Introducing the Brass-SEO Research Index
Most SEO advice traces back to vendor studies. Semrush publishes a report. Ahrefs publishes a report. The claims travel from blog post to blog post, shedding their sources along the way. By the time the stat reaches you, good luck finding the original research.
Brass-SEO built something different.
Quick Navigation
- What the Research Index Is
- What's in Each Topic
- The Inclusion Bar
- Why This Matters for GEO
- Frequently Asked Questions
What the Research Index Is
Brass-SEO now maintains a public index of primary literature on SEO and generative-engine optimization at brass-seo.com/research. Five topics. Twenty-six entries. Every source is a peer-reviewed academic paper, an official specification from the IETF or W3C, or an independent benchmark from the HTTP Archive. No vendor white papers. No paywalled-only sources.
Each entry gets an answer capsule — what Brass-SEO draws from it and why — plus a direct link to the primary source. The research behind the recommendation is visible and checkable.
The index is free to access. No subscription required.
What's in Each Topic
Brass-SEO built the index around five topics that cover the evidence base for our recommendations.
AI Citation & Generative Engine Optimization has six entries covering the peer-reviewed research on how AI search engines select and cite sources. The foundational study — Aggarwal et al., published at ACM KDD 2024 — found that adding concrete numerical statistics to a page raised AI search visibility by up to 40% in their test matrix. A 2025 analysis of over 366,000 citations across ChatGPT, Perplexity, and Google AI Overviews found that generative engines cite a small, heavily concentrated set of established sources regardless of query type. A December 2025 study found that 37% of domains cited by LLM-based search engines don't appear in traditional search results for the same queries.
Crawling & Indexing has six entries — all official specifications. RFC 9309, the IETF's formal robots.txt standard published in August 2022, establishes that allow rules take precedence over conflicting disallow rules. The Sitemaps XML protocol sets a 50,000 URL / 50 MB per-file hard limit. Google's Googlebot documentation states that Googlebot fetches only the first 2 MB of any file.
Core Web Vitals covers Google's official threshold definitions alongside independent benchmark data. The Web Almanac 2024 Performance chapter measured that 43% of mobile sites pass all three CWV metrics — down from 48% after Google replaced FID with INP in March 2024. The five entries here explain why a Lighthouse lab score and a GSC field data reading can point in opposite directions.
Structured Data includes the schema.org specification (maintained by a W3C Community Group founded by Google, Microsoft, Yahoo, and Yandex), the JSON-LD 1.1 W3C Recommendation, and a 2026 paper finding that standard JSON-LD alone didn't produce the AI retrieval gains practitioners expected — enhanced entity pages with dereferenceable linked data produced a 29.6% RAG accuracy improvement instead.
Ranking Signals covers Google's own confirmed disclosures. The Guide to Ranking Systems lists 17 named active systems. The How Search Works documentation establishes the three-stage model — crawl, index, serve — that explains why a page can pass crawling and still never rank. The 2024 Content Warehouse API documentation leak, which Google confirmed as authentic, documented NavBoost click signals and a siteAuthority attribute as explicit internal attributes.
The Inclusion Bar
Brass-SEO applies a three-part bar to every source before it gets added. The source must be a primary document — a peer-reviewed paper, official specification, or non-vendor independent benchmark. The URL must resolve and the work must not have been retracted. And it must carry a concrete applied implication for practitioners, not just an interesting finding.
That last requirement cuts a lot. A paper with no clear consequence for how you run a site doesn't make the index, even if it's well-cited.
Hard excludes: vendor white papers, tool-vendor benchmarks, paywalled-only papers, abandoned tools. Brass-SEO has no financial relationship with any listed source. All URLs were fetch-verified in June 2026.
Why This Matters for GEO
Brass-SEO's AI Citability analysis already audits pages for generative engine optimization signals. The research index is the evidence base behind those recommendations.
The 40% visibility gain from adding numerical statistics — that comes from Aggarwal et al. The schema.org markup and metadata freshness recommendations come from the GEO-16 framework, which measured citations across Brave Search, Google AI Overviews, and Perplexity. When Brass-SEO flags a specific signal in a page audit, there's a primary source behind it.
There's a second reason this matters. AI search engines, as the citation research shows, favor sources they can verify. A page that attributes its claims to named authors with checkable URLs signals a different kind of credibility than a page that says "studies show." The research index is built to that standard.
Frequently Asked Questions
What is the Brass-SEO Research Index?
The Brass-SEO Research Index is a curated collection of primary literature on SEO and generative-engine optimization at brass-seo.com/research. It covers 26 entries across five topics: AI citation research, crawling and indexing, Core Web Vitals, structured data, and ranking signals. Sources are limited to peer-reviewed papers, IETF and W3C standards, and independent benchmarks from organizations with no stake in SEO software sales.
Is the research index free to access?
Yes. The research index is public and requires no Brass-SEO account or subscription.
How is this different from a typical SEO blog post about research?
Each entry links to the primary source and explains the specific factual point Brass-SEO draws from it. The inclusion bar excludes vendor white papers — which make up the majority of sources cited in typical SEO content — and requires that every source carry a concrete practitioner implication. The URLs are fetch-verified and retraction-checked at each review cycle.
What does "primary literature" mean in this context?
Primary literature means the original document: the peer-reviewed paper (not a blog post summarizing it), the official IETF RFC (not an article explaining it), the actual HTTP Archive Web Almanac chapter (not a tweet quoting one number from it). Every entry in the research index links to that original source, not an intermediary.
How often is the index updated?
Sources are fetch-verified every six months and the retraction status of academic papers is checked at each cycle. New entries are added when a source passes the three-part inclusion bar. The last verified date appears on the pillar page and each topic page.
How does the research index connect to Brass-SEO recommendations?
Brass-SEO's AI Citability analysis and page audit recommendations are grounded in the primary literature indexed here. The GEO paper by Aggarwal et al. (KDD 2024) is the basis for recommendations about numerical evidence. The GEO-16 framework is the basis for schema.org and metadata freshness signals. The research index makes that connection explicit.