SQRG (Search Quality Rater Guidelines)

Originally published at thatdevpro.com. This framework reference is part of the 14-tier Engine Optimization stack from ThatDevPro, an SDVOSB-certified veteran-owned web + AI engineering studio. You are reading the dev.to mirror; the source-of-truth canonical version with embedded validation tools lives at the link above.

Google's 182-Page Quality Rater Manual — How Human Raters Actually Evaluate Content

A comprehensive installation and audit reference for understanding how Google's human Search Quality Raters evaluate websites, and structuring a website to score well against the Page Quality (PQ) and Needs Met (NM) rating scales raters use. This document is dual-purpose: installation manual and audit document.

Cross-stack implementation note: the code samples in this framework are written in plain HTML for clarity. For React, Vue, Svelte, Next.js, Nuxt, SvelteKit, Astro, Hugo, 11ty, Remix, WordPress, Shopify, and Webflow equivalents of every pattern below, see framework-cross-stack-implementation.md. For pure client-rendered SPAs (no SSR/SSG) see framework-react.md. For Tailwind-specific concerns (purge, dynamic classes, dark-mode CLS, focus accessibility) see framework-tailwind.md.

1. Document Purpose & How to Use This Document

1.1 What This Document Is

This is the canonical reference for the Search Quality Rater Guidelines (SQRG) — Google's internal manual used by thousands of contracted human evaluators worldwide to assess search result quality. The full document is publicly available at services.google.com/fh/files/misc/hsw-sqrg.pdf. It runs 182 pages, was last updated September 11, 2025, and is the most authoritative public document on what Google considers high-quality content.

This framework document distills the SQRG into actionable installation and audit procedures. Where the SQRG describes what raters look for, this document specifies what to put on a website to give those raters (and the algorithms trained on rater data) what they're scoring.

The SQRG itself is not a ranking algorithm. Quality raters do not directly affect rankings — Google has explicitly stated this. But rater feedback trains and validates the algorithms that do affect rankings. Sites that score well against SQRG criteria are sites that score well in Google's actual ranking systems. The SQRG is the most direct window into Google's quality philosophy.

1.2 Three Operating Modes

Mode A — Install Mode: Building SQRG-aligned infrastructure into a website. Follow Sections 2 → 14.

Mode B — Audit Mode: Self-rating the site as if you were a Google quality rater. Skip to Section 11.

Mode C — Hybrid Mode: Audit then install for failing items.

1.3 How Claude Code CLI Should Consume This Document

1. Read Section 2 — collect client variables
2. Read Section 3 — understand the SQRG structure (PQ ratings + NM ratings + EEAT + YMYL all integrate here)
3. Apply Section 4 — Page Quality rating criteria for site-wide and per-page evaluation
4. Apply Section 5 — Needs Met rating criteria for query-specific evaluation
5. Install per Sections 6-9
6. Validate per Section 11
7. Generate report per Section 14

1.4 Conflict Resolution Rules

1.5 Required Tools

• The actual SQRG — read services.google.com/fh/files/misc/hsw-sqrg.pdf cover-to-cover
• Search results comparison — to evaluate Needs Met, compare your page against top 10 results for the query
• Multiple devices — raters evaluate on mobile and desktop; both must score well
• Honest internal raters — people who can score your content harshly without bias

2. Client Variables Intake

# ============================================
# SQRG FRAMEWORK CLIENT VARIABLES
# ============================================

# --- Business & Site Identity (REQUIRED) ---
business_name: ""
primary_domain: ""
site_purpose: ""                     # The single sentence that describes why this site exists
site_purpose_obvious_to_raters: false # Could a rater determine the site's purpose in 30 seconds?

# --- Page-Level Identity (REQUIRED for audit) ---
audited_page_url: ""
page_purpose: ""                     # What is this specific page trying to accomplish?
page_purpose_obvious: false
page_main_content_quality: ""        # "high", "medium", "low" (be honest)
page_target_query: ""                # What query is this page intended to satisfy?

# --- Site-Wide Content Quality (REQUIRED) ---
percentage_pages_high_quality: 0     # Percentage of indexed pages that would score "High" in PQ
percentage_pages_low_quality: 0      # Percentage that would score "Low" or "Lowest"
has_lowest_quality_pages: false      # If true, these need immediate attention
lowest_quality_examples: []          # URLs of pages that would score "Lowest"

# --- Critical SQRG Page Elements (REQUIRED) ---
has_clear_who_runs_site: false       # Can a rater determine who's behind the site?
has_clear_who_authors: false         # Can a rater determine who wrote each piece?
has_customer_service_info: false     # For e-commerce/transactional sites
has_clear_contact_info: false
has_purpose_statement_on_about_page: false

# --- Reputation Research (REQUIRED) ---
business_search_reputation: ""       # What appears when a rater searches "{{brand}} reviews"?
business_external_reviews_summary: ""  # Aggregate sentiment from Google, BBB, Trustpilot
known_negative_reputation_issues: []
reputation_research_completed_date: ""

# --- YMYL Status (REQUIRED) ---
ymyl_classification: ""              # "full", "partial", "lite", "non" — affects PQ thresholds
ymyl_categories: []

# --- E-E-A-T Status (REQUIRED — see framework-eeat.md) ---
eeat_self_assessment_score: 0        # Out of 130 from EEAT framework audit
eeat_critical_fails: 0

# --- Helpful Content System Status (REQUIRED — see framework-hcs.md) ---
hcs_self_assessment_score: 0         # Out of 54
hcs_critical_fails: 0

3. What the SQRG Is

The Search Quality Rater Guidelines is the manual Google provides to its thousands of contracted human evaluators worldwide. These evaluators ("quality raters" or "search quality evaluators") are paid contractors — typically fluent in their assigned language and trained extensively on the SQRG — who assess Google search results for quality. Their evaluations help Google understand whether ranking algorithm changes are improving or degrading search quality.

Critical clarification about how raters affect rankings: Raters do not directly change rankings. As Google's Danny Sullivan has stated: "Our systems aren't looking for [E-E-A-T]. Our raters are using that to see if our systems are working well to show good information." Rater data is feedback — like restaurant feedback cards — that helps Google adjust its algorithms. Aggregate rater feedback over time trains the ranking systems.

The SQRG covers two main rating tasks raters perform:

1. Page Quality (PQ) rating — How high-quality is this individual page? PQ uses a 5-level scale: Lowest, Low, Medium, High, Highest. Raters apply the PQ scale to evaluate any web page, regardless of search query.

2. Needs Met (NM) rating — How well does this page satisfy the user who issued a specific search query? NM uses a 5-level scale: Fails to Meet (FailsM), Slightly Meets (SM), Moderately Meets (MM), Highly Meets (HM), Fully Meets (FullyM). NM is query-specific.

The SQRG also covers:

• E-E-A-T evaluation criteria (see framework-eeat.md)
• YMYL classification and elevated standards (see framework-ymyl.md)
• Reputation research procedures
• Spam, deception, and harm evaluation
• AI-generated content evaluation criteria (added in late 2024 and expanded in September 2025)

The September 2025 SQRG update was significant. It:

• Added evaluation criteria for AI Overviews (parallel to existing guidance for featured snippets)
• Expanded YMYL to explicitly include elections, civic institutions, and government trust
• Updated guidance on AI-generated content evaluation
• Refined criteria for reputation research

The SQRG is publicly available. Reading it cover-to-cover is one of the most under-utilized SEO research activities. This framework document distills the operational implications, but the SQRG itself is the source of truth.

4. Page Quality (PQ) Rating

The PQ rating scale is what raters apply to evaluate any individual page. Sites should score "High" or "Highest" on every important page.

4.1 The Five PQ Levels

Lowest — The page should not exist in search results. Reasons to rate Lowest:

• Harmful or hateful content
• Highly inaccurate or deceptive content
• YMYL content from creators clearly lacking expertise
• Content that demonstrates malicious intent
• Pages with the explicit purpose of spreading misinformation
• Pages that financially harm users (scams)
• Pages with extreme inadequacy of MC quality
• Pages that are hacked, deceptive, or completely unable to fulfill their stated purpose

Low — The page has significant problems. Reasons to rate Low:

• Inadequate MC for the page's purpose
• Untrustworthy creators
• Negative reputation for the site or creator
• Significantly distracting or disruptive ads
• Inadequate level of E-E-A-T for the topic
• Misleading or exaggerated titles
• Content that doesn't match the title
• For YMYL: insufficient expertise, accuracy concerns

Medium — The page is acceptable but unremarkable. Could be either:

• Mixed signals — some good, some bad
• Solid foundational quality but nothing exceptional
• Generic, unremarkable content that does the job

High — The page is well-made and trustworthy. Indicators:

• Satisfying amount of high-quality MC
• Demonstrates E-E-A-T appropriate for the topic
• Positive reputation of website and content creators
• Helpful and clear information about the website
• Functional design with good user experience
• Appropriate ads that don't interfere with MC
• For YMYL: appropriate expertise demonstrated

Highest — The page is exceptional. Indicators:

• Very high level of MC quality
• Exceptional level of E-E-A-T
• Very positive reputation
• Demonstrably the best on the topic for the query
• Original, substantial, comprehensive content
• Clear, prominent expertise of creators
• For YMYL: very high level of expertise, accuracy, and trust

4.2 What Raters Examine for PQ

For every page being PQ-rated, raters investigate:

4.2.1 Page purpose

Raters first determine: what is this page trying to do? Inform? Sell? Entertain? Help? Express opinion? Share experience?

The page's purpose must be obvious within seconds. Pages with unclear purpose fail PQ regardless of MC quality.

4.2.2 Main Content (MC) quality and quantity

MC is the content the page exists to deliver — the article body, product description, video, tool, etc. Raters evaluate:

• Is there enough MC for the purpose?
• Is the MC accurate?
• Is the MC original or unique value?
• Is the MC well-produced (not sloppy)?
• Does the MC demonstrate effort, talent, accuracy?
• Does the MC fulfill the purpose well?

For YMYL pages, MC quality requires demonstrably accurate information from credentialed sources.

4.2.3 Supplementary Content (SC) quality

SC is everything else on the page — navigation, related links, ads, comments, sidebars. Raters evaluate:

• Does SC support the MC purpose?
• Is SC distracting from MC?
• Are ads disruptive?
• Are pop-ups annoying?

4.2.4 Information about the website

Raters research the website itself. Required elements that boost PQ:

• About page that explains the site's purpose
• Contact information (real, verifiable)
• Customer service information for transactional sites
• Author/creator information
• Editorial process if applicable
• Owner identity transparency

Sites with no information about who runs them fail PQ research.

4.2.5 Information about authors/creators

Raters specifically research the people behind content:

• Who wrote this?
• What's their expertise?
• Is the byline real?
• What's their reputation?
• For YMYL: do they have appropriate credentials?

4.2.6 Reputation research

Raters perform off-site research about the site and creators:

• Search "{{brand}} reviews" — what do users say?
• Check BBB, Trustpilot, Sitejabber, Google reviews
• For YMYL: check professional licensing, regulatory actions
• News about the site or creators (positive or negative)
• Industry recognition

Negative reputation can reduce PQ even if the page itself is well-made. Strongly negative reputation can drop PQ to Low or Lowest.

4.2.7 Page experience signals

Raters consider:

• Mobile usability
• Page load speed
• Intrusive ads or interstitials
• Layout shifts
• Functional links and features
• Working forms
• Working search

Page experience problems can reduce PQ rating.

4.3 PQ-Boosting Implementation Requirements

To score High or Highest on PQ, ensure:

4.3.1 Every page has obvious purpose

The H1 and first paragraph must make the page's purpose obvious. Examples:

• Service page: "We help [audience] [accomplish goal] through [our service]"
• Article: "This guide covers [specific topic] for [specific audience]"
• Product page: "[Product name] is [what it is] for [who]"
• Tool: "Use this tool to [accomplish specific task]"

If the purpose isn't obvious in 5-10 seconds, the page fails this criterion.

4.3.2 MC is substantive and adequate for purpose

MC must match the page's stated purpose:

• Comprehensive guide → comprehensive content
• Product page → complete product information
• Service page → clear service description
• News article → complete news reporting
• Tool → functional tool

Thin MC for a substantial purpose claim is a PQ failure.

4.3.3 Author and creator transparency

Every content page has:

• Visible author byline
• Link to author bio page
• Author credentials displayed
• Real photo (not stock, not generic icon)
• Verifiable identity

See framework-eeat.md Section 4.2 for full author implementation.

4.3.4 Information about the website

Required pages, all linked from footer:

• /about/ with site purpose, owner identity, history
• /contact/ with real contact information
• /authors/ (if multi-author site)
• Privacy policy
• Terms of service
• Customer service info if transactional

See framework-eeat.md Section 5.2 for full requirements.

4.3.5 Reputation infrastructure

Take active steps to ensure reputation research returns positive results:

• Maintain Google Business Profile with reviews
• Active management of Trustpilot, BBB, industry-specific review sites
• Press kit and earned media easily found
• Wikipedia/Wikidata presence if notable enough
• Respond to all reviews (positive and negative) professionally
• Monitor brand mentions and address misinformation

See framework-eeat.md Section 4.3 (Authoritativeness) for reputation infrastructure.

4.3.6 Functional design and clean experience

• Mobile-responsive design
• Fast loading (Core Web Vitals "Good")
• No intrusive ads or interstitials
• Working features (forms, search, navigation)
• Clean, professional appearance
• No broken images or layout issues

5. Needs Met (NM) Rating

NM is query-specific. For each query, raters evaluate whether each result meets the user's needs.

5.1 The Five NM Levels

Fully Meets (FullyM) — The result completely satisfies the user's specific need. Reserved for queries with a clear, single best answer that the page provides perfectly. Example: query "weather in San Francisco today" → result that shows San Francisco weather.

Highly Meets (HM) — The result is very helpful and likely to be what most users want. The result satisfies a primary user intent.

Moderately Meets (MM) — The result is helpful for some users, satisfies a secondary intent, or partially satisfies the primary intent.

Slightly Meets (SM) — The result is somewhat related to the query but doesn't really answer it, or answers a less likely interpretation.

Fails to Meet (FailsM) — The result doesn't satisfy the query at all. Off-topic, inaccurate, foreign-language, or otherwise irrelevant.

5.2 What Raters Consider for NM

5.2.1 User intent interpretation

Raters first determine likely user intents for the query. Most queries have multiple possible intents:

• Query "apple" could mean: the company, the fruit, applications, etc.
• Query "best running shoes" could mean: research mode, purchase mode, specific scenario (trail, road, marathon)
• Query "tax deadline" could mean: federal, state-specific, business, individual

Pages that satisfy the most likely intent score higher.

5.2.2 Result usefulness

Beyond relevance, raters consider:

• Is this result useful right now, on this device?
• Is the page accessible to this user?
• Is the answer clear or buried?
• Does the page require unnecessary additional action?

A page that buries the answer 2000 words deep when the user just needs a quick fact rates lower than a page that surfaces the answer immediately.

5.2.3 Query freshness needs

Raters consider whether the query implies need for fresh content:

• News queries → freshness matters
• "Best laptop 2026" → recent content rates higher
• Evergreen topics → freshness less important

Pages with dates that match the query's freshness expectation score higher.

5.2.4 Localization

For local-intent queries, results in the user's location score higher.

5.2.5 YMYL stricter NM

For YMYL queries, raters apply higher standards. A result might be relevant but rate lower if it's from an untrustworthy source on a YMYL topic.

5.3 NM-Optimization Implementation

5.3.1 Match content to genuine user intent

For every published page, document:

• Primary intent the page serves
• Secondary intents it can also satisfy
• Intents it does NOT satisfy

Don't create pages for intents you can't actually satisfy.

5.3.2 Surface the answer prominently

If users searching for the page's target query need a specific answer, deliver it in the first paragraph or hero section. Then expand below.

<article>
  <h1>{{HEADLINE_MATCHING_QUERY_INTENT}}</h1>

  <section class="answer-summary">
    <p>{{DIRECT_ANSWER_TO_QUERY_IN_2_3_SENTENCES}}</p>
  </section>

  <section class="full-detail">
    {{COMPREHENSIVE_TREATMENT}}
  </section>
</article>

5.3.3 Match freshness to intent

For queries with freshness needs:

• Update content regularly
• Display update dates prominently
• Date-stamp time-sensitive sections

For evergreen queries:

• Don't fake freshness
• Focus on comprehensive depth

5.3.4 Multi-intent pages handled correctly

For broad queries with multiple reasonable intents, address them with clear navigation:

<nav aria-label="Choose your path">
  <h2>What are you looking for?</h2>
  <ul>
    <li><a href="#beginners">If you're new to {{TOPIC}}</a></li>
    <li><a href="#advanced">If you're experienced with {{TOPIC}}</a></li>
    <li><a href="#troubleshooting">If you're troubleshooting</a></li>
  </ul>
</nav>

5.3.5 Mobile-first NM

Most rater work happens on mobile. The mobile experience must:

• Surface the answer immediately
• Have tap targets meeting 48px minimum
• Load quickly on cellular
• Render correctly on small screens

6. PQ-Optimization Site-Wide Installation

6.1 Phase 1: Site Identity Foundation

6.1.1 Homepage that demonstrates purpose

The homepage must communicate within 5 seconds:

• Who runs the site
• What the site does
• Who it's for
• Why visitors should trust it

<main>
  <section class="hero">
    <h1>{{SITE_PURPOSE_AS_VALUE_PROPOSITION}}</h1>
    <p class="subhead">{{WHO_ITS_FOR_AND_WHAT_THEY_GET}}</p>
  </section>

  <section class="who-we-are-summary">
    <h2>Who We Are</h2>
    <p>{{SHORT_INTRODUCTION_OF_FOUNDER_OR_TEAM}}</p>
    <p><a href="/about/">More about us →</a></p>
  </section>

  <section class="what-we-do">
    <h2>What We Cover</h2>
    <ul>{{LIST_OF_PRIMARY_TOPICS_OR_SERVICES}}</ul>
  </section>

  <section class="trust-indicators">
    {{REVIEW_AGGREGATE}}
    {{FEATURED_IN_LOGOS}}
    {{KEY_CREDENTIALS}}
  </section>
</main>

6.1.2 Comprehensive About page

See framework-hcs.md Section 6.3 and framework-eeat.md Section 5.2.1 for full About page requirements. SQRG specifically rewards comprehensive About pages.

6.1.3 Author hub and individual author pages

See framework-eeat.md Section 7.

6.1.4 Customer service infrastructure for transactional sites

E-commerce or service sites need:

• /customer-service/ page with:
  • Hours of operation
  • Multiple contact methods
  • Response time expectations
  • Common questions answered
• Live chat or contact form on every transactional page
• Order tracking, return policy, refund policy linked from footer
• Real human responsiveness (24-48 hour response SLA)

6.2 Phase 2: Reputation Infrastructure

6.2.1 Active review management

• Claim Google Business Profile
• Set up Trustpilot, BBB, industry-specific review platforms
• Implement review acquisition workflow (request reviews from satisfied customers)
• Respond to every review within 48 hours
• Address negative reviews professionally with offer to resolve

6.2.2 Earned media and press

See framework-eeat.md Section 4.3 (Authoritativeness pillar) for full implementation.

6.2.3 Wikipedia and Wikidata presence

For sites with sufficient notability, presence on Wikipedia and Wikidata is one of the strongest reputation signals raters check. See framework-eeat.md Section 4.3.4.

6.2.4 Industry recognition

• Submit to relevant industry awards
• Pursue certifications and accreditations
• Speaking opportunities at industry conferences
• Guest articles on industry-leading publications

6.3 Phase 3: Content Quality at Scale

6.3.1 No "Lowest" pages allowed

Audit the site for any pages that would rate "Lowest":

• Hacked or compromised pages
• Pages with malware
• Deceptive or scammy content
• Inaccurate YMYL content
• Pages that cannot fulfill their stated purpose

Remediate or remove all "Lowest" pages immediately. Their existence drags site-wide PQ.

6.3.2 No "Low" pages without remediation plan

Audit for pages that would rate "Low":

• Inadequate MC for the page's purpose
• Significantly distracting ads
• Misleading titles
• Untrustworthy creators

Either remediate to bring up to Medium or higher, or remove.

6.3.3 Aim for High or Highest on important pages

Top-traffic pages, money pages, and topic pillar pages should all aim for Highest. Apply:

• Original research and insights (Information Gain — see framework-infogain.md)
• Comprehensive coverage
• Clear E-E-A-T demonstration
• Excellent page experience
• Substantial MC

6.4 Phase 4: Page Experience

Per Section 4.2.7 of this document and Tier 1 of the 14-tier framework:

• Mobile-responsive
• Core Web Vitals "Good"
• No intrusive ads or interstitials
• Functional features
• Clean, professional design
• Working forms, search, navigation

6.5 Phase 5: AI Content Evaluation Per SQRG

The September 2025 SQRG update added specific guidance for AI-generated content. Key principles:

• AI involvement is not automatically "Lowest" — quality of output matters more than method of creation
• AI content presented as expert-authored without disclosure can be rated "Lowest"
• AI-generated YMYL content without expert review is rated lower
• Mass-produced AI content without genuine value is rated "Low" or "Lowest"

Implementation:

• Disclose AI use clearly per framework-eeat.md Section 4.4.6
• Apply expert review per framework-ymyl.md Section 5.8 for YMYL content
• Don't pass AI content off as expert-authored
• Quality every published piece through human review

7. NM-Optimization Per-Page Installation

For each indexable page, configure for the queries it's targeting.

7.1 Document Target Query and Intent

For every page, document in the CMS or content management system:

• Primary target query
• Primary user intent (informational, commercial, navigational, transactional)
• Secondary queries served
• Freshness requirement (high, medium, low, none)
• Local relevance (yes/no)

7.2 Match Page Structure to Intent

7.2.1 Informational queries

User wants knowledge. Page structure:

• Direct answer in first paragraph
• Comprehensive treatment below
• Sources cited
• Author with relevant credentials
• Related questions answered

7.2.2 Commercial queries

User is researching purchase. Page structure:

• Comparison or evaluation framework
• Specific recommendations with reasoning
• Pros and cons
• Real testing or experience
• Pricing information
• Where to buy / how to get

7.2.3 Transactional queries

User wants to take action. Page structure:

• Action-oriented hero (Buy, Sign Up, Get Started)
• Trust signals near CTA
• Pricing clear
• Process clear
• Customer service accessible
• Reviews/testimonials visible

7.2.4 Navigational queries

User wants a specific destination. Page structure:

• Confirms they're in the right place
• Directs to specific resource quickly
• Provides search or sitemap if not exact match

7.2.5 Local queries

User wants location-specific result. Page structure:

• Location prominent
• Address, hours, phone
• Map embedded
• Get directions CTA
• Local-specific content (not generic)

7.3 Freshness Signal Calibration

For pages serving freshness-sensitive queries:

• Display dateModified prominently
• Note specific time-sensitive elements ("As of {{MONTH_YEAR}}")
• Update on a cadence matching the freshness need
• Don't game date refreshing — see framework-hcs.md Section 9.6

7.4 Mobile-First Page Design

Every page must work well on mobile because raters work on mobile:

• Tap targets ≥ 48px
• Font size ≥ 16px body
• Single-column layout for content
• Hero CTA visible above fold on small screens
• No horizontal scrolling
• Forms work with mobile keyboards
• Phone numbers are clickable (tel: links)

8. Reputation Research Optimization

Raters perform off-site research about every site they evaluate. The goal: ensure their research finds positive, supporting information.

8.1 Brand Search Audit

Search Google for these queries about your business and document what comes up:

• {{business_name}}
• {{business_name}} reviews
• {{business_name}} complaints
• {{business_name}} {{primary_service}}
• is {{business_name}} legit
• {{founder_name}}
• {{founder_name}} {{business_name}}

Top 10 results for each should be:

• Owned (your website, official social profiles)
• Favorable (positive reviews, earned media)
• Neutral at minimum (industry directory listings)

If negative results appear in top 10:

• Address the underlying issue if legitimate
• Drown out via positive content production if outdated/unfair
• Pursue removal if defamatory or false

8.2 Review Platform Coverage

Maintain active presence on:

• Google Business Profile — most important for local
• Trustpilot — increasingly important for AI engines
• BBB — important for trust signals
• Industry-specific — Clutch (B2B), Healthgrades (medical), Avvo (legal), etc.
• Sitejabber — general consumer reviews
• G2 / Capterra — for B2B SaaS

Average rating target: 4.5+. Volume target: 100+ reviews per platform where applicable. Response rate: 100%.

8.3 Wikipedia and Wikidata

For notable businesses and individuals:

Wikidata (lower notability bar):

• Create entry at https://www.wikidata.org/wiki/Special:NewItem
• Populate properties: instance of, founded by, founding date, official website, headquarters, sameAs links
• Add references to verifiable sources

Wikipedia (higher notability bar):

• Requires multiple independent secondary sources
• Don't create your own Wikipedia article (conflict of interest)
• Wait for organic creation, or work with experienced Wikipedia editor

8.4 Earned Media Inventory

Document all earned media at /press/ page:

• Logos of publications that have featured the business
• Direct links to each piece of coverage
• Dates and quotes from coverage
• Press kit downloadable
• Journalist contact form

8.5 Negative Reputation Management

If known negative reputation exists:

Legitimate complaints:

• Address the underlying issue
• Improve processes that caused complaints
• Document remediation in updated content

Outdated negative information:

• Produce more current positive content
• Request updates from publishers where reasonable
• Drown out with high-volume positive material

Defamatory or false content:

• Document evidence of falsity
• Send legal demand for removal where appropriate
• Pursue legal remedy if necessary
• Engage reputation management firm if pattern is severe

9. AI Content Per SQRG (September 2025 Update)

The September 2025 SQRG update added specific guidance for evaluating AI-generated content. Key implementation requirements:

9.1 Disclosure Requirements

Per framework-eeat.md Section 4.4.6 — site-wide AI policy + per-article disclosure callouts.

9.2 Expert Review Requirements

For YMYL AI content: per framework-ymyl.md Section 5.8.

For non-YMYL AI content: still requires human review, factual verification, and bylined responsibility.

9.3 Quality Threshold

AI content must meet the same quality bar as human-written content:

• Original insight, not generic synthesis
• Specific examples, not abstract treatments
• Verifiable claims with citations
• Edge cases and nuance
• Demonstration that the publishing entity has genuine expertise

9.4 What's Never Acceptable

• AI content with fake bylines
• AI content fabricating quotes, statistics, or sources
• AI content on YMYL topics without credentialed reviewer
• AI mass-production strategies (publishing dozens of articles per day with minimal review)

10. PQ Audit Criteria

Score the site against PQ rating criteria:

Score: 40 max. World-class PQ: 36+/40 with zero Critical fails.

11. NM Audit Criteria

For sample pages, evaluate against query-specific NM criteria:

Score per page: 16 max. World-class NM: 14+/16 with zero Critical fails.

12. Common SQRG Mistakes & Anti-Patterns

12.1 Generic About Pages

Anti-pattern: "We are a leading provider of solutions for businesses..." — generic, says nothing specific.

Why it fails: Raters research the site. Generic about pages signal a lack of genuine identity.

Fix: Specific, concrete, real. Who are you actually? What do you actually do? Why?

12.2 Hidden Owner Identity

Anti-pattern: Site doesn't say who runs it. Footer just shows business name with no individual.

Why it fails: SQRG specifically requires transparency about who's behind the site. Anonymity is a Trust failure.

Fix: Real names, real photos, real bios.

12.3 No Customer Service for Transactional Sites

Anti-pattern: E-commerce site with no phone number, no live chat, just a contact form going to a black hole.

Why it fails: SQRG explicitly requires customer service info for transactional sites. Lack of it can drop PQ to Low.

Fix: Real customer service, real responsiveness, accessible from every transactional page.

12.4 Misleading or Exaggerated Titles

Anti-pattern: "Ultimate Guide" that's 800 words. "Complete Tutorial" that covers basics only. "Best X in 2026" that's just a list.

Why it fails: Title-content mismatch is explicit PQ failure criteria.

Fix: Title matches content. If you can't deliver "ultimate," don't claim it.

12.5 Negative Reputation Ignored

Anti-pattern: Negative BBB rating, low Google reviews, complaints in industry forums — none addressed.

Why it fails: Reputation research drives PQ. Negative reputation drops PQ regardless of on-page quality.

Fix: Active reputation management. Address legitimate complaints. Drown out outdated negatives with current positives.

12.6 YMYL Without Expertise

Anti-pattern: Health blog written by non-credentialed bloggers, finance advice from unlicensed authors, legal information from non-attorneys.

Why it fails: SQRG holds YMYL content to expert standards. Lacking expertise is automatic Low or Lowest rating.

Fix: Either get credentialed authors and reviewers, or reframe content as personal experience (with proper disclaimers), or stop publishing in that category.

12.7 Site-Wide Quality Drag

Anti-pattern: Site has good content but also has hundreds of thin SEO experiments, abandoned categories, expired content.

Why it fails: PQ is evaluated holistically. Low-quality pages drag the entire site's evaluation.

Fix: Aggressive content audit. Remove or remediate all Low and Lowest pages. Quality > quantity.

12.8 Intrusive Ad Patterns

Anti-pattern: Pop-ups on entry, full-screen interstitials before content, ads that move during reading, pre-roll ads before short content.

Why it fails: Page experience explicitly impacts PQ. Intrusive ads drop PQ rating.

Fix: Ads that don't disrupt MC. Modest ad density. No interstitials before MC.

12.9 No Reputation Effort for Notable Businesses

Anti-pattern: Established business with no Wikipedia, no Wikidata, no consistent earned media coverage.

Why it fails: Notable businesses should have reputation infrastructure. Absence is suspicious.

Fix: Actively build reputation infrastructure per Section 8.

12.10 AI Content Pretending to Be Expert-Authored

Anti-pattern: AI-generated articles bylined to a person who didn't write them. No disclosure of AI involvement.

Why it fails: SQRG's September 2025 update explicitly addresses this. Detection results in Lowest rating.

Fix: Disclose AI use. Have credentialed reviewer sign off. Don't pretend.

13. Maintenance Schedule

13.1 Monthly

• Review newly added content for PQ markers
• Check brand search results for new negative content
• Monitor review platforms for new reviews; respond to all

13.2 Quarterly

• Site-wide PQ audit (sample 5% of pages, score each)
• Reputation research refresh — search brand queries, document state
• Update About page if business has evolved
• Audit AI content compliance if AI is in use

13.3 Annually

• Read the most recent SQRG cover-to-cover
• Comprehensive PQ self-audit across all major page types
• Strategic review of identified low-PQ patterns
• Update reputation infrastructure (Wikipedia, Wikidata, press kit)

13.4 On SQRG Update

When Google publishes an updated SQRG (typically annually):

1. Read the update document
2. Identify new criteria or changed criteria
3. Update this internal framework document accordingly
4. Audit sites against new criteria
5. Implement changes within 30 days

14. Implementation/Audit Report Templates

14.1 SQRG Implementation Report Template

# SQRG Framework Implementation Report

**Site**: {{BUSINESS_NAME}}
**Implementation Date**: {{TODAY}}

## Summary
- Self-rated PQ score: {{X}}/40
- Estimated PQ rating: {{HIGHEST/HIGH/MEDIUM/LOW/LOWEST}}
- YMYL pages reviewed: {{COUNT}}
- "Lowest" pages identified and remediated: {{COUNT}}
- "Low" pages identified and on remediation path: {{COUNT}}

## Site Identity Improvements
{{LIST}}

## Author Identity & Credentials Improvements
{{LIST}}

## Reputation Infrastructure Status
{{LIST}}

## Customer Service Status (if transactional)
{{LIST}}

## AI Content Compliance Status
{{STATUS}}

## Validation Results
{{AUTOMATED_AND_MANUAL_RESULTS}}

## Sign-Off

14.2 SQRG Audit Report Template

# SQRG Framework Audit Report

**Site**: {{BUSINESS_NAME}}
**Audit Date**: {{TODAY}}

## Executive Summary
{{SUMMARY}}

**Estimated Site PQ Rating**: {{RATING}}
**PQ Score**: {{X}}/40
**NM Score (Sample)**: {{X}}/16 average

## PQ Findings
### Site Identity ({{PASS/FAIL}})
{{FINDINGS}}

### MC Quality Distribution
- Highest: {{COUNT}} pages
- High: {{COUNT}} pages
- Medium: {{COUNT}} pages
- Low: {{COUNT}} pages
- Lowest: {{COUNT}} pages

### Reputation Research Findings
- Top 10 brand search results: {{POSITIVE/MIXED/NEGATIVE}}
- Review platform coverage: {{ASSESSMENT}}
- Earned media: {{ASSESSMENT}}

### Page Experience
{{ASSESSMENT}}

## NM Findings (Sample Pages)
{{PER_PAGE_FINDINGS}}

## Critical Failures
{{LIST_WITH_REMEDIATION}}

## Recommended Remediation Order
{{PRIORITIZED_LIST}}

## Sign-Off

End of Framework Document

Document version: 1.0
Last updated: 2026-04-29
Maintained by: ThatDeveloperGuy

The SQRG is the most comprehensive public document on what Google considers high-quality content. This framework distills it into operational requirements, but the SQRG itself should be read directly. The PDF is updated periodically — typically annually — and the most recent version (September 2025) added important guidance on AI Overviews and expanded YMYL to include civic content.

The SQRG is the framework that connects E-E-A-T, YMYL, and Helpful Content. Where those frameworks describe specific dimensions of quality, the SQRG describes how human raters synthesize all of them into Page Quality and Needs Met evaluations.

Companion documents:

• framework-eeat.md
• framework-ymyl.md
• framework-hcs.md
• framework-coreupdates.md
• framework-infogain.md
• framework-entitysalience.md
• framework-knowledgegraph.md
• framework-aicitations.md

From the ThatDevPro Engine Optimization framework library. Studio: ThatDevPro (SDVOSB veteran-owned web + AI engineering). Sister property: ThatDeveloperGuy. Source: https://www.thatdevpro.com/insights/framework-sqrg/.