Methodology

Apex Sentinel Scoring v2.0 is rubric-based and grounded in published industry standards. Every category uses deterministic signals from sources we actually collect (no invented numbers). Where a signal isn't yet available, we say so explicitly — null is preferred over fake data. Weights come from cited industry benchmarks (Google documentation, Lighthouse scoring, AWR CTR study, Whitespark LSRF, Google Quality Rater Guidelines).

Core principles:

No arbitrary weights — every weight traces to a cited source or is explicitly "emerging".
null beats fake data. Missing signal = "—", not zero.
Field data preferred over lab data (RUM > Lighthouse).
Threshold-based where industry has thresholds (CrUX buckets for CWV).
Public methodology — this page is the single source of truth, editable in git.

Technical

Server errors, SSL/TLS, broken links, redirect chains, robots.txt, sitemap, canonical tags. Penalty-based: start at 100, deduct per Sentinel finding from the last 7 days (P0=−25, P1=−10, P2=−3). Aligned with Google's Technical SEO fundamentals and Lighthouse SEO score thresholds (90+ = good).

Weight breakdown

HTTP 5xx rate	25%
SSL/TLS grade (Qualys SSL Labs)	15%
Broken internal links	15%
robots.txt validity	10%
Sitemap validity + freshness	10%
Canonical tag presence	10%
Redirect chain depth	10%
Mixed content	5%

Data sources (in use)

Sentinel crawler tiers 1 + 2 + 5
SSL Labs Qualys grade
Google Technical SEO guidelines

Industry references

Content

Content quality proxied via real search engagement + first-party editorial hygiene. CTR-vs-expected uses AWR's 2024 CTR-by-position benchmarks (position 1 ≈ 31.7%, position 5 ≈ 9.5%). Editorial QA and content depth are measured by Sentinel's own crawler (title/meta/H1 presence + in-range %, median word count).

Weight breakdown

CTR / expected-CTR ratio (AWR 2024)	30%
Editorial QA (first-party title/meta/H1 in-range %)	25%
Query coverage breadth (queries ranking top 20)	15%
Impression growth 7d-vs-prior-7d	15%
Content depth (median word count)	15%

Data sources (in use)

Google Search Console
Sentinel site-crawl (first-party HTML parse)
AWR 2024 CTR-by-position benchmarks

Pending additions

Topic coverage / entity breadth (NLP)
Readability grade (Flesch-Kincaid)

Industry references

Authority

GSC-derived authority proxy pending Ahrefs API access. Strong correlations: low avg position across niche queries, many top-3 + top-10 rankings, and high organic click volume. Will blend in Ahrefs Domain Rating + referring domains when available.

Weight breakdown

Average rank position (lower = stronger authority)	30%
# of queries ranking in top 3	25%
# of queries ranking in top 10	25%
Total organic clicks (28d)	20%

Data sources (in use)

GSC queries history (28d)

Pending additions

Ahrefs Domain Rating
Moz Domain Authority
Referring-domain profile + spam score

Industry references

UX / Performance

Google's Core Web Vitals — the perf metrics Google uses for ranking. Field data (RUM pixel) preferred; falls back to Lighthouse lab below 100 samples. Thresholds are Google's own CrUX buckets: LCP <2500ms, CLS <0.1, INP <200ms = 'good' (score 100). Poor = 20.

Weight breakdown

LCP (Largest Contentful Paint) p75	40%
CLS (Cumulative Layout Shift) p75	30%
INP (Interaction to Next Paint) p75	30%

Data sources (in use)

Apex Sentinel RUM pixel (field)
Lighthouse lab (fallback)

Pending additions

Google CrUX API for fleet-wide p75 benchmark comparison

Industry references

Local SEO

Local-intent discoverability. Whitespark Local Search Ranking Factors methodology. Measures LocalBusiness / CannabisDispensary schema (first-party crawl), license display, and GSC performance on local-intent queries (queries containing 'near me', 'dispensary', or ' in '). NAP consistency proxy via schema + license alignment.

Weight breakdown

LocalBusiness / CannabisDispensary JSON-LD coverage (first-party crawl)	30%
Local-intent query avg rank (GSC)	30%
License displayed (compliance rule)	20%
NAP consistency proxy	20%

Data sources (in use)

Sentinel site-crawl (JSON-LD extraction)
Sentinel compliance scanner
GSC local-query analysis

Pending additions

Google Business Profile API
Local pack position tracker
Citation audit (BrightLocal / Whitespark)

Industry references

On-Page

Per-URL HTML + metadata fundamentals, measured first-party by Sentinel's crawler fetching every site URL and parsing the DOM with cheerio. Weighted by Google ranking-impact: title (20-60 char) + H1 discipline + schema + canonical lead. meta_keywords excluded (Google ignores since 2009).

Weight breakdown

Title in-range (20-60 chars)	20%
Single H1 per page	15%
JSON-LD present	15%
Canonical self-referencing	12%
Meta description in-range (50-160 chars)	10%
Viewport meta	8%
OG tags complete (title+desc+image)	8%
Twitter Card tag	5%
<html lang> attribute	4%
Images with alt text	3%

Data sources (in use)

Sentinel site-crawl (cheerio HTML parse — title/meta/H1/canonical/viewport/og/twitter/lang/img/JSON-LD)

Pending additions

Internal linking analyzer (anchor text breadth)

Industry references

Mobile

Mobile-specific experience (Google uses mobile-first indexing). Mobile RUM Core Web Vitals + WCAG 2.2 AA accessibility violations from axe-core. Mobile-LCP threshold is the dominant signal.

Weight breakdown

Mobile LCP p75 (RUM device=mobile)	30%
Mobile INP p75	25%
Accessibility violations (WCAG 2.2 AA)	20%
Mobile CLS p75	15%
Mobile JS errors (RUM)	10%

Data sources (in use)

RUM pixel (device=mobile)
Sentinel tier9 axe-core a11y audit
WCAG 2.2 AA ruleset

Pending additions

Dedicated mobile tap-target audit (WCAG 2.5.5)

Industry references

AI Readiness

How likely AI search (Perplexity, ChatGPT, Google AI Overview, Claude) is to cite your pages. Emerging discipline — measures first-party JSON-LD breadth, FAQ/QA + Speakable schema presence, and answer-first query ranking. Framework evolves as AEO best practices stabilize in 2025-2026.

Weight breakdown

Structured data breadth (first-party JSON-LD detection)	35%
Answer-first query rankings (question-form queries in top 10)	30%
FAQ/QAPage schema coverage	25%
Speakable schema coverage	10%

Data sources (in use)

Sentinel site-crawl (JSON-LD @graph walker)
GSC question-form query analysis

Pending additions

AI crawler allowlist check (GPTBot, ClaudeBot, PerplexityBot in robots.txt)
Inbound AI citation tracking (Perplexity / ChatGPT cite monitor)
Google AI Overview appearance tracker

Industry references

Schema %

Structured data deployment breadth + type diversity, measured first-party by fetching every page and extracting all <script type="application/ld+json"> blocks (including @graph arrays). Google Rich Results Test API pending — will add validation-pass-rate weighting when wired.

Weight breakdown

Any JSON-LD present %	70%
Distinct schema.org type diversity (normalized to 5 types)	30%

Data sources (in use)

Sentinel site-crawl (JSON-LD @graph walker — detects LocalBusiness, Organization, Breadcrumb, FAQPage, QAPage, HowTo, Speakable, Article, Product, Person, etc.)

Pending additions

Google Rich Results Test API
Schema.org spec validator

Industry references

E-E-A-T

Google's Quality Rater Guidelines (December 2022 update) — 4 equal-weighted pillars: Experience, Expertise, Authoritativeness, Trustworthiness. Trustworthiness has the richest signal set today (HTTPS, privacy/terms links, license display, contact visibility).

Weight breakdown

Trustworthiness (compliance passes: HTTPS, privacy, terms, license, age gate)	30%
Authoritativeness (GSC proxy — pending Ahrefs)	30%
Expertise (first-party: author byline 45% · Organization schema 35% · single-H1 20%)	20%
Experience (GA4 bounce rate — low bounce = users finding value)	20%

Data sources (in use)

Sentinel compliance scans (5 trust rules)
GSC queries (authority proxy)
Sentinel site-crawl (author byline + Organization JSON-LD + H1 discipline)
GA4 engagement

Pending additions

Original-imagery ratio
Ahrefs inbound authority (DR/DA)
Topical consistency NLP
Review aggregate (GBP/TrustPilot)

Industry references

AEO

Answer-engine / voice / zero-click readiness. First-party JSON-LD detection for FAQPage, QAPage, HowTo and Speakable schema plus GSC query-intent analysis. Emerging discipline — evolves as answer engines mature.

Weight breakdown

FAQ / QAPage schema coverage (first-party)	25%
Question-form queries ranking top 20	25%
Schema foundation (any JSON-LD coverage)	20%
Featured-snippet eligibility (position ≤ 1.5)	15%
HowTo schema coverage	8%
Speakable schema coverage	7%

Data sources (in use)

Sentinel site-crawl (FAQPage/QAPage/HowTo/Speakable JSON-LD detection)
GSC question-form query analysis

Pending additions

Voice-search query tracker
PAA (People Also Ask) presence scraper
Zero-click SERP-feature coverage tracker

Industry references

Version history: v2.0 (2026-04-19) — hard-swap from v1 composite to rubric-based scoring. v1 used hand-tuned weights with no citations; v2 is sourced to industry standards and transparent at every signal.