Chapter 5

Writing Semantic Content — Koray's Framework

50+ rules for creating content that Google's NLP systems can fully understand, extract factual triples from, and rank with high confidence. Complete author guidelines, rule index, and technical on-page elements.

5A. Core Semantic Content Writing Rules

Semantic content writing requires a fundamentally different approach from traditional SEO writing. These rules, derived from Koray Tugberk GUBUR's Semantic SEO framework and expanded by Mubashir Hassan, govern how to write content that Google's NLP systems can fully understand and rank with high confidence.

RULE 01

Include More Entities and Attribute Values

Include more entities and attribute values, as long as they are relevant to the macro context of the page. Every relevant entity mention strengthens the semantic signal.

RULE 02

Single Macro Context Per Page

Each web page should have exactly one macro context — one primary topic that all content on the page serves. Multi-topic pages confuse Google's semantic understanding.

RULE 03

Use Factual Sentence Structures

Use direct, factual structures: 'X does Y' instead of 'X is known for Y'. Change indirect statements to direct action sentences — 'Apple creates innovative products' not 'Apple is known for creating innovative products.'

RULE 04

Use Research and University Studies

Support claims with research papers, university studies, government data, and authoritative sources. Evidence-backed content signals expertise and trustworthiness to both users and Google.

RULE 05

Be Concise — Eliminate Fluff

Replace vague superlatives with specific data. 'SEO is the most common marketing channel' → 'SEO is a marketing channel with 45% of total digital marketing spend in [Year].'

RULE 06

Do Not Break Context Across Paragraphs

Maintain unbroken contextual flow within and between paragraphs. Each paragraph should follow logically from the previous one without abrupt topic changes or confusing transitions.

RULE 07

Optimize Discourse Integration

Structure content so that every sentence, paragraph, and section connects logically and semantically — helping search engines clearly understand how ideas relate, why they matter, and what the content achieves.

RULE 08

Use Question and Answer Format

Opt for Q&A format where possible. Google converts all headings into questions internally, so writing in Q&A format aligns with how Google processes and displays content in featured snippets and PAA.

RULE 09

Do Not Distance the Question from the Answer

Answer the question immediately after posing it, then expand. A definitive answer should appear in the first sentence of each section. Never bury the answer after lengthy preamble.

RULE 10

Expand Evidence with Variations

Provide multiple examples, perspectives, and types of evidence for each information point. Instead of 'Exercise is good for health' → 'Exercise improves heart function, boosts mental well-being, and reduces all-cause mortality risk by 35% (WHO, 2023).'

RULE 11

Do Not Break the Information Graph

Follow a logical order in declarations. Avoid confusing jumps between ideas. Bad: 'Cats are playful. The sky is blue. Cats like milk.' Good: 'Cats are playful animals. They enjoy activities like chasing toys and climbing.'

RULE 12

Use Shorter Sentences

Use shorter sentences as much as possible. Short, clear sentences are easier for NLP systems to parse and extract triples from. Long compound sentences with multiple clauses reduce parsing accuracy.

RULE 13

Decrease Contextless Words

Reduce words that don't add meaning. Original: 'The document is really quite interesting and has a lot of stuff in it.' Improved: 'The document provides detailed insights on [topic].'

RULE 14

Delete All Words That Add Nothing

Delete all words that, if deleted, don't change the meaning of the text. Remove: really, very, quite, just, simply, basically, essentially, generally, typically (when used as filler).

RULE 15

Give More Information Per Section

Maximize information density, especially unique information. More factual density per unit of text signals semantic richness to Google's quality assessment systems.

RULE 16

Create a Proper Context Vector

Build a coherent context vector that runs from H1 through every subsequent heading to the end of the document. Contextual flow must be linear — each heading deepens the context established by the heading above it.

RULE 17

Always Give Consistent Declarations

Never change opinions, statements, or factual claims from web page to web page within the same site. This extends to all first-party properties (videos, social media, GBP). Consistency builds entity trust with Google.

RULE 18

Use Consistent Style Across Documents

Maintain consistent brand voice, terminology, and writing style across all website content. Consistency signals brand identity and helps Google understand your entity as a coherent, reliable source.

RULE 19

Have Unique N-Grams and Phrase Combinations

Create unique N-gram combinations to demonstrate originality. Unique phrases signal that your content adds new information to the web rather than just rephrasing existing content.

RULE 20

Include Same N-Grams at Beginning and End

Including the same N-grams (key phrases) at both the start and end of a web document reflects contextual consistency and strengthens the contextual vector throughout.

RULE 21

Include Multiple Examples, Data and Percentages

Add multiple examples, data sets, and percentages for each information point. Experts use numbers, statistics, and evidence. Specific numeric data improves extraction accuracy for Google's fact-checking systems.

RULE 22

Complete a Single Topic with Every Detail

Even if a detail doesn't appear in queries or competitor documents, include it. Less important inclusions can be added as lower-order headings to signal they are part of the topic but not the primary focus.

RULE 23

Use Fewer Links Per Document

Minimize unnecessary links while maximizing information density. Excessive linking can dilute page authority and distract from the semantic focus of the content.

RULE 24

Do Not Promote Products in Information Pages

Information pages should be unbiased. Mixing sales content into informational pages confuses Google's intent classification and reduces the page's ability to rank for informational queries.

RULE 25

Use Ordered and Unordered Lists

Use both ordered (numbered) and unordered (bulleted) lists appropriately. Lists help Google extract structured information and can trigger list-format featured snippets in SERPs.

RULE 26

Use Definitions, Signifiers and Qualifiers

Entity signifiers help Google identify exactly what entity you are discussing. Qualifiers specify the precise scope of your claims. Together, they make content more machine-readable and extractable.

RULE 27

Key Terms Must Connect to Title and H1

All key terms used must connect to the title and H1, which in turn must connect to the topical map. Context is vital — never break the context governed by the vectors created by your title and primary heading.

RULE 28

Stick to a 40-Word Limit Per Answer

40-word answers for key points help trigger featured snippets. Concise, self-contained answers directly below headings align with how Google selects and displays featured snippets.

RULE 29

Be Certain in Declarations

Use certain declarative language: 'The sun rises every day' not 'The sun usually rises.' Sentences with 'will, should, need to, have to' are less valid for fact extraction than direct declarations.

RULE 30

Cut the Fluff — Delete Contextless Words

Every word you use must have contextual relevance. Delete any word that, if removed, does not change the meaning — adverbs as modifiers, hedging language, redundant adjectives, and filler transitions.

RULE 31

Use Numeric Values — Tell the Exact Number

Experts are specific. Replace vague quantities with exact numbers: Not 'many users prefer...' but '67% of users prefer...' (Source: [Organization], [Year]).

RULE 32

Qualify the Instances

Numbers alone are sometimes insufficient. Specify the exact nature of items in lists — 'There are 6 severe symptoms of X disease' and separately 'There are 9 rare symptoms of X disease.' This precision improves semantic extraction.

RULE 33

Understand Context of Verbs — Choose Predicates Wisely

Determine predicates wisely. In Semantic SEO, Google uses verbs (predicates) to understand intent, action, and relationships between entities. The choice of verb directly affects how Google categorizes the relationship expressed.

RULE 34

Give Examples After a Plural Noun

After stating a plural category, immediately provide concrete examples: 'There are X different cryptocurrencies to trade on Coinbase, including Bitcoin, Ethereum, and Dogecoin.' This enables Google to extract entity-attribute relationships.

RULE 35

Use Same Part of Speech in Listings

When writing lists, every point should start with the same part of speech. All verbs: 'Enhance clarity, Improve accuracy, Ensure efficiency.' Not mixed: 'Enhance clarity, Accurate, For efficiency.' Consistent POS tags make lists semantically coherent.

RULE 36

Optimize Subordinate Text First Sentence

The opening sentence of any subordinate text should immediately convey its purpose or relevance. Before: 'This section explains the benefits of SEO.' After: 'Understanding SEO can dramatically increase website traffic and visibility by 47% within 6 months.'

RULE 37

Put 'If' Statements in the Second Part

Conditional statements should place the condition after the main clause: 'SEO drives more organic traffic if implemented with a semantic content strategy.' This ensures the primary information appears first.

RULE 38

Bold the Answer, Not the Search Term

When using bold text for emphasis, bold the answer or key fact — not the search keyword. Bold signals importance to both users and Google's NLP systems. Use it to highlight extractable facts, not to stuff keywords.

5B. Complete Author Rule Index

The following is the complete index of semantic content writing rules for authors, as catalogued in the Semantic SEO Playbook. Use this as a checklist when auditing or creating semantic content.

Use Entities, Attributes & Popular N-Grams
Focus on Entities, not Keywords
Single Macro Context per Webpage
Factual Sentence Usage
Perspective Richness
Fake vs Genuine Originality
Using Research Papers to Prove Points
Maintain Context
Discourse Integration Optimization
Topic in Q&A Format
Timely Answer Delivery
Expand Evidence with Variations
Maintain Information Graph
Use Shorter Sentences
Reduce Contextless Words
More Information per Page/Section
Context Vector Hierarchy
Consistent Document Style
Consistent Declarations
Unique N-Grams
Same N-Grams at Start & End
Single Topic with Every Detail
Fewer Links
Avoid Product Promotion in Info Articles
Use Ordered & Unordered Lists
Long Form Answering
Key Term in Title & Heading
40-Word Featured Snippet
Avoid Opinion in Articles
Avoid Everyday Language
Avoid Analogies
Avoid Entity Stuffing
Avoid Coreference Error
Match Tenses
Modality
Citing Authoritative Sources
Content Length Rules
Avoid Unnecessary Sentences
Enhance Paragraph Perspective
Avoid Copy Pasting Questions
Avoid Linking to Citations
Table Context
Use Abbreviations
Avoid Confusing Users or Bots
Provide Safe Answers
Signal Answer Section
Match Anchor Text with Target Page Title
Placement of 'if' in Second Statement
Optimize Subordinate Text First Sentence
Give Examples After Plural Nouns
Choose Predicates Wisely
Be Specific in Descriptions
Use Numeric Values
Cut the Fluff Out
Be Certain
Consistent Part of Speech in Listings
Prioritize Attributes & Contexts
Use Diverse Measurement Units
Answer Boolean Questions
Avoid Uncertain Words
Grammar & Spelling Accuracy
Google Fact Verification
Truth Ranges
Optimize Content for NLP Relevance
Answering Short vs Long Form
Importance of TOC in Ranking
Parts of Speech

5C. Technical On-Page Semantic Elements

Technical on-page elements are the structural signals that communicate semantic meaning to Google beyond the content itself. These elements must be optimized consistently with the content's entity and contextual strategy.

Title Tag — Semantic Rules

Include the central entity, primary attribute, and context modifier. Follow the formula: [Entity] + [Attribute] + [Context/Location/Modifier]. Avoid clickbait; be descriptive and specific.

Meta Description — Semantic Rules

Include the main entity, one supporting entity, the key benefit/outcome, and a natural call-to-action. Keep within 155 characters. Match the macro context of the page.

H1 Optimization

The H1 defines the macro context. It must include the central entity and its most important attribute. Must differ slightly from the title tag while maintaining the same macro context.

URL as Root, Seed, Node

URLs should reflect the topical map hierarchy: Root (domain/central entity) → Seed (primary subtopic) → Node (specific topic). Clean, entity-based slugs with hyphens, no stop words, no keyword stuffing.

Internal Linking — Semantic Rules

Internal links should use anchor text that describes the relationship between source and target pages. Contextual anchors communicate entity connections. Link from related entities to target entities with descriptive relationship anchors.

Image SEO

Image file names should include entity names and attributes. Alt text should describe the image content using entity-attribute language, not just keyword stuffing. Images should reinforce the entity context of the page.

LSI Keywords Myth

Latent Semantic Indexing (LSI) keywords are a myth in modern SEO — Google does not use LSI technology. Focus on entity coverage, attribute completeness, and contextual relevance instead.

Anchor Text Rules

Anchor text should match the target page's title or primary entity — not a generic keyword. Partial match anchors that describe the relationship are preferable to exact-match anchors. Vary anchor text naturally across multiple links to the same page.

Article Posting Momentum

Maintain consistent publication momentum over time. Google rewards sites that publish quality content at a sustainable, predictable pace. Sudden bursts of mass publication followed by silence disrupt topical authority signals.

Semantic HTML

Use semantic HTML elements (article, section, aside, header, footer, main, nav) to reinforce the structural meaning of content. Semantic HTML helps Google's parser understand the role and hierarchy of different content sections.

Connected Schema Markup

Apply schema markup that connects entities across the page: Article schema, BreadcrumbList, Author schema, Organization schema, FAQ schema. Connected schema creates a knowledge graph within the page that reinforces entity relationships.

Brand Schema Markup

Define your brand as an entity using Organization/Brand schema with: name, URL, logo, sameAs (social profiles), description, and contactPoint. Brand schema helps establish your entity in Google's Knowledge Graph.

Understanding EVA Model

Entity-Attribute-Value is the fundamental unit of semantic content: Entity (what you're describing), Attribute (the property you're describing), Value (the specific data about that property). Every factual sentence should express an EVA relationship.

What is a Contextual Layer?

A semantic level of meaning added to content through entity mentions, attribute specifications, factual triples, or schema markup. Content with multiple contextual layers provides richer semantic signals than surface-level keyword content.

Reviving Expired Domains

Expired domains with existing topical authority can accelerate new site authority building. The domain's historical topic must match your new site's topic, and the topical coverage must be relevant and substantial.

★ Chapter 5 Summary
  • Every sentence should serve the semantic purpose of the document — eliminating fluff maximizes NLP extraction accuracy.
  • Consistent declarations, entity usage, and writing style across all pages build entity trust with Google's systems.
  • Technical elements (title, H1, URL, schema, internal links) must be semantically aligned with the content's entity and contextual strategy.
  • The EVA model (Entity-Attribute-Value) is the foundational unit — every factual claim should be expressible as a structured triple.