Working copy: this standard is maintained in aaron-marketing-skills —
references/core-eeat-benchmark.mdand published here as its citable home. Framework changes land there first, then sync here (scripts/sync-family.sh).📦 Merged into aaron-marketing-skills — CORE-EEAT now ships inside the unified marketing skills library (v10.0.0) alongside the CITE and C³ frameworks and 38 marketing skills. This repo remains as the standalone source.
8 dimensions × 10 items = 80 evaluation criteria for optimizing content visibility across AI engines (GEO) and search engines (SEO).
Version: 1.1 | Author: Aaron | Updated: 2026-02-10
Sister Project: CITE Domain Rating — domain-level authority assessment (40 items)
This document uses progressive disclosure — read as deep as you need:
| Depth | Sections | You Will Learn |
|---|---|---|
| Skim | Parts 1–2 | What CORE-EEAT is + all 80 check items at a glance |
| Apply | Parts 3–5 | How to score content, adapt to content types, and run quality gates |
| Master | Parts 6–7 | Detailed Pass/Partial/Fail criteria for every item with examples |
| Optimize | Part 8 | AI engine preferences, implementation mapping, common errors |
CORE-EEAT is a content quality benchmark for the AI era. It evaluates content across two complementary systems:
| System | Optimization Target | Core Question | Items |
|---|---|---|---|
| CORE | GEO (Generative Engine Optimization) | Can AI understand, extract, and cite this content? | 40 |
| EEAT | SEO (Search Engine Optimization) | Is the source behind this content trustworthy? | 40 |
| Total | GEO + SEO | — | 80 |
The two systems are mutually exclusive and collectively exhaustive:
Quick rule: If you can see it in the article/page → CORE. If you need to check the author/org/site → EEAT.
CORE-EEAT evaluates content; its sister project CITE Domain Rating evaluates the domain behind the content. Together they form a complete 120-item assessment:
| Benchmark | Evaluates | Level | Items | Core Question |
|---|---|---|---|---|
| CORE-EEAT | Content quality | Single page/article | 80 | Is this content worth citing? |
| CITE | Domain authority | Entire domain | 40 | Is this domain worth trusting as a source? |
| Combined | Full assessment | Content + Domain | 120 | Should AI engines cite this source? |
| System | Dim | Full Name | Core Question |
|---|---|---|---|
| CORE | C | Contextual Clarity | Does content match intent? Is the answer upfront? |
| CORE | O | Organization | Is structure clear? Can AI parse it efficiently? |
| CORE | R | Referenceability | Can data and evidence be verified and cited? |
| CORE | E | Exclusivity | What unique value does this add? Why cite you over others? |
| EEAT | Exp | Experience | Does the creator have first-hand experience? |
| EEAT | Ept | Expertise | Does the creator have professional competence? |
| EEAT | A | Authority | Do third parties recognize this source? |
| EEAT | T | Trust | Is the site technically, legally, and governance-wise trustworthy? |
Each check item carries a priority tag indicating its primary optimization target:
| Tag | Meaning | When to Prioritize |
|---|---|---|
| GEO-First 🎯 | Critical for AI engine citation | Content targeting AI search visibility |
| SEO-First 🔍 | Critical for traditional search ranking | Content targeting Google/Bing organic traffic |
| Dual ⚡ | Important for both channels | All content should aim for these |
One-line standard for every check item. Use this as your evaluation reference.
| ID | Dim | Check Item | Priority | One-Line Standard |
|---|---|---|---|---|
| C01 | C | Intent Alignment | Dual ⚡ | Title promise = content delivery |
| C02 | C | Direct Answer | GEO 🎯 | Core answer in first 150 words |
| C03 | C | Query Coverage | Dual ⚡ | Covers ≥3 query variants (synonyms, long-tail) |
| C04 | C | Definition First | GEO 🎯 | Key terms defined on first use |
| C05 | C | Topic Scope | GEO 🎯 | Explicitly states what is and isn't covered |
| C06 | C | Audience Targeting | Dual ⚡ | States "this article is for..." |
| C07 | C | Semantic Coherence | GEO 🎯 | Logical flow between paragraphs, no jumps |
| C08 | C | Use Case Mapping | GEO 🎯 | Decision framework: when to choose A vs B |
| C09 | C | FAQ Coverage | GEO 🎯 | Structured FAQ covering long-tail follow-ups |
| C10 | C | Semantic Closure | Dual ⚡ | Conclusion answers the opening question + next steps |
| O01 | O | Heading Hierarchy | Dual ⚡ | H1→H2→H3, no level skipping |
| O02 | O | Summary Box | GEO 🎯 | Has TL;DR or Key Takeaways section |
| O03 | O | Data Tables | GEO 🎯 | Comparisons and specs presented in tables |
| O04 | O | List Formatting | GEO 🎯 | Parallel items use bullet or numbered lists |
| O05 | O | Schema Markup | GEO 🎯 | Appropriate JSON-LD (Article/FAQ/HowTo/etc.) |
| O06 | O | Section Chunking | GEO 🎯 | Each section has single topic; paragraphs 3–5 sentences |
| O07 | O | Visual Hierarchy | SEO 🔍 | Key concepts bolded or highlighted |
| O08 | O | Anchor Navigation | Dual ⚡ | Table of contents with jump links |
| O09 | O | Information Density | GEO 🎯 | No filler; consistent terminology throughout |
| O10 | O | Multimedia Structure | Dual ⚡ | Images/videos have captions and carry information |
| R01 | R | Data Precision | GEO 🎯 | ≥5 precise numbers with units (%, $, ms) |
| R02 | R | Citation Density | GEO 🎯 | ≥1 external citation per 500 words |
| R03 | R | Source Hierarchy | GEO 🎯 | Primary sources first; ≥3 Tier 1–2 sources |
| R04 | R | Evidence-Claim Mapping | GEO 🎯 | Every claim backed by evidence immediately after |
| R05 | R | Methodology Transparency | GEO 🎯 | Sample size, steps, and criteria documented |
| R06 | R | Timestamp & Versioning | Dual ⚡ | Last updated <1 year; version changes noted |
| R07 | R | Entity Precision | GEO 🎯 | Full names for people/orgs/products; no "a company" |
| R08 | R | Internal Link Graph | SEO 🔍 | Descriptive anchor texts forming topic clusters |
| R09 | R | HTML Semantics | GEO 🎯 | Uses <article>, <figure>, <time>, <cite> |
| R10 | R | Content Consistency | Dual ⚡ | Data self-consistent; no broken links (404) |
| E01 | E | Original Data | GEO 🎯 | First-party surveys, experiments, or statistics |
| E02 | E | Novel Framework | GEO 🎯 | Named, citable original framework or model |
| E03 | E | Primary Research | GEO 🎯 | Original experiments/surveys with documented process |
| E04 | E | Contrarian View | GEO 🎯 | Challenges consensus with evidence |
| E05 | E | Proprietary Visuals | Dual ⚡ | ≥2 original infographics, charts, or diagrams |
| E06 | E | Gap Filling | GEO 🎯 | Covers questions competitors don't |
| E07 | E | Practical Tools | Dual ⚡ | Downloadable templates, checklists, or calculators |
| E08 | E | Depth Advantage | GEO 🎯 | Deeper than competing content on same topic |
| E09 | E | Synthesis Value | GEO 🎯 | Cross-domain knowledge combination (A+B=C) |
| E10 | E | Forward Insights | GEO 🎯 | Data-backed predictions and trend analysis |
| ID | Dim | Check Item | Priority | One-Line Standard |
|---|---|---|---|---|
| Exp01 | Exp | First-Person Narrative | SEO 🔍 | Contains "I tested" or "We found" + action verbs |
| Exp02 | Exp | Sensory Details | SEO 🔍 | ≥10 sensory words (smooth, heavy, bright) |
| Exp03 | Exp | Process Documentation | Dual ⚡ | Step-by-step process with timeline |
| Exp04 | Exp | Tangible Proof | SEO 🔍 | ≥2 original photos/screenshots with timestamps |
| Exp05 | Exp | Usage Duration | SEO 🔍 | States "after X months of use..." |
| Exp06 | Exp | Problems Encountered | Dual ⚡ | Shares ≥2 real problems + solutions |
| Exp07 | Exp | Before/After Comparison | SEO 🔍 | Shows change, improvement, or difference |
| Exp08 | Exp | Quantified Metrics | Dual ⚡ | Measurable experience data (time, cost, success rate) |
| Exp09 | Exp | Repeated Testing | SEO 🔍 | Multiple tests or long-term tracking |
| Exp10 | Exp | Limitations Acknowledged | GEO 🎯 | States "we only tested X scenario" |
| Ept01 | Ept | Author Identity | SEO 🔍 | Byline + avatar + bio (>30 words) |
| Ept02 | Ept | Credentials Display | SEO 🔍 | Relevant degrees, certs, years of experience |
| Ept03 | Ept | Professional Vocabulary | Dual ⚡ | Accurate industry jargon, no misuse |
| Ept04 | Ept | Technical Depth | Dual ⚡ | Parameters, thresholds, examples are actionable |
| Ept05 | Ept | Methodology Rigor | GEO 🎯 | Analysis method is reproducible |
| Ept06 | Ept | Edge Case Awareness | Dual ⚡ | Discusses ≥2 exceptions or "when this doesn't apply" |
| Ept07 | Ept | Historical Context | SEO 🔍 | Shows knowledge of the field's evolution |
| Ept08 | Ept | Reasoning Transparency | GEO 🎯 | "We chose A over B because..." with tradeoffs |
| Ept09 | Ept | Cross-domain Integration | Dual ⚡ | Connects knowledge across fields |
| Ept10 | Ept | Editorial Process | SEO 🔍 | "Reviewed by" or "Fact-checked by" labels |
| A01 | A | Backlink Profile | SEO 🔍 | Cited by authoritative sites (.edu, .gov, leaders) |
| A02 | A | Media Mentions | SEO 🔍 | "Featured in" with media logos |
| A03 | A | Industry Awards | SEO 🔍 | Displays relevant industry awards or recognition |
| A04 | A | Publishing Record | SEO 🔍 | Conference talks, publications, patents |
| A05 | A | Brand Recognition | Dual ⚡ | Brand has search volume |
| A06 | A | Social Proof | SEO 🔍 | Authentic user testimonials with real details |
| A07 | A | Knowledge Graph Presence | Dual ⚡ | Has Wikipedia entry or Google Knowledge Panel |
| A08 | A | Entity Consistency | GEO 🎯 | Brand/author info consistent across the web |
| A09 | A | Partnership Signals | SEO 🔍 | Shows partnerships with authoritative organizations |
| A10 | A | Community Standing | SEO 🔍 | Active and influential in professional communities |
| T01 | T | Legal Compliance | SEO 🔍 | Privacy Policy + Terms of Service present |
| T02 | T | Contact Transparency | SEO 🔍 | Physical address or ≥2 contact methods |
| T03 | T | Security Standards | SEO 🔍 | Site-wide HTTPS, no security warnings |
| T04 | T | Disclosure Statements | Dual ⚡ | Affiliate links disclosed (veto if missing) |
| T05 | T | Editorial Policy | SEO 🔍 | Content standards and review process published |
| T06 | T | Correction & Update Policy | Dual ⚡ | Has corrections page or changelog |
| T07 | T | Ad Experience | SEO 🔍 | Ads <30% of page; no intrusive popups |
| T08 | T | Risk Disclaimers | Dual ⚡ | YMYL topics have necessary disclaimers |
| T09 | T | Review Authenticity | Dual ⚡ | Reviews show authenticity signals |
| T10 | T | Customer Support | SEO 🔍 | Clear return policy, complaint channels, response SLA |
| Status | Score | Meaning |
|---|---|---|
| Pass | 10 | Fully meets criteria |
| Partial | 5 | Partially meets criteria |
| Fail | 0 | Does not meet criteria |
Each dimension: sum of 10 items = 0–100 points.
GEO Score = (C + O + R + E) / 4
SEO Score = (Exp + Ept + A + T) / 4
Total Score = (GEO Score + SEO Score) / 2
For specific content types, apply dimension weights instead of equal averaging:
Weighted Score = Σ (dimension_score × weight)
| Dim | Product Review | How-to Guide | Comparison | Landing Page | Blog Post | FAQ Page | Alternative | Best-of | Testimonial |
|---|---|---|---|---|---|---|---|---|---|
| C | 10% | 20% | 10% | 20% | 25% | 25% | 10% | 10% | 10% |
| O | 10% | 20% | 20% | 10% | 10% | 25% | 15% | 25% | 5% |
| R | 15% | 10% | 25% | 5% | 10% | 15% | 25% | 20% | 15% |
| E | 20% | 5% | 10% | 5% | 20% | 5% | 5% | 15% | 10% |
| Exp | 20% | 5% | 5% | 5% | 10% | 5% | 15% | 5% | 30% |
| Ept | 5% | 20% | 15% | 5% | 10% | 10% | 5% | 10% | 5% |
| A | 5% | 5% | 5% | 25% | 5% | 5% | 5% | 5% | 5% |
| T | 15% | 15% | 10% | 25% | 10% | 10% | 20% | 10% | 20% |
Example (Product Review):
Score = C×0.10 + O×0.10 + R×0.15 + E×0.20 + Exp×0.20 + Ept×0.05 + A×0.05 + T×0.15
| Score | Rating | Meaning |
|---|---|---|
| 90–100 | Excellent | Industry benchmark quality |
| 75–89 | Good | Outperforms most competitors |
| 60–74 | Medium | Meets baseline with room to improve |
| 40–59 | Low | Needs focused optimization |
| 0–39 | Poor | Needs complete restructuring |
These items trigger a risk alert regardless of total score:
| Item | Trigger | Impact |
|---|---|---|
| T04 Disclosure Statements | Affiliate links present without disclosure | Trust dimension = Fail |
| C01 Intent Alignment | Clickbait; content doesn't match title | Content credibility destroyed |
| R10 Content Consistency | Core data contradicts itself | Citation reliability = zero |
What is the primary goal?
│
├── Teach users how to do something → Blog (guides)
├── Your product vs one competitor → Alternative
├── Objective comparison of 3+ products → Best-of
├── Show product fits a persona → Use-case
├── Show verified customer results → Testimonial
├── Answer common questions → FAQ (or embed as faqs[] field)
├── Describe product features → Landing
└── Share industry insights or trends → Blog (insights)
| Confusion Pair | Type A | Type B | Deciding Factor |
|---|---|---|---|
| 1v1 vs multi-product | Alternative | Best-of | Alternative = persuade to switch; Best-of = help choose |
| User story vs customer proof | Use-case | Testimonial | Use-case = fictional persona; Testimonial = verified customer data |
| FAQ quantity | Embed as faqs[] |
Dedicated FAQ page | <5 embed; 5–9 either; 10+ dedicated page |
| Product intro vs tutorial | Landing | Blog | Landing = short, feature-focused; Blog = long, educational |
Source tiers: Tier 1 (.gov/.edu/official docs) | Tier 2 (industry authorities: Moz, Ahrefs, Forbes) | Tier 3 (community blogs)
| Content Type | Tier 1 | Tier 2 | Tier 3 | Notes |
|---|---|---|---|---|
| Blog (guides) | 1+ | 0–1 | 0 | Official tool documentation |
| Blog (tools) | 2+ | 1+ | 0–1 | Each tool's pricing page required |
| Blog (insights) | 2+ | 2+ | 0–1 | Research and data sources |
| Alternative | 1+ | 0–1 | 0 | Competitor's official pricing page |
| Best-of | 3+ | 1+ | 0–1 | All tools' pricing pages |
| Use-case | 0–1 | 1+ | 0 | Industry statistics |
| FAQ | 0 | 0 | 0 | User-generated content primarily |
| Testimonial | 0 | 0 | 0 | Verified customer data |
| Landing | 0–1 | 0 | 0 | Own product data |
| Content Type | First Review | Regular Review | Re-review Trigger |
|---|---|---|---|
| Blog | Before publish | Quarterly | Ranking drops, info expires |
| Alternative | Before publish | Monthly | Competitor price/feature changes |
| Best-of | Before publish | Monthly | New tools launch, pricing changes |
| Use-case | Before publish | Quarterly | Industry data updates |
| Testimonial | Before publish | Annually | Customer info changes |
| FAQ | Before publish | Quarterly | New frequent questions appear |
| Landing | Before publish | As needed | Product feature updates |
┌──────────────┬──────────────┬──────────────┬──────────────────────┐
│ Gate 1 │ Gate 2 │ Gate 3 │ Gate 4 │
│ Preparation │ CORE Review │ EEAT Review │ Score & Action │
├──────────────┼──────────────┼──────────────┼──────────────────────┤
│ Set content │ Review │ Review │ Calculate GEO/SEO/ │
│ type │ C-O-R-E │ Exp-Ept-A-T │ Total scores │
│ Select │ (40 items) │ (40 items) │ Identify top 5 │
│ weights │ Mark each │ Mark each │ improvements │
│ Prepare │ Pass/Partial │ Pass/Partial │ Check veto items │
│ competitor │ /Fail │ /Fail │ Generate action plan │
│ references │ │ │ │
├──────────────┼──────────────┼──────────────┼──────────────────────┤
│ Output: │ Output: │ Output: │ Output: │
│ Config sheet │ CORE card │ EEAT card │ Report + priorities │
└──────────────┴──────────────┴──────────────┴──────────────────────┘
These conditions require immediate action — do not wait for full evaluation:
| Trigger | Severity | Immediate Action |
|---|---|---|
| T04 Fail: Affiliate links without disclosure | Critical | Add disclosure banner at page top |
| C01 Fail: Clickbait title | Critical | Rewrite title and first paragraph |
| R10 Fail: Core data contradicts itself | High | Verify all data before publishing |
| 3+ Fails in same dimension | High | Systematic restructure of that dimension |
Draft ──→ Review ──→ Live ──→ Monitor ──→ Update / Archive
│ │ │ │ │
Write + Gate 1-4 Publish Track KPI ┌──┴──┐
self-check quality + rankings Update Archive
Update vs Archive Decision:
Is the content still accurate?
├── YES → Update lastVerified date, continue monitoring
└── NO → Can it be updated?
├── YES (minor fix) → Update content, note revision
├── YES (major rewrite) → Full Gate 1-4 process
└── NO (topic obsolete) → Archive:
① Add expiration notice at page top
② Set 301 redirect to replacement content
③ Remove from site navigation
| Metric | Poor | Average | Good | Excellent |
|---|---|---|---|---|
| Bounce Rate | >70% | 50–70% | 30–50% | <30% |
| Time on Page | <1 min | 1–2 min | 2–4 min | >4 min |
| Scroll Depth | <25% | 25–50% | 50–75% | >75% |
| SERP CTR | <1% | 1–3% | 3–5% | >5% |
| AI Citation Rate | 0% | 5–15% | 15–30% | >30% |
Detailed Pass/Partial/Fail criteria for all 40 CORE check items. Items marked with ❌/✅ include wrong/correct examples.
Does content clearly answer user intent? Can AI quickly grasp the core value and boundaries?
C01: Intent Alignment | Dual ⚡ - Pass: Title promise fully delivered in content; intent type (Informational/Transactional/Commercial) is clear - Partial: Title and content mostly aligned with minor drift - Fail: Clickbait; content doesn't match title promise - ❌ Title "10 Best Free Tools" → content only covers 3 paid tools - ✅ Title "How to Download YouTube Subtitles (3 Free Methods)" → covers exactly 3 free methods
C02: Direct Answer | GEO-First 🎯 - Pass: First 150 words contain a clear definition sentence or conclusion phrase (directly citable by AI) - Partial: Answer within first 300 words but with lengthy preamble - Fail: Answer buried in the middle or end; opening is all background - ❌ "YouTube is the world's largest video platform... [answer in paragraph 4]" - ✅ "The fastest way to download YouTube subtitles is using NoteLM — paste URL, select format, click download. Free, no account required."
C03: Query Coverage | Dual ⚡ - Pass: Covers ≥3 query variants (synonyms, long-tail terms, related questions); appropriate entity density - Partial: Covers 1–2 variants - Fail: Only targets a single exact query
C04: Definition First | GEO-First 🎯 - Pass: All key terms defined on first use ("X is Y..." pattern) - Partial: Most terms defined - Fail: Terms used without definition; readers or AI may misinterpret
C05: Topic Scope | GEO-First 🎯 - Pass: Explicitly states "This article covers X, not Y"; reaches AI's expected completeness threshold - Partial: Implied boundaries, not explicitly stated - Fail: Scope unclear; content sprawls
C06: Audience Targeting | Dual ⚡ - Pass: Explicitly states target reader (e.g., "for beginners"); language style matches audience - Partial: Implied through content difficulty - Fail: Audience unclear; inconsistent difficulty level
C07: Semantic Coherence | GEO-First 🎯 - Pass: Logical connectors between paragraphs (therefore, however, given that); no semantic jumps; follows NLP prediction logic - Partial: Mostly coherent with occasional jumps - Fail: Frequent logic breaks and topic jumps
C08: Use Case Mapping | GEO-First 🎯 - Pass: Clearly states applicable and inapplicable scenarios; provides decision framework (when to choose A vs B) - Partial: Mentions some scenarios but incomplete - Fail: No use case guidance
C09: FAQ Coverage | GEO-First 🎯 - Pass: Has structured FAQ (FAQPage Schema or explicit Q&A format) covering high-frequency long-tail follow-ups - Partial: Has Q&A content but not structured - Fail: No FAQ or Q&A content - ❌ No FAQ section, or only 1–2 generic questions - ✅ 8+ FAQs with FAQPage Schema, questions in natural language ("Can I download subtitles without an account?")
C10: Semantic Closure | Dual ⚡ - Pass: Conclusion explicitly answers the opening question and provides next steps or exploration direction - Partial: Has conclusion but doesn't fully loop back to opening - Fail: No conclusion, or conclusion unrelated to opening
Is content structure clear? Does it reduce AI token processing burden and improve extraction efficiency?
O01: Heading Hierarchy | Dual ⚡ - Pass: Single H1; H2→H3 properly nested; no level skipping; logical progression - Partial: Minor level skipping but overall clear - Fail: Chaotic hierarchy, multiple H1s, or severe skipping
O02: Summary Box | GEO-First 🎯 - Pass: Has prominent TL;DR, key takeaways, or highlighted summary box - Partial: Has summary but not prominent - Fail: No summary of any kind - ❌ Article starts with 500-word introduction, no summary anywhere - ✅ "Key Takeaways: • NoteLM is fastest (2–5 sec) • Tool B best for batch • All support SRT format"
O03: Data Tables | GEO-First 🎯 - Pass: Uses HTML tables for comparisons, specs, and data with clear headers - Partial: Has tables but unclear structure - Fail: Uses prose where tables would be better - ❌ "Tool A is cheaper than Tool B which is cheaper than Tool C" (prose comparison) - ✅ Structured table: | Tool | Price | Speed | Accuracy | Winner |
O04: List Formatting | GEO-First 🎯 - Pass: Appropriate list usage (~1–2 lists per 500 words); parallel items listed - Partial: Insufficient or slightly excessive list usage - Fail: Lists overused or completely absent
O05: Schema Markup | GEO-First 🎯 - Pass: Correct JSON-LD (Article/FAQ/HowTo/Product/etc.) matching content type - Partial: Has Schema but type doesn't match well - Fail: No Schema markup - ❌ FAQ content on page but no FAQPage JSON-LD markup - ✅ FAQPage + Article + BreadcrumbList JSON-LD matching page content type
O06: Section Chunking | GEO-First 🎯 - Pass: Each H2/H3 section has a single clear topic; paragraphs 3–5 sentences; easy to scan - Partial: Most sections clear; some paragraphs too long - Fail: Mixed topics within sections; walls of text - ❌ Single 500-word paragraph covering setup, features, and pricing - ✅ Separate H2 sections for Setup (3 sentences), Features (4 sentences), Pricing (3 sentences)
O07: Visual Hierarchy | SEO-First 🔍 - Pass: Important content emphasized through formatting (bold, boxes, color); key concepts and conclusions bolded - Partial: Some visual emphasis - Fail: No visual hierarchy throughout
O08: Anchor Navigation | Dual ⚡ - Pass: Has table of contents with anchor links; clear breadcrumb navigation - Partial: Has TOC but no anchor links - Fail: Long-form content without any navigation aids
O09: Information Density | GEO-First 🎯 - Pass: High information density; no filler content; same concept uses same term throughout (no synonym rotation) - Partial: Minor repetition or filler - Fail: Significant filler content
O10: Multimedia Structure | Dual ⚡ - Pass: Images/videos/code blocks have clear captions; positioned purposefully; convey step or structural information - Partial: Has multimedia but lacks descriptions - Fail: No multimedia, or decorative-only images
Are data, evidence, and logic chains sufficient, precise, and verifiable by AI?
R01: Data Precision | GEO-First 🎯 - Pass: ≥5 precise data points with units (%, USD, kg, ms); directly extractable - Partial: 2–4 precise data points - Fail: No precise data; all vague descriptions ("many", "approximately", "a lot") - ❌ "NoteLM is very fast and highly accurate with excellent support" - ✅ "NoteLM processes videos in 2–5 seconds with 94% accuracy, supporting 100+ languages"
R02: Citation Density | GEO-First 🎯 - Pass: ≥1 external citation per 500 words; ≥3 source types (academic, media, official, industry) - Partial: ≥1 citation per 1000 words; 2 source types - Fail: Insufficient citations or single source type - ❌ 2000-word article with zero external citations - ✅ Every 500 words has ≥1 external source ("According to Wyzowl's 2026 Report...")
R03: Source Hierarchy | GEO-First 🎯 - Tier 1: .gov, .edu, .org, Wikipedia, PubMed, original papers and standards - Tier 2: Industry authority sites (Moz, Ahrefs, Forbes, etc.) - Tier 3: Notable industry blogs - Tier 4: Short links, affiliate links (negative signal) - Pass: Primary sources prioritized; ≥3 Tier 1–2 sources - Partial: 1–2 Tier 1–2 sources - Fail: No authoritative sources or multiple Tier 4 links
R04: Evidence-Claim Mapping | GEO-First 🎯 - Pass: Every core claim immediately followed by evidence, data, or citation - Partial: Most claims backed by evidence - Fail: Multiple claims without evidence ("empty assertions")
R05: Methodology Transparency | GEO-First 🎯 - Pass: Sample size, test steps, and statistical criteria documented; conclusions reproducible - Partial: Partial methodology description - Fail: Conclusions without methodological support
R06: Timestamp & Versioning | Dual ⚡ - Pass: Last updated <1 year; date visible; version change notes present - Partial: Updated 1–3 years ago - Fail: >3 years old or date not visible
R07: Entity Precision | GEO-First 🎯 - Pass: Full names for people, organizations, product models, locations, dates; no vague references ("a company", "a product") - Partial: Most entities precise; occasional vagueness - Fail: Frequent vague references
R08: Internal Link Graph | SEO-First 🔍 - Pass: Descriptive anchor texts building conceptual connections; forms topic clusters - Partial: Has internal links but anchor texts are non-descriptive - Fail: No internal links, or "click here" style anchors
R09: HTML Semantics | GEO-First 🎯
- Pass: Correct use of <article>, <aside>, <figure>, <time>, <cite> semantic tags
- Partial: Some semantic tags used
- Fail: Pure <div> markup; no semantic structure
R10: Content Consistency | Dual ⚡ - Pass: All data self-consistent; no broken links (404); no redirect chains; has corrections or revision log - Partial: Mostly consistent with minor discrepancies - Fail: Data contradicts itself or significant broken links
Does content provide information scarce in LLM training data? Why would AI cite you instead of others?
E01: Original Data | GEO-First 🎯 - Pass: Has first-party data (e.g., "We analyzed 10,000 samples"); dataset is citable - Partial: Some original data - Fail: All data cited from others - ❌ Restating manufacturer specs from official docs - ✅ "In our testing of 50 samples across 8 categories, NoteLM achieved 94% accuracy" (original test data)
E02: Novel Framework | GEO-First 🎯 - Pass: Proposes a named, citable original framework or model (e.g., "the CORE-EEAT Framework") - Partial: Innovates on an existing framework - Fail: No framework innovation
E03: Primary Research | GEO-First 🎯 - Pass: Has documented research process (experimental conditions, metrics, control groups) - Partial: Some primary research - Fail: No primary research
E04: Contrarian View | GEO-First 🎯 - Pass: Challenges mainstream consensus with data or logic backing - Partial: Some differentiated views - Fail: Entirely follows conventional wisdom
E05: Proprietary Visuals | Dual ⚡ - Pass: ≥2 original infographics, flowcharts, or data visualizations - Partial: 1 original visualization - Fail: No original visuals
E06: Gap Filling | GEO-First 🎯 - Pass: Covers niche questions or edge cases that competitors miss - Partial: Partially fills content gaps - Fail: Content highly similar to competitors
E07: Practical Tools | Dual ⚡ - Pass: ≥1 downloadable or copyable template, checklist, calculator, or interactive component - Partial: Has examples but not actionable enough - Fail: No practical tools
E08: Depth Advantage | GEO-First 🎯 - Pass: Depth clearly exceeds competing content on same topic - Partial: Comparable depth to competitors - Fail: Shallower than competitors
E09: Synthesis Value | GEO-First 🎯 - Pass: Combines cross-domain knowledge to produce new insights - Partial: Some cross-domain content but combination isn't novel - Fail: Knowledge siloed; no cross-domain integration
E10: Forward Insights | GEO-First 🎯 - Pass: Data-backed predictions or trend analysis with clear reasoning - Partial: Some forward-looking content - Fail: Only discusses past and present
Detailed Pass/Partial/Fail criteria for all 40 EEAT check items. Items marked with ❌/✅ include wrong/correct examples.
Proves the creator "was there" with first-hand, real-world experience.
Exp01: First-Person Narrative | SEO-First 🔍 - Pass: First-person + action verb combinations ("I tested", "We analyzed") - Partial: Has only first-person OR only action verbs - Fail: Entirely third-person narration - ❌ "This tool provides fast transcription with high accuracy" - ✅ "I tested NoteLM on 50 YouTube videos over 2 weeks and found 94% accuracy in English content"
Exp02: Sensory Details | SEO-First 🔍 - Pass: ≥10 sensory words (smooth, heavy, bright, crisp, sluggish) - Partial: 5–9 sensory words - Fail: <5 sensory words
Exp03: Process Documentation | Dual ⚡ - Pass: Detailed operation or usage process with steps, timeline, and key decision points - Partial: Partial process description - Fail: No process documentation
Exp04: Tangible Proof | SEO-First 🔍 - Pass: ≥2 clearly original images (photos/screenshots/hand-drawn) with timestamps or environmental context - Partial: 1 original image - Fail: No original images - ❌ Stock photo of a laptop with generic overlay text - ✅ Original screenshot with timestamp showing actual transcription result + accuracy comparison photo
Exp05: Usage Duration | SEO-First 🔍 - Pass: Explicitly states testing or usage duration ("after 3 months of continuous use") - Partial: Implied duration but not explicit - Fail: No duration stated
Exp06: Problems Encountered | Dual ⚡ - Pass: Honestly shares ≥2 problems encountered, with solutions or workarounds - Partial: Mentions 1 problem - Fail: All positive descriptions; no problems mentioned - ❌ "NoteLM is perfect — no issues at all, 10/10 recommended!" - ✅ "Two issues: 1) Heavy accents dropped accuracy to ~80%, 2) Videos over 3 hours sometimes timeout"
Exp07: Before/After Comparison | SEO-First 🔍 - Pass: Clear before/after comparison or side-by-side comparison with alternatives - Partial: Implied comparison - Fail: No comparison
Exp08: Quantified Metrics | Dual ⚡ - Pass: Quantified experience data (time spent, cost, success rate, performance numbers) - Partial: Some quantified data - Fail: Purely subjective experience with no quantification
Exp09: Repeated Testing | SEO-First 🔍 - Pass: States that multiple tests or long-term tracking were conducted - Partial: Implied repeat testing - Fail: Single test only
Exp10: Limitations Acknowledged | GEO-First 🎯 - Pass: Explicitly states experience limitations ("We only tested X scenario", "Not tested on Y platform") - Partial: Partially acknowledges limitations - Fail: No limitations acknowledged - ❌ "This is the best tool for everyone in every situation" - ✅ "Limitation: We only tested English and Spanish. Results may differ for tonal languages like Mandarin"
Proves the creator has the knowledge and skills to get the subject right.
Ept01: Author Identity | SEO-First 🔍 - Pass: Byline + avatar + bio (>30 words) with clear role boundaries - Partial: Has 1–2 of the above - Fail: No author information - ❌ "By Admin" with no photo or bio - ✅ "By Sarah Chen, AI Product Researcher (5+ years). Previously at Google AI, published in ACM. [Photo + Social links]"
Ept02: Credentials Display | SEO-First 🔍 - Pass: Displays relevant professional qualifications (degrees, certificates, years of experience, project history) - Partial: Has credentials but weak relevance - Fail: No credentials displayed
Ept03: Professional Vocabulary | Dual ⚡ - Pass: Accurate use of industry jargon; average word length >5.5 characters; no terminology misuse - Partial: Moderate professionalism - Fail: Vocabulary too simple or terms misused
Ept04: Technical Depth | Dual ⚡ - Pass: Technical details accurate and deep; parameters, thresholds, and examples are actionable - Partial: Has technical content but shallow - Fail: Technical content superficial or contains obvious errors - ❌ "The AI uses advanced technology to transcribe" - ✅ "Uses Whisper large-v3 model (1550M params), WER of 5.2% on LibriSpeech, supports beam_size=5 for accuracy"
Ept05: Methodology Rigor | GEO-First 🎯 - Pass: Analysis methodology is clear, reproducible, and follows industry standards - Partial: Has methodology but not rigorous enough - Fail: No methodology or methodology has obvious flaws
Ept06: Edge Case Awareness | Dual ⚡ - Pass: Discusses ≥2 edge cases or exceptions, stating "when this doesn't apply" - Partial: Mentions 1 edge case - Fail: No edge cases considered
Ept07: Historical Context | SEO-First 🔍 - Pass: Demonstrates understanding of the field's development history; has timeline or evolution explanation - Partial: Some historical background - Fail: Lacks historical perspective
Ept08: Reasoning Transparency | GEO-First 🎯 - Pass: Explicitly states cause-effect relationships and tradeoffs ("We chose A over B because...") - Partial: Some reasoning but incomplete - Fail: Conclusions given without reasoning - ❌ "NoteLM is the best choice" (no reasoning given) - ✅ "We recommend NoteLM over Otter.ai because: free vs $16.99/mo, 94% vs 91% accuracy, no account needed"
Ept09: Cross-domain Integration | Dual ⚡ - Pass: Effectively integrates cross-domain knowledge, generating new perspectives - Partial: Some cross-domain content - Fail: Knowledge siloed; single domain
Ept10: Editorial Process | SEO-First 🔍 - Pass: Has "Reviewed by" or "Fact-checked by" labels clearly visible - Partial: Has editorial review but no visible labels - Fail: No editorial process
Proves that third parties — especially high-authority entities — recognize this source.
A01: Backlink Profile | SEO-First 🔍 - Pass: Naturally cited by authoritative sites (.edu, .gov, or industry leaders) - Partial: Some backlinks - Fail: No notable backlinks
A02: Media Mentions | SEO-First 🔍 - Pass: Has "Featured in" with media logos or mainstream news coverage - Partial: Minor media mentions - Fail: No media mentions
A03: Industry Awards | SEO-First 🔍 - Pass: Displays relevant industry awards or certification body recognition - Partial: Has awards but weak relevance - Fail: No awards
A04: Publishing Record | SEO-First 🔍 - Pass: Conference talks, professional publications, academic papers, or patents - Partial: Some publishing record - Fail: No publishing record
A05: Brand Recognition | Dual ⚡ - Pass: Brand has industry recognition; brand terms have search volume - Partial: Some brand awareness - Fail: Brand unknown
A06: Social Proof | SEO-First 🔍 - Pass: Authentic user reviews and recommendations (real photos, specific content, natural timing distribution) - Partial: Reviews present but credibility uncertain - Fail: No social proof
A07: Knowledge Graph Presence | Dual ⚡ - Pass: Has Wikipedia entry or Google Knowledge Panel - Partial: Partially indexed in knowledge graphs - Fail: Not in any knowledge graph
A08: Entity Consistency | GEO-First 🎯 - Pass: Brand and author name and description consistent across the entire web; facilitates AI attribution - Partial: Mostly consistent - Fail: Inconsistent information across web; contradictions exist
A09: Partnership Signals | SEO-First 🔍 - Pass: Displays partnerships with authoritative organizations or industry body endorsements - Partial: Some partnership signals - Fail: No partnership signals
A10: Community Standing | SEO-First 🔍 - Pass: Active and influential in professional communities; high-quality discussions and ratings traceable - Partial: Some community participation - Fail: No community presence
Is the site technically, legally, and governance-wise trustworthy?
T01: Legal Compliance | SEO-First 🔍 - Required: Privacy Policy, Terms of Service - Bonus: Cookie Policy, GDPR compliance, Affiliate Disclosure - Pass: Required pages present plus bonus items - Partial: Required pages present only - Fail: Missing required pages - ❌ No Privacy Policy or Terms page; footer only has copyright notice - ✅ Footer links to Privacy Policy, Terms of Service, Cookie Policy, GDPR compliance page
T02: Contact Transparency | SEO-First 🔍 - Pass: Physical address or ≥2 contact methods; verifiable real entity behind the site - Partial: Email only - Fail: No contact information (anonymous site)
T03: Security Standards | SEO-First 🔍 - Pass: Site-wide HTTPS; no security warnings; extra security for payment and login pages - Partial: Has HTTPS but some pages insecure - Fail: Uses HTTP
T04: Disclosure Statements | Dual ⚡ — VETO ITEM - Pass: Affiliate links present AND clearly disclosed (at page top or next to links) - Partial: No affiliate links (not applicable) - Fail: Affiliate links present WITHOUT disclosure — automatic veto - ❌ Article contains affiliate links but no disclosure anywhere — veto triggered - ✅ Page top: "Disclosure: Some links are affiliate. We earn commission at no extra cost. All opinions are our own."
T05: Editorial Policy | SEO-First 🔍 - Pass: Content standards, review process, and editorial team information published - Partial: Some editorial guidelines - Fail: No editorial policy
T06: Correction & Update Policy | Dual ⚡ - Pass: Has corrections page, update principles, and revision history - Partial: Has update dates but no formal correction mechanism - Fail: No correction or update mechanism - ❌ Article published in 2024 with no update date or correction history - ✅ "Last updated: 2026-02-01. Correction (2026-01-15): Updated Otter.ai pricing from $12.99 to $16.99/mo"
T07: Ad Experience | SEO-First 🔍 - Pass: Ads <30% of page area; no fullscreen popups; doesn't obstruct content - Partial: Ads 30–50% of page area - Fail: Ads >50% or has fullscreen/deceptive popups
T08: Risk Disclaimers | Dual ⚡ - Pass: YMYL topics (finance/health/legal) have necessary disclaimers and risk warnings - Partial: Some disclaimer coverage - Fail: YMYL content with no disclaimers
T09: Review Authenticity | Dual ⚡ - Pass: User reviews show authenticity signals (real photos, specific details, natural timing distribution) - Partial: Reviews present but authenticity uncertain - Fail: Reviews obviously fake or absent
T10: Customer Support | SEO-First 🔍 - Pass: Clear return/refund policy, complaint channels, and defined response SLA - Partial: Policy exists but unclear - Fail: No customer support or return policy
| Engine | Citation Style | Preferred Content | Priority Items |
|---|---|---|---|
| Google AI Overview | Snippet extraction from paragraphs, lists, tables, FAQs | Structured paragraphs, definition sentences, comparison tables, FAQ pairs | C02, O03, O05, C09 |
| ChatGPT Browse | Conversational answers with citation links | Complete information blocks, data statements with citations, standalone conclusions | C02, R01, R02, E01 |
| Perplexity AI | Multi-source synthesis + inline citations | Original research data, tiered sources, methodology-transparent conclusions | E01, R03, R05, Ept05 |
| Claude | Precision-first with nuanced arguments | Argument + evidence chains, balanced views, transparent reasoning | R04, Ept08, Exp10, R03 |
These items have the highest impact on AI engine visibility across all engines:
| Rank | Item | Why It Matters |
|---|---|---|
| 1 | C02 Direct Answer | All engines extract from the first paragraph; first 150 words determine citation |
| 2 | C09 FAQ Coverage | FAQ structure directly matches user follow-ups; heavily extracted by Google AI Overview and Perplexity |
| 3 | O03 Data Tables | Comparison data is the most extractable structured format; significantly higher citation rate than prose |
| 4 | O05 Schema Markup | JSON-LD helps AI understand content type and structure; increases correct parsing probability |
| 5 | E01 Original Data | AI prefers citing exclusive, verifiable data sources over second-hand summaries |
| 6 | O02 Summary Box | Key Takeaways are often the first choice for AI summary citations |
Maps check items to UI components, Schema types, and data fields.
| Check Item | Component | Schema | Data Field |
|---|---|---|---|
| C02 Direct Answer | DirectAnswerBox | — | excerpt, directAnswer |
| C09 FAQ Coverage | FAQAccordion (includeSchema) | FAQPage | faqs[] |
| O02 Summary Box | SummaryBox (7 variants) | — | keyTakeaways[] |
| O03 Data Tables | DataTable (5 variants) | — | comparisons, tools |
| O05 Schema Markup | JSON-LD generator | Per content type | Auto-generated |
| O08 Anchor Nav | TableOfContents | — | Auto-extract H2/H3 |
| R02 Citations | CitationsList | — | eeat.externalCitations[] |
| R03 Source Tiers | CitationsList (showTierBadge) | — | eeat.externalCitations[].tier |
| R05 Methodology | MethodologySection | — | methodology |
| R06 Timestamp | VerificationBadge | — | lastVerified, updateDate |
| R08 Internal Links | RelatedResourcesSection | BreadcrumbList | relatedPosts[], relatedTools[] |
| Exp06 Problems | ProConsList | — | cons[] |
| Exp07 Before/After | DataTable (before-after) | — | beforeAfter |
| Ept01 Author | AuthorCard | Person | eeat.authorInfo |
| A06 Social Proof | TestimonialCard | Review | testimonials[] |
| T04 Disclosure | Disclosure banner | — | eeat.dataDisclaimer |
| Content Type | Required Schema | Conditional Schema |
|---|---|---|
| Blog (guides) | Article, Breadcrumb | FAQ, HowTo |
| Blog (tools) | Article, Breadcrumb | FAQ, Review |
| Blog (insights) | Article, Breadcrumb | FAQ |
| Alternative | Comparison*, Breadcrumb, FAQ | AggregateRating |
| Best-of | ItemList, Breadcrumb, FAQ | AggregateRating per tool |
| Use-case | WebPage, Breadcrumb, FAQ | — |
| FAQ | FAQPage, Breadcrumb | — |
| Landing | SoftwareApplication, Breadcrumb, FAQ | WebPage |
| Testimonial | Review, Breadcrumb | FAQ, Person |
*Comparison = custom implementation extending Product Schema.
How CORE-EEAT content-level items feed into CITE Domain Rating domain-level scores.
When evaluating holistically, pair CORE-EEAT with CITE:
| CORE-EEAT Items | Feeds Into CITE Item | Relationship |
|---|---|---|
| C02 (Direct Answer), O02 (Summary Box), E01 (Original Data) | CITE C05-C08 (AI Citations) | Citable content drives domain-level AI citations |
| A07 (Knowledge Graph), A08 (Entity Consistency) | CITE I01 (Knowledge Graph Presence) | Content-level authority signals build domain identity |
| O05 (Schema Markup), R09 (HTML Semantics) | CITE I04 (Schema.org Coverage) | Page-level Schema contributes to domain-wide coverage |
| Ept01 (Author Identity), Ept02 (Credentials) | CITE I05 (Author Entity Recognition) | Content author signals build domain author recognition |
| T03 (Security Standards) | CITE T07 (Technical Security) | Same signal, different scope (page vs domain) |
| C03 (Query Coverage), E08 (Depth Advantage) | CITE E07-E08 (Topical Authority) | Content depth and coverage build domain topical authority |
Cross-evaluation pattern:
| Pattern | CITE Score | CORE-EEAT Score | Diagnosis | Action |
|---|---|---|---|---|
| Strong domain, strong content | High | High | Ideal state | Maintain and expand |
| Strong domain, weak content | High | Low | Authority wasted on poor content | Prioritize content quality (CORE-EEAT) |
| Weak domain, strong content | Low | High | Great content, invisible domain | Build domain authority (CITE) |
| Weak domain, weak content | Low | Low | Fundamental issues | Start with CORE-EEAT, then CITE |
| # | Error | Item | Wrong | Right |
|---|---|---|---|---|
| 1 | Answer buried | C02 | 500-word preamble before answer | Core answer in first 150 words |
| 2 | Clickbait title | C01 | "Shocking! This method is insane" | Title accurately describes content |
| 3 | Heading skip | O01 | H1 jumps to H3, no H2 | H1→H2→H3 sequential nesting |
| 4 | Wall of text | O06 | Single 300-word paragraph | 3–5 sentences per paragraph, by topic |
| 5 | No summary | O02 | No TL;DR or key takeaways | Key Takeaways box at article top |
| 6 | Vague numbers | R01 | "very fast", "very cheap" | "120ms response", "$29/mo" |
| 7 | Unsourced stats | R02 | "Studies show 90% of users..." | Include source link and year |
| 8 | Stale pricing | R06 | 2023 prices shown, no date | "Pricing as of 2026-02" |
| 9 | Broken links | R10 | External links to 404 pages | Regular link audits; fix or remove |
| 10 | No unique value | E01 | Restates official docs verbatim | Add original test data or comparisons |
| 11 | No first person | Exp01 | Entirely third-person tone | Add "I tested" or "We found" |
| 12 | Only positives | Exp06 | Zero negatives; reads like an ad | Honestly share ≥2 downsides |
| 13 | No author info | Ept01 | No byline, avatar, or bio | Name + photo + 30-word bio |
| 14 | Undisclosed affiliate | T04 | Affiliate links without disclosure | Disclosure banner at page top |
| 15 | Missing FAQ Schema | O05 | Q&A content without JSON-LD | Add FAQPage Schema markup |
When unsure which dimension a check item belongs to:
| Rule | Assign To |
|---|---|
| Visible in the article or page | CORE |
| Requires checking author, org, or site | EEAT |
| "Can this sentence be verified?" | CORE-R |
| "Is this entity worth trusting?" | EEAT-T |
| "Does this content add unique value?" | CORE-E |
| "Does the value come from personal experience?" | EEAT-Exp |
| Pair | Disambiguation |
|---|---|
| R (Citation Density) vs A (Backlinks) | R = sources YOU cite (outbound); A = others citing YOU (inbound) |
| CORE-E (Exclusivity) vs Exp (Experience) | E = is the content unique (information gain); Exp = did the creator personally experience it |
| R (Timestamp) vs T (Correction Policy) | R = is info current (freshness); T = is there a maintenance mechanism (governance) |
| O (Schema) vs R (HTML Semantics) | O = page-level JSON-LD; R = semantic tags in content (<article>, <time>) |
| R (Method Transparency) vs Ept (Method Rigor) | R = was the method documented (transparency); Ept = is the method sound (quality) |
| Ept (Reasoning) vs R (Evidence Mapping) | Ept = is reasoning explainable; R = does each claim have an evidence chain |