Campaign Research: The Five Datasets That Decide GEO and AEO Strategy
Chapter 6
Continued from The Test Test Test Method
Chapter 6 of 14 in LLM Mastery: How AI Recommends Brands for LLM Domination
Every LLM answer you log contains five datasets, and together they are the complete research layer of a GEO and AEO campaign.
What does Chapter 6 argue?
Every logged AI answer contains five datasets: fan-out queries, reasoning data, cited sources, winners and sub-categories.
Together they show where the game is played, how it is scored, where to show up, what to beat and where a brand can win now. When a main category is locked, the chapter's escape route is to create and own a sub-category, seed it everywhere and get mentioned as its leader.
Chapter 6 is where the book defines the five datasets.
The five datasets
Chapter 6 opens with the book's definition of the five datasets, quoted at the start of this page.
The five datasets are the research the book says every logged AI answer contains: fan-out queries, reasoning data, cited sources, winners and sub-categories.
Glossary entries for terms it uses: Fan-out queries, Reasoning data, Cited sources, Sub-categories and Prompt bank.
What is in Chapter 6?
Chapter 6 works through six named parts:
- Dataset one: fan-out queries
- Dataset two: reasoning data
- Dataset three: the cited sources
- Dataset four: the winners
- Dataset five: the sub-categories
- The five together
Who does Chapter 6 name?
Chapter 6 names James Dooley, the person behind the book's digital avatar.
James Dooley, whose digital avatar co-authors this book, coined an entire discipline on exactly this logic, and it works at every scale.
The digital avatar is one of the book's seven authors.
How does Chapter 6 close?
Chapter 6 closes on this line:
The answers are not just the scoreboard. They are the research department, and it works for free.
Chapter 7, Seeding the Training Layer, follows it.