FeatureQuery Fan-Out Analysis
When you ask ChatGPT a question, it silently fans out into multiple sub-queries. We capture every fan-out, classify its purpose, and show which ones mention your brand — and which don't.
Six PurposesEvery sub-query is doing something
Rewords the prompt to maximize retrieval recall
Bolts on a constraint — geography, regulation, price
Pits brands head-to-head in a comparison
Searches for specific brands by name
Targets a trusted publication for sourcing
Adds a recency modifier to filter stale content
The Hidden LayerOne question becomes a dozen
ChatGPT averages 8-15 sub-queries per search-triggering prompt. Perplexity Deep Research can fire 20-50. None of them search for the exact words you typed.
If the model never searches for your brand by name, no amount of content polish will fix it. Fan-out analysis tells you whether you have a content problem, a categorization problem, or an entity recognition problem.
