The GEO Lab — Engineering Search for the Generative Era
Last updated: 3 March 2026 · Revised with latest GEO research and updated comparison data.
The GEO Lab studies how content is retrieved, extracted and synthesised in AI-driven search systems — and how optimisation must evolve beyond ranking.
From document-level scoring to section-level retrieval. From position tracking to inclusion modelling.
Why Has Search Changed?
The Shift from Ranking to Retrieval in GEO
GEO addresses a fundamental shift in search architecture: modern AI systems retrieve and synthesise individual content blocks rather than scoring whole pages. Traditional document-level ranking is giving way to section-level retrieval.
For two decades, search engines evaluated entire documents. Pages were scored, rankings were assigned, and positions determined visibility. In my 20+ years of SEO practice, I have observed that model becoming increasingly insufficient as AI systems transform how users find information.
Modern search systems retrieve sections, not just pages. They compress multiple sources into summaries and synthesise answers instead of listing links. Visibility is shifting from ranking to inclusion. If a section is not retrieved, it cannot be cited. If it cannot be parsed cleanly, it cannot be extracted. If it cannot survive compression, it disappears.
The scale of this shift is significant. According to SE Ranking’s 2025 research, AI platforms generated over 1 billion referral visits in June 2025 — a 357% increase year-over-year. Research from Ahrefs shows that Google AI Overviews reduce organic click-through rates by up to 58% for top-ranking pages.
This is a structural change — not a feature update. Generative Engine Optimisation (GEO) is the study and engineering of that transition.
What Is Generative Engine Optimisation?
GEO: Beyond Traditional SEO Rankings
GEO focuses on whether content participates in AI-generated answers, not where it ranks. This fundamental shift requires new optimisation strategies and measurement approaches.
Generative Engine Optimisation (GEO) is the practice of designing content to maximise its probability of being retrieved, extracted, and synthesised within AI-driven search systems. Traditional SEO optimises for position. GEO optimises for participation in the answer. The GEO Lab documents this transition through controlled experiments, framework development, and public tooling.
How Does GEO Compare to Traditional SEO?
GEO extends traditional SEO rather than replacing it. Understanding the differences helps practitioners prioritise interventions:
| Aspect | Traditional SEO | GEO |
|---|---|---|
| Optimisation unit | Entire pages | Individual sections |
| Primary goal | Rankings and click-through rate | Inclusion in AI-generated answers |
| Key signals | Backlinks, domain authority | Extractability, entity clarity |
| Success metric | Position and organic traffic | Retrieval and citation rate |
| Orientation | Document-centric | Retrieval-centric |
In my testing, I found that pages ranking #1 in Google often fail to appear in AI-generated answers when their content structure prevents clean extraction — this is the gap GEO addresses.
What Is The GEO Stack?
Five Layers of GEO Visibility
The GEO Stack organises optimisation variables by layer, allowing practitioners to diagnose issues and prioritise fixes systematically rather than applying ad-hoc changes.
The GEO Stack is a five-layer framework for engineering generative visibility, developed by Artur Ferreira at The GEO Lab. The five layers are:
- Retrieval Probability — the likelihood that a section is selected for inclusion
- Extractability — how cleanly content can be parsed and compressed
- Entity Reinforcement — consistent naming and entity density throughout content
- Structural Authority — trust signals, citations, and expertise markers
- System Memory — how AI systems remember and reference your content over time
Each layer addresses a distinct aspect of how generative search systems select, parse, and cite content. In my experience auditing hundreds of pages, I have found that optimisation must address all five layers sequentially — a deficiency in a lower layer limits the performance of every layer above it.
What Will You Find at The GEO Lab?
GEO Research and Public Tooling
The GEO Lab publishes controlled experiments, case studies, and diagnostic tools for practitioners implementing generative search optimisation strategies.
The GEO Lab publishes research and tools for generative search optimisation:
- Controlled experiments testing content restructuring for AI retrieval
- Case studies on extractability and summarisation performance
- Schema and structured data implementation analysis
- Internal linking strategies for entity density
- Measurement frameworks for AI visibility tracking
Every experiment follows a hypothesis, intervention, observation, and business implication structure. See Experiment 001 for a real example of this methodology.
What Is The GEO Lab Console?
GEO Console: Section-Level Diagnostics
The GEO Console measures content against the five GEO Stack layers, providing section-level scoring that reveals exactly where content fails and what interventions will improve visibility.
The GEO Lab Console is a diagnostic tool developed by Artur Ferreira to measure content extractability and retrieval readiness at the section level. The Console analyses pages against The GEO Stack framework and scores content across five dimensions:
- Retrieval Probability scoring
- Extractability analysis
- Entity Reinforcement measurement
- Structural Authority evaluation
- System Memory indicators
The Console is currently in development. For methodology documentation, see the GEO Field Manual.
GEO Revision History
The GEO Lab homepage is revised regularly to reflect new research findings and framework updates. Major revisions are documented below with version notes.
- March 2026: Updated with Round 2 structural improvements, added cross-references, refreshed data points.
- February 2026: Initial publication with GEO Stack framework and baseline research.
Ready to start? Get in touch · Browse the ebook library