User Behavior Signals Vintage
Whether Google uses user behavior data to influence rankings has been one of the most debated topics in SEO. Bill Slawski cut through the debate by analyzing the actual patents that describe these systems, documenting exactly what user behavior signals are captured and how they may be used.
User Behavior Data Collection
Click-Through Feedback
Google tracks which results users click and uses this data as a ranking signal.
Click Signal Analysis
| Signal | What It Indicates | Likely Effect |
|---|---|---|
| High CTR | Result title/snippet appeals to users | Positive ranking signal |
| Low CTR | Result not attractive for this query | May trigger snippet change |
| Position-normalized CTR | CTR relative to expected CTR for that position | Truer quality signal |
| Click skip patterns | Users skipping a result to click a lower one | Negative signal for skipped result |
Long Clicks vs. Short Clicks
Pogo-Sticking Detection
Pogo-sticking occurs when a user:
- Clicks a result
- Quickly returns to the search results
- Clicks a different result
This pattern suggests the first result did not satisfy the user's query, and the patent describes this as a negative signal for the first result.
Source: User Behavior Data Google May Use to Influence Search Rankings (2011)
Query Sequences and Refinement
Google analyzes what users search for after their initial query.
Query Refinement Patterns
What Query Refinement Signals
| Pattern | Interpretation |
|---|---|
| Adding specificity | Original results were too broad |
| Changing topic | User's need was satisfied, moved on |
| Similar but different query | Original results did not satisfy intent |
| No refinement | User was satisfied with results |
| Adding location | User wanted local results |
| Adding time period | User wanted recent results |
Search Quality Measurement
Patents describe how Google uses user behavior data to evaluate overall search quality, not just individual result ranking.
Quality Evaluation Metrics from Patents
Using Behavior for Content Classification
The 2006 patent on using queries to detect adult content showed how user behavior can classify content:
- If users who search for adult terms consistently click on a page, that page gets classified accordingly
- The same principle applies to other content categories
- User behavior serves as implicit classification feedback
Source: Ask.com Using Queries to Detect and Filter Adult Content (2006)
User Feedback Integration
Google's 2011 announcement about incorporating user feedback was analyzed by Bill in the context of existing patents.
Types of Explicit User Feedback
| Feedback Type | How It Is Used |
|---|---|
| Site blocking | Users blocking a site from their results inform quality scoring |
| Spam reporting | User reports feed into spam detection systems |
| Result quality feedback | Thumbs up/down data on search results |
| Knowledge Panel edits | Suggested corrections to entity data |
Implicit vs. Explicit Feedback
Bot vs. Human Traffic Detection
Patents describe systems for distinguishing between genuine user behavior and bot activity:
Detection Methods
| Method | What It Catches |
|---|---|
| Behavioral patterns | Click speeds, browsing patterns that are too regular |
| Session analysis | Non-human session characteristics |
| JavaScript execution | Bots that do not execute JavaScript |
| Mouse movement | Automated vs. natural mouse patterns |
| Request patterns | Unusual request frequencies or sequences |
Bot traffic is filtered from behavioral signals so that click data used for ranking reflects genuine user interest.
Personalization Through Behavior
User behavior data is also used for search personalization:
Personalization Factors from Behavior
Source: How to Personalize Web Search (2006)
Key Takeaways
- User behavior signals are real — Patents explicitly describe systems that use click data, dwell time, and query sequences.
- Long clicks are better than short clicks — Time spent on a page after clicking is a quality indicator.
- Pogo-sticking is a negative signal — Users quickly returning to search results suggests the result was unsatisfying.
- Query refinement patterns inform Google — What users search next after your query reveals whether results satisfied intent.
- Bot traffic is filtered — Only genuine user behavior influences ranking signals.
- Behavior drives personalization — Individual user patterns shape future search results for that user.