About Artur Ferreira — The GEO Lab

Why I Built The GEO Lab

Artur Ferreira founded The GEO Lab after 20 years (since 2004) in SEO and organic growth strategy. The shift from traditional search optimisation to Generative Engine Optimisation did not happen overnight, but there was a moment the change became undeniable.

I was reviewing performance data for a client who’d done everything right: two years of consistent publishing, a strong backlink profile, technically clean architecture. Rankings were solid. Then AI Overviews started appearing above the fold for their core queries, and organic click-through dropped sharply — without a single ranking change. The page was still in position one. The traffic had simply gone somewhere else: into a synthesised answer at the top of the page, assembled from multiple sources, none of which were the client’s.

That’s when I understood that the game had changed at a structural level, not a tactical one.

For most of my career, search was a ranking problem. We optimised for keyword targeting, link authority, technical crawl efficiency, and SERP position. But search is no longer purely ranked. It is retrieved, interpreted, compressed, and synthesised. Large language models, AI Overviews, and generative search systems have introduced a new layer: visibility is now partly determined by whether your content is retrieved, understood, and cited inside machine-generated answers.

Most industry discussion about this shift has been reactive — tool comparisons, prompt tactics, surface-level speculation. Very little focuses on the structural mechanics: how retrieval systems decide what to extract, how entity clarity affects summarisation, how content architecture reinforces machine confidence. That gap is why The GEO Lab exists.

This site documents my transition from traditional SEO strategy to Generative Engine Optimisation (GEO) — publicly, rigorously, and experimentally. Not as a hype cycle. As a systems problem.

What You Will Find Here

The GEO Lab is not a marketing blog. It is a documented research effort. Everything published here starts from a hypothesis, goes through a defined intervention, and produces documented observations with commercial implications.

In practice that means:

• Controlled GEO experiments — with documented results, including a 24-percentage-point improvement in citation rates from structural changes alone. See Experiment 001.
• Content restructuring case studies — real before-and-after examples of extractability improvements, including the PageSpeed case study showing how technical performance affects retrieval probability.
• Structured data implementation tests — schema markup effectiveness in AI retrieval contexts.
• AI visibility measurement frameworks — methods to track retrieval and citation across platforms where standard analytics fall short.
• Free practitioner resources — browse the complete GEO Lab Library for guides, frameworks, and reference tools.
• The GEO Lab Console — a diagnostic tool measuring content extractability and retrieval readiness at the section level, currently in development.

All experiments and audit results are published openly in The GEO Log.

My Approach

Twenty years in SEO teaches you to be sceptical of trend cycles. Most of what gets labelled “the future of search” turns out to be a feature update with a good PR team. Generative search is different — not because it’s newer or shinier, but because it changes the unit of optimisation from the document to the section, and from position to retrieval. That’s a structural shift, not a surface one.

Four principles shape everything published here:

• Systems over tactics AI search is architectural. The retrieval pipeline — query processing, section selection, extraction, compression, citation — is consistent in what it rewards. Surface tricks don’t scale against a system; understanding the system does.
• Evidence over hype If a GEO claim can’t be tested, it isn’t strategy — it’s opinion. Every framework published at The GEO Lab is grounded in documented experiments with measurable outcomes.
• Commercial impact over vanity metrics Retrieval rate and citation frequency are interesting. Revenue and strategic positioning are what matter. GEO research here is always anchored to business outcomes, not AI novelty.
• Evolution over nostalgia Traditional SEO fundamentals — domain authority, technical health, link equity — still matter and still feed the retrieval pipeline. But optimisation that stops at the page level is now incomplete. Both layers are required.

The framework behind everything published here is the GEO Stack — a five-layer model I developed for engineering generative visibility systematically. It covers retrieval probability, extractability, entity reinforcement, structural authority, and system memory. Each layer targets a specific stage of the retrieval pipeline. Read the full framework on the GEO Stack page.

Current Projects

The GEO Field Manual — a 90-page practitioner guide to Generative Engine Optimisation, published February 2026. It covers the full GEO Stack implementation, section-level audit methodology, entity reinforcement strategies, and AI visibility measurement. Written for practitioners who need a working system, not a theoretical overview.

GEO Lab Console — a diagnostic tool measuring content extractability and retrieval readiness at the section level. Currently in development. It surfaces the exact GEO Stack layer where a content section fails — so practitioners can fix the right thing, not just the most obvious thing.

GEO Brand Citation Index — a public tracker measuring how frequently brands are cited in AI-generated search responses across platforms, updated regularly.

Get in Touch

If you have questions about GEO, want to discuss the research, or are interested in collaboration, visit the Contact page to reach me directly. I respond within 48 hours.

You can also follow the ongoing research on X/Twitter (@TheGEO_Lab) or connect on LinkedIn.

Frequently Asked Questions