Zach Cohen and Seema Amble — a16z - How Generative Engine Optimization (GEO) Rewrites the Rules of Search - May 2025

Free

How Generative Engine Optimization (GEO) Rewrites the Rules of Search

FreeFree tier

Type

Open Source

Company

Andreessen Horowitz

Visit Website GitHub (38 stars)

About Zach Cohen and Seema Amble — a16z - How Generative Engine Optimization (GEO) Rewrites the Rules of Search - May 2025

This a16z blog post by Zach Cohen and Seema Amble analyzes the paradigm shift from traditional SEO to Generative Engine Optimization (GEO) as AI-native search engines (e.g., Perplexity, Claude, ChatGPT) replace browser-based search. It explains that visibility now depends on how often LLMs cite your content (reference rates) rather than click-through rates or page rankings. The post covers the fragmentation of search across platforms (Instagram, Amazon, Siri), longer query lengths (23 words average vs. 4), deeper sessions, and the need to optimize content for model extractability (e.g., bullet points, summary phrases). It discusses the business model differences between ad-driven search and subscription LLMs, namedrops emerging GEO platforms like Profound, Goodie, and Day, and provides actionable insights for marketers adapting to this new landscape.

Key Features

Explains the shift from SEO (link ranking) to GEO (LLM reference rates) for visibility in AI-native search

Discusses fragmentation of search across LLMs, social platforms, and voice assistants

Provides data on query length (23 words avg) and session depth (6 minutes) in AI-native search

Offers content optimization tips for LLM extractability: bullet points, summary phrases, clear structure

Analyzes business model differences between ad-based search and subscription LLMs

Names early GEO platforms: Profound, Goodie, and Day

Highlights ChatGPT's referral traffic to tens of thousands of domains as a signal of LLM-driven discovery

Pros & Cons

Pros

Provides a clear, data-driven explanation of the transition from SEO to GEO
Offers practical content formatting recommendations for LLM citation
Includes specific metrics (query length, session depth) that marketers can benchmark
Names real-world platforms and referral traffic data to ground the analysis
Written by a16z partners with credibility in tech venture investing

Cons

The analysis is forward-looking and may not apply to all industries equally
Limited in concrete implementation steps – more conceptual than tactical
Does not provide a step-by-step GEO audit or tool recommendations
Focuses on early observations; the GEO ecosystem is still nascent

Best For

Content marketers adapting SEO strategies to AI-native search enginesBrands measuring visibility through citation/reference rates instead of page rankingsProduct teams designing content that LLMs can easily parse and citeSEO professionals learning the new rules of generative engine optimizationInvestors understanding the structural shift in the $80B+ SEO market

FAQ

What is Generative Engine Optimization (GEO)?

GEO is a new paradigm for visibility in AI-native search engines (like ChatGPT, Perplexity, Claude). Instead of optimizing for page rank (SEO), brands optimize for reference rates – how often their content is cited by LLMs in generated answers.

How is GEO different from traditional SEO?

Traditional SEO focused on ranking high on search results pages via keywords, backlinks, and engagement. GEO focuses on content being well-organized, easy for LLMs to parse (e.g., bullet points, summary phrases), and dense with meaning. Visibility means appearing in the AI's answer itself, not a links list.

Why is the business model for AI-native search different?

Traditional search monetized through ads; users paid with data/attention. Most LLMs are subscription-based, reducing incentives to surface third-party content unless it adds user value. This changes how content gets referenced.

What are examples of AI-native search platforms mentioned?

The article mentions Perplexity, Claude, ChatGPT, and notes fragmentation across Instagram, Amazon, and Siri. It also names emerging GEO platforms Profound, Goodie, and Day.

How should content be optimized for GEO?

Use clear formatting like bullet points, include summary phrases (e.g., 'in summary'), and ensure content is dense with meaning rather than just keywords. LLMs favor well-structured, easily extractable text.

What is a reference rate?

A reference rate measures how often a brand or content is cited as a source in model-generated answers. It replaces click-through rate as a key metric in the GEO era.