Fame gets you used. Methodology gets you remembered.
The first AI citation study run as a controlled experiment — six weeks in, and the findings upend the standard advice.
Your buyers have moved their research into AI assistants. On one topic we measured, AI gets asked 74× more often than Google gets searched — in full questions, not keywords. So when AI answers your buyers’ questions: whose content does it use? Whose name does it say? And what actually decides who gets remembered?
Mid-study findings · Live on four AI platforms · Seats are limited
Not a snapshot survey. A controlled study.
Ten live websites — some treated with a specific methodology, some deliberately left alone, including two of the most famous names in marketing — tracked against roughly 1,100 real buyer questions on four AI platforms, on a fixed schedule, with every raw AI answer stored. Six weeks into an 84-day run, with predictions filed in writing before the results came in.
Six weeks in, the standard advice is failing the test.
Fame gets your content used — but almost never gets your name said.
The famous sites appear as anonymous links under AI answers again and again; the AI has never once told the buyer who they are. One of their field’s signature ideas is now repeated by AI nearly word for word — crediting no one.
SEO strength does not carry over.
The two strongest SEO operations in the study — sites that dominate traditional search rankings — have never been used as a source by ChatGPT, the assistant buyers use most. If you’ve been told SEO = AI visibility, this study is the test of that claim.
The “volatility” everyone fears is predictable.
The industry says 70% of AI visibility vanishes in six months and calls it volatility. We watched which 70%. It isn’t random — it’s predictable from three things a content team controls.
93% of the winnable ground isn’t what budgets optimize for.
93% of what the winners win is not what most marketing budgets are optimizing for. We’ll show you where the winnable ground actually is — and the trap query that looks like a buyer question but isn’t.
You’ll leave with more than takeaways.
These are mid-study findings — the patterns that have held checkpoint after checkpoint.
The full study report comes when the study concludes, and the next study will accept new participant sites. If you want your content measured the way these ten sites are being measured, this webinar is where that conversation starts.
Register now — seats are limited.
Join us for the mid-study findings: where the winnable ground is, what makes an idea ownable, and what to tell your team to do differently next quarter.
