Graphite

Ethan Smith has been working in SEO since 2007—18 years of expertise in a field he now believes is undergoing its second-most-significant transformation ever. The first was when Google introduced Panda and similar algorithms to kill spam; this second shift is the rise of Answer Engine Optimization (AEO), where LLMs like ChatGPT, Claude, and Gemini are becoming the new search interface.

The insight is elegant: ChatGPT is already driving more traffic to some newsletters than Twitter. This isn't a distant future—it's happening now. But most companies don't understand how to win at this new game.

Graphite emerged as a consultancy and (increasingly) a platform to help companies optimize for answer engines. The core realization was that AEO isn't just "better SEO"—it's a fundamentally different discipline. In Google Search, ranking #1 for a keyword wins you the click. But in LLM answers, the LLM summarizes many citations, so you need to be mentioned as *many times as possible* across different sources.

This unlocks a massive opportunity for early-stage companies. With traditional SEO, you can't rank without domain authority, which takes months or years to build. With AEO, a brand-new startup can get cited tomorrow—through a Reddit comment, a YouTube video, or a mention on a blog—and start showing up in answers immediately. The entire game theory changes.

Webflow became an exemplar case. Working with them, Graphite discovered that the three most effective citation channels were: (1) traditional SEO landing pages optimized for specific questions, (2) YouTube/Vimeo videos, and (3) Reddit comments. The secret to Reddit wasn't automation or spam—it was authenticity. Real Webflow employees making real accounts, identifying threads relevant to their product, saying who they were, and providing genuinely useful information. Even five high-quality Reddit comments could move the needle.

The results were stunning: **a 6x conversion rate difference** between LLM traffic and Google search traffic. People coming from ChatGPT had high intent (they'd already had a conversation, narrowed their thinking) compared to cold Google searchers.

Ethan is emphatic about the gap between common "best practices" and actual results. Most SEO advice online is wrong because people repeat things without doing analysis. The solution: run controlled experiments. Take 200 questions. Leave 100 alone (control group). Intervene on 100 others (test group). Track results for two weeks before and after. Compare. Reproduce the winning strategy multiple times to confirm it works.

What didn't work: automated spam. Reddit's moderation and LLMs' own quality filters quickly surface and suppress inauthentic content. What worked: specificity, authenticity, and long-tail questions that didn't exist in Google Search (e.g., "How do I build a payment processing API for X use case?").

Ethan also emphasized that 19 out of 20 landing pages drive nearly zero traffic. Most companies waste effort on pages that don't matter. Instead, focus on the few high-impact pages and write them really well, answering all follow-up questions and subtopics comprehensively.

Graphite is positioned as the go-to consultancy for AEO strategy. Ethan is building answer-tracking tools (60+ exist already) and a Webflow app to help companies block training on their content while allowing indexing. He's also researching citation overlap across different LLMs—finding that ChatGPT and Google Search only overlap ~35% of citations, while Perplexity and Google overlap ~70%.

The core framework Ethan shares is deceptively simple: (1) Figure out what questions you want to rank for (use competitor paid search + ChatGPT to convert keywords to questions). (2) Set up a tracker to measure your current share of voice. (3) Analyze who's showing up and design landing pages to answer those questions comprehensively. (4) Run targeted off-site campaigns (affiliates, YouTube, Reddit, Quora). (5) Run controlled experiments to confirm what works. (6) Scale the winners. (7) Hire a hybrid team—SEO specialists plus a marketing/community generalist.

For early-stage founders, Ethan's advice is counterintuitive: skip traditional SEO entirely. Focus only on citation optimization and long-tail questions. You can win quickly, with minimal domain authority, by being authentic and specific.