Blog

Methodology, not opinion.

Substantive pieces on technical SEO, GEO, AEO, and AI search visibility. Each post is a working artifact — the kind of document you could hand to a developer and they could ticket directly. We publish one post per month minimum, more when there is something genuinely new to say.

How AI Search Engines Decide Which Brands to Cite — 12 Technical SEO Levers That Matter

GEO is not a new discipline that replaces SEO. It is technical SEO applied to a new consumer — the retriever that builds an AI's answer. This piece maps the 12 technical SEO levers that most directly determine whether ChatGPT, Claude, Perplexity, and Google AI Overviews cite your brand, and shows where the work compounds.

Recent Posts

The GEO Evidence Problem: 602 of 769 SEO Tactics Have No Controlled Test

A review of 9,249 industry documents counted how much SEO advice has actually been tested. For GEO the answer is barely any — and what we changed because of it.

Test 01: Does FAQPage Schema Affect AI Citation?

A randomized test across seven matched pairs and four AI engines, with the design, seed and decision rule published before any data. Result due December 2026.

The Citation Oligopoly: Why Fewer Than 1,000 Domains Own AI Search

40–55% of ChatGPT and Perplexity citations flow to under 1,000 domains. Reddit alone appears in 18% of ChatGPT answers. The data on concentration — and the three routes that actually work.

Localized GEO: Why Your Brand Gets Cited in English But Not in German

AI citation is locale-specific in a way most US brands have never measured. The cause is almost never translation quality — and the fix is a 30-day schema parity sprint.

How to Measure GEO: Citation Rate, Share of Answer, and What Actually Matters

Every retainer conversation arrives at the same question: how do we know this is working? The three-layer measurement framework, with 2026 benchmarks across all four AI surfaces.