SEO & Web Development Wiki — Overview

Evolving dual-pillar synthesis. Updated as sources are ingested and each thesis develops.

Sources ingested: 61 Last updated: 2026-09-04

Pillar 1 — Web Development

Working Thesis

Web-dev “best practice” in 2026 is defined by survival under real-world constraints (traffic, security audits, accessibility law, performance budgets, team turnover), not framework choice. The engineering core — server-first rendering (React Server Components (RSC)/Rendering Strategies (CSR / SSR / SSG / ISR / RSC)), modular-monolith + edge architecture, field-measured Core Web Vitals, security-by-design, Web Accessibility (WCAG 2.2 AA), strict-typed maintainability, and CD — is independently corroborated: the vendor Pagepro guide is backed by Smashing Magazine’s measured case studies (e.g. Barry Pollard on Core Web Vitals). The pillar has broadened into design (Responsive & Intent-Based Design, Design Quality & Craft), process (AI-Assisted Development — Claude Code, Vercel-internal agents, Greg Isenberg’s agent-ops), and a third SSG (Astro 7.2 incremental static builds). Site Search is no longer single-approach: hosted-crawl (Site Search 360), build-time (Pagefind), and a decision rule for when to reach for Algolia.

Key Findings

1. Server-first + field-measured performance is settled practice; CWV is both a UX and ranking signal, and field ≠ lab (measure the users you actually have). — Core Web Vitals, Improving Core Web Vitals, A Smashing Magazine Case Study
2. Security, accessibility, and maintainability shift left into the pipeline (OWASP 2025, WCAG 2.2 AA, strict TS, DORA). Keyboard Navigation and Web Accessibility (WCAG 2.2 AA) are now double-sourced.
3. The frontier is the process, not just the artifact: designing for intent/fluidity (Responsive & Intent-Based Design) and building with AI agents (AI-Assisted Development, Agent Orchestration) are the emerging, less-settled edges.

Pillar 2 — SEO

Working Thesis

The SEO corpus is local-SEO-first and now has an official Google pole. The triad is documented in-wiki (Tips to improve your local ranking on Google): relevance, distance, prominence (links + reviews). Caleb’s Three-Step Method (Low-Competition Ranking Method) (GBP audit → landing page → verified citations) and Core 30 are operationalizations of that triad, not Google’s instructions. Independent practitioners (James Dooley, Joy Hawkins, Nico (AI Ranking), Jono Catliff, Ahrefs) corroborate the GBP ingredients and contest the branded packaging.

Named disagreements are sharper, not fewer:

• Duplicate Content (for Service-Area Pages) is three-way: official (not a penalty; canonicalize), Dooley+Joy (identical location pages outrank unique rewrites — Joy said this to Caleb on 2026-08-06), Caleb (unique Core 30 per location). Official spam policy separately bans thin city-gate / doorway farms.
• GBP categories: official “fewest, IS-a not HAS-a” (Guidelines for representing your business on Google) vs Caleb’s many secondaries / 20–30 services.
• GEO / AEO (Getting Recommended by AI): official Google gen-AI guide is retrieval of unique non-commodity pages; Ahrefs measures branded mentions (0.664) and freshness; Caleb’s 250 Authority Protocol remains an unsupported training-data implant. ChatGPT local uses Bing web results, not Bing Places (How does ChatGPT conduct local searches?).
• Brand & User Signals as Ranking Drivers: leak (An Anonymous Source Shared Thousands of Leaked Google Search API Documents with Me; Everyone in SEO Should See Them) shows click/Chrome attributes exist; stored ≠ weighted. Still contested as “#1 factor.”
• PBN as a line item in Agency Pricing (Scope-Based) vs official link-spam policy and Caleb’s own “PBNs harm” YouTube claims.

Key Findings

1. Established (now official + practitioner): the local triad; GBP completeness; reviews and links as prominence; people-first unique content; E-E-A-T (Experience, Expertise, Authoritativeness, Trust) is not a ranking factor; same-site duplicates are not a spam penalty; AI content is fine if people-first and spam if scaled without value. ChatGPT local reads Bing web results, not Bing Places for Business.
2. Contested: Duplicate Content (for Service-Area Pages) (three-way, including Joy-to-Caleb); Brand & User Signals as Ranking Drivers; GBP category count; whether Bing Places is a ChatGPT input (it isn’t, directly).
3. Corrected: 250 Authority Protocol remains an unsupported training-data claim against both Anthropic and Google’s 2026 GEO guide. ChatGPT-via-Bing is web results, not Places.

Caveat: still creator-heavy on tactics (Caleb course + YouTube). Official docs now grade those tactics. Branded frameworks (Core 30, super-citations, 250-protocol, 8-pass, three-step) remain creator-opinion/emerging.

Where They Intersect

• Rendering ↔ crawlability (Rendering Strategies (CSR / SSR / SSG / ISR / RSC)); Core Web Vitals as a ranking signal; Schema Markup (Structured Data) approached from technical SEO and local SEO; site structure = ranking strategy (Core 30 ↔ pillar-cluster Technical SEO from Architecture).
• AI as the shared frontier: Data Poisoning & LLM Backdoors/GEO / AEO (Getting Recommended by AI) (SEO) and AI-Assisted Development (web-dev) both grapple with LLMs — one for being recommended by them, one for building with them. The Anthropic study sits at the seam (AI-security fact correcting an SEO tactic).
• Craft/metrics tension: Design Quality & Craft’s metric-skepticism is a values counterweight to the SEO corpus’s rank/conversion maximization — a healthy internal tension to preserve, not resolve.

Open Questions

• Duplicate Content (for Service-Area Pages) — official “not a penalty” + Joy/Dooley “identical pages win” + Caleb unique-per-location + spam-policy doorways. Still the sharpest dispute; needs a controlled test, not more anecdotes.
• GBP category count — official fewest vs practitioner many. Suspension risk vs ranking-coverage incentive is untested in-wiki.
• Brand & User Signals as Ranking Drivers — leak proves collection, not weight.
• Algolia/Meilisearch how-to — decision rule is in; an implementation source is not.
• Web-dev design/process edges still thin vs the SEO pillar.
• File an updated analysis of the duplicate-content debate (the 2026-07-01 analysis predates official docs + the Joy/Caleb interview).