innovation
US10229166B1	Modifying ranking based on implicit behavior	Click data re-ranking
US20130339350A1	Ranking Based on CTR	CTR-based ranking
US8082246B2	Ranking using click distance property	Click distance metrics
US8843477B1	Onsite and offsite search ranking	Website click data
US9684697B1	Ranking based on click data	Repeat click counts
US8959093B1	Ranking based on anchors	Click data quality stats
US9009146B1	Ranking based on similar queries	Click data as quality
US8938463B1	Modifying ranking based on implicit behavior	Weighted click data
US8117209B1	Ranking based on user behavior/feature	User behavior data
US9092510B1	Modifying ranking based on implicit behavior	Return behavior analysis
US8250065B1	Ranking based on click throughs	Click-through counting
US8818977B1	Context sensitive ranking	Contextual click models
US8676790B1	Ranking using selection data	Selection/click scoring
US8930350B1	Autocompletion using query data	Processed click data
US20050021397A1	Content-targeted advertising	Click/conversion feedback
US8938787B2	Detecting identity of user	Click behavior patterns
US20080162475A1	Click-fraud detection method	Click pattern monitoring
US8880441B1	Click stream analysis for fraud	Click pattern analysis
US7933984B1	Detecting click spam	Click spamming detection
US7788260B2	Ranking based on click frequency	Click frequency metrics

Click-Through Rate Processing

NavBoost (DOJ Confirmed)

The DOJ antitrust case confirmed Google uses a system called NavBoost that processes click data:

---

Part 2: User Engagement Signals (14 Patents)

Engagement beyond clicks - shares, comments, and interactions.

Patent	Title	Key Innovation
US10296642B1	Ranking content for engagement	Engagement scoring
US20110208585A1	Measurement of Engagement	Engagement score computation
US11245966B2	Matching and ranking content	Engagement for ranking
US20170193544A1	Modification by engagement	Engagement level adjustment
US8626823B2	Page ranking using sharing	User sharing signal
US7904303B2	Engagement-oriented recommendation	Engagement predictions
US8965883B2	Ranking user generated content	User credential scores
US10936986B2	Engagement recommendations	Creator engagement
US10049138B1	Reputation and engagement system	Community engagement
US8972391B1	Recent interest based scoring	Interest-based scoring
US20240311558A1	Comment section analysis	Viewer engagement indicators
US10084732B1	Ranking social connections	Social interaction metrics
US9773256B1	User-based ad ranking	Aggregate ad performance
US11065542B2	User engagement in games	Remote engagement data

Engagement Score Components

---

Part 3: Dwell Time & Session Data (10 Patents)

Time on page and session behavior are critical ranking signals.

Patent	Title	Key Innovation
US8255413B2	Responding to queries	Dwell time metrics
US8601023B2	Identifying documents	User activity over time
US9558233B1	Quality measure for resource	Dwell time quality score
US8838587B1	Propagating query classifications	Session data with dwell
US8959093B1	Ranking based on anchors	Document dwell tracking
US9092510B1	Modifying ranking based on behavior	Time-based weighting
US7752201B2	Recommendation based on sessions	Session duration pairing
US10885039B2	ML based search improvement	Session timing
US11170059B2	Personalized content for sessions	Session-based personalization
US7729940B2	Analyzing advertising ROI	Minimum dwell filter

Dwell Time Impact on Ranking

Session Analysis

Google analyzes entire search sessions, not just individual clicks:

1. Query Reformulation - Many reformulations = low satisfaction
2. Pogo-Sticking - Clicking back quickly = poor result
3. Task Completion - No return to search = task completed
4. Session Duration - Longer sessions on a topic = high engagement

---

Part 4: Personalized Search & Recommendations (25 Patents)

How Google tailors search results to individual users.

Patent	Title	Key Innovation
US20060064411A1	Search engine using user intent	User behavior compilation
US11263278B2	Triggering personalized queries	User profile queries
EP2050020A1	Personalized search indexing	Custom indexing
US8321278B2	Targeted ads based on profile	User + page profiles
US9449105B1	User-context-based search	Context determination
US7716225B1	Ranking based on user behavior	User behavior model
US7505964B2	Ranking using related queries	Related query data
US7113917B2	Personalized recommendations	Similar user ID
US20110066497A1	Personalized advertising	User GUID assignment
US11392993B2	Personalized recommendations	Algorithm selection
US9483778B2	Generating a user profile	Dynamic profile compression
US11164105B2	Intelligent recommendations	Deep learning multimodal
US9842358B1	Personalized recommendations	Profile comparison
US12380115B1	Providing recommendations	Adaptive recommendations
US11281734B2	Personalized recommender	Limited data handling
US9792366B2	Content recommendation	Third party profiles
US9251527B2	Personalized recommendations	Population preferences
US6853982B2	Content personalization	Action-based customization
US8032506B1	User-directed recommendations	Profile persistence
US12293401B2	Personalized context-aware recs	Explanation generation
US7908183B2	Recommendation system	Ratings-based profiles
US7593921B2	User profile and preferences	Feature-based profiles
US20050131762A1	User info for targeted ads	Profile inference
US20120233142A1	Personalization using term profiles	Term-based personalization
US20080208705A1	Personalized shopping assistant	Purchase recommendations

Personalization Architecture

---

Part 5: User Intent Modeling (18 Patents)

Understanding what users really want from their queries.

Patent	Title	Key Innovation
US8868548B2	User intent from query patterns	Pattern-based intent
US20140207622A1	Intent prediction recommendation	Intent identification
US20220051665A1	AI-based intent analyzer	AI intent system
US9465833B2	Disambiguating user intent	Intent disambiguation
US20170075988A1	Automatic query resolution	Intent analysis
US11562028B2	Concept prediction for intents	Intent creation
US11663201B2	Query variants generation	Intent understanding
US11144730B2	End to end dialogue modeling	Intent analyzer
US20230153365A1	User intents and sentiments	Intent + sentiment
US20220107802A1	Context-aided search intent	Intent detection
US10706450B1	Intent-aware search results	Semantic intent
US20200293874A1	Intent with transfer learning	Intent identification
US8843470B2	Meta classifier for intent	Non-linear ensemble
US8918354B2	Intent from social messages	Social intent detection
WO2017143338A1	User intent and context search	Syntactic parsing
EP3005168A1	NL search for intent queries	Intent templates
US20200159790A1	Intent-oriented browsing	ML intent detection
EP3824400A1	Visual intent triggering	Visual intent ML

Intent Classification Categories

---

Part 6: Behavioral Signals for Ranking (12 Patents)

How behavior directly affects ranking positions.

Patent	Title	Key Innovation
US12259895B1	Behavior-driven query similarity	Behavioral signals
US6012053A	User-controlled relevance ranking	User behavior ranking
US8661029B1	Modifying ranking based on behavior	Click weighting
US8818995B1	Ranking based on trust	Trust from behavior
US10394830B1	Sentiment detection as ranking	Sentiment signals
US8762373B1	Personalized result ranking	Past selection activity
US8825644B1	Adjusting ranking of results	Behavior-based adjustment
US6546388B1	Metadata search ranking	Click behavior monitoring
US20121430814A1	Search engine result ranking	R metric for satisfaction
US7693825B2	Ranking implicit search results	Event-based ranking
US7818315B2	Re-ranking based on query log	Query log matching
US8224827B2	Ranking based on classification	Classification ranking

---

Part 7: A/B Testing for Search (7 Patents)

How Google tests changes to search ranking.

Patent	Title	Key Innovation
US20060162071A1	A/B testing	Split testing
US11132700B1	Identifying effects in A/B tests	Direct/indirect effects
US9032282B2	A/B testing for web content	Split testing methodology
US20140278747A1	Stratified sampling A/B tests	Sampling techniques
US11593667B2	A/B testing with sequential hypothesis	Sequential testing
US8296643B1	Multiple experiments on test page	Simultaneous tests
US9906612B2	Long term metrics multivariate	Long-term testing

A/B Testing Framework

---

Part 8: User Satisfaction Prediction (5 Patents)

Measuring whether users are satisfied with search results.

Patent	Title	Key Innovation
US20060224554A1	Query revision using high-ranked queries	Satisfaction from quality
US20140019199A1	Automatically evaluating satisfaction	Satisfaction metrics
US8442984B1	Website quality signal generation	Rating relationships
US20180330001A1	Processing user behavior data	Post-click data
US20120130814A1	Search engine result ranking	Searcher satisfaction R metric

User Satisfaction Indicators

---

Summary Statistics

Category	Patent Count
Click-Through Rate & Click Data	20
User Engagement Signals	14
Dwell Time & Session Data	10
Personalized Search & Recs	25
User Intent Modeling	18
Behavioral Signals for Ranking	12
A/B Testing for Search	7
User Satisfaction Prediction	5
TOTAL	111

---

Key Insights for SEO

1. Click-Through Rate Matters

• CTR is normalized by position (position 1 is expected to get more clicks)
• Abnormally high CTR for lower positions is a strong positive signal
• Compelling title tags and meta descriptions improve CTR

2. Dwell Time is Critical

• Long dwell time indicates content relevance
• Quick returns to SERP (pogo-sticking) are negative signals
• Engaging, comprehensive content keeps users on page

3. Engagement Beyond Clicks

• Shares and social engagement are positive signals
• Comments and interactions indicate quality content
• Multiple page views in a session show value

4. Personalization Impact

• Results vary significantly by user
• Location, search history, and device all matter
• Building repeat visitors helps personalized ranking

5. User Satisfaction

• Google measures overall search satisfaction
• Task completion is the ultimate goal
• Reducing query reformulations is a success metric

---

Actionable Optimization Strategies

Improving CTR

1. Write compelling, accurate title tags
2. Create engaging meta descriptions
3. Use structured data for rich snippets
4. Match search intent in titles

Increasing Dwell Time

1. Create comprehensive, valuable content
2. Use clear navigation and structure
3. Include multimedia (images, videos)
4. Answer the query thoroughly upfront

Boosting Engagement

1. Include social sharing buttons
2. Enable and moderate comments
3. Create shareable content formats
4. Build email lists for return visits

Supporting User Intent

1. Match content to query intent
2. Provide clear answers early
3. Include related topics for exploration
4. Use internal linking strategically

---

Related Modules

• Module 7: User Behavior - Core behavior signals
• Module 11: Neural Networks & AI - ML for ranking
• Module 5: Content Quality - Quality signals

---

Pro Tip

User behavior signals are relative, not absolute. What matters is how your page performs compared to other results for the same query. Focus on being measurably better than competitors in your SERP.