Skip to content

How Google Ranks Pages According to Patents Vintage

This guide synthesizes Bill Slawski's patent analyses into a structured overview of the ranking signals Google has patented. This is not speculation — these are signals described in granted patents and patent applications filed by Google engineers.

The Ranking Signal Taxonomy

SOP: How Google Ranks a Page

Step 1: Query Processing

When a user submits a query, Google first processes the query itself:

  1. Spelling correction — Fix obvious typos
  2. Query classification — Determine intent (informational, navigational, transactional, local)
  3. Entity recognition — Identify known entities in the query
  4. Query expansion — Add synonyms, related terms
  5. Data source selection — Route to appropriate index (web, news, images, local, shopping)

Step 2: Initial Retrieval

Google retrieves candidate documents from its index:

  1. Inverted index lookup — Find all documents containing query terms
  2. Phrase matching — Identify documents with query phrases (not just individual words)
  3. Semantic matching — Find documents matching query meaning via context vectors
  4. Entity matching — Find documents about the entity referenced in the query

Step 3: Scoring

Each candidate document is scored on multiple dimensions:

Step 4: Reranking

After initial scoring, results may be reranked based on:

  • Diversity — Ensuring variety in results for ambiguous queries
  • Personalization — Adjusting based on user history
  • Freshness boost — Promoting recent content for time-sensitive queries
  • De-duplication — Removing substantially similar results
  • Blending — Inserting images, videos, news, maps into web results

Major Ranking Signal Categories

1. Content Signals

SignalDescriptionPatent Source
Keywords in titlesTitle tag relevance to queryRanking Signals (2018)
Keywords in headingsH1-H6 heading relevanceRanking Signals (2018)
Related phrasesCo-occurring terms that establish topical coveragePhrase-Based Indexing
Content depthComprehensive coverage of a topicQuality Score Patent
Context termsWords that establish the context/meaning of contentContext Vectors (2016)
N-gram language modelsNatural language patterns in contentRanking Signals (2018)
SignalDescriptionPatent Source
Link authority (PageRank)Quality and quantity of inbound linksPageRank Patent
Anchor text relevanceDescriptive link text matching queryAnchor Text Patents
Link positionWhere the link appears on the pageReasonable Surfer
Link contextSurrounding text of the linkReasonable Surfer
Link velocityRate of link acquisition over timeHistorical Data Patent
Co-citationBeing linked alongside authoritative sourcesCo-Citation (2012)

3. User Behavior Signals

SignalDescriptionPatent Source
Click-through ratePercentage of searchers clicking a resultUser Behavior (2011)
Dwell timeTime spent on page after clickingUser Behavior (2011)
Pogo-stickingReturning to search results quicklyUser Behavior (2011)
Query refinementSearching again after visiting a resultUser Behavior (2011)
Page watch timeTime spent consuming page contentRanking Signals (2018)
Site blockingUsers choosing to block a sitePanda Integration

4. Quality Signals

SignalDescriptionPatent Source
Quality scoreAlgorithmic quality assessmentPanda Patent (2011)
Content uniquenessNon-duplicate, original contentQuality Score Patent
Ad-to-content ratioAmount of advertising vs. useful contentQuality Score Patent
Template detectionIdentifying auto-generated contentQuality Score Patent
Expert evaluation alignmentMatching quality rater assessmentsPanda Patent

5. Freshness Signals

SignalDescriptionPatent Source
Document ageWhen the content was first publishedHistorical Data Patent
Content update frequencyHow often content is modifiedHistorical Data Patent
Link freshnessWhen links to the page were createdHistorical Data Patent
Query deserves freshnessWhether the query expects recent resultsQuery Classification

6. Entity Signals

SignalDescriptionPatent Source
Entity recognitionContent about known entitiesKnowledge Graph Patents
Entity completenessCoverage of entity attributesEntity Patents
Entity relationshipsConnections to related entitiesCo-occurrence Analysis
Structured dataSchema markup providing entity dataStructured Data Patents

Deep Dive Pages

Key Takeaways

  1. No single signal dominates — Ranking is always a combination of multiple weighted signals
  2. Signal weights change per query — A news query weights freshness heavily; a research query weights authority
  3. User behavior is a real feedback loop — Click patterns and dwell time do influence rankings
  4. Quality is algorithmically measured — Google has specific, trainable systems for scoring content quality
  5. Entities complement keywords — Content that maps to known entities gets additional ranking opportunities
  6. History creates context — Your site's historical patterns form a baseline against which current signals are evaluated

A tribute to Bill Slawski (1958-2022) — the foremost authority on search engine patent analysis.