How ChatGPT recommends running shoes: no brand wins — you win an intent
Ask ChatGPT five ways a runner shops and you get five different leaders. A category teardown of who wins where, the sources behind it, and the buy box that excludes Amazon.
Get found when buyers ask AI.
Field notes, teardowns and playbooks on generative engine optimization — how ChatGPT, Perplexity and Google AI Overviews decide who to recommend.
Featured
Teardown
Anatomy of a ChatGPT search: how it decides which brands to recommend
We ran real shopping questions through ChatGPT and recorded everything it did — the shortlist, the question it rewrites before searching, the reviews it trusts over your site, and the buy box.
Read the teardownWHAT DECIDES THE SHORTLISTlive
the sources it trusts92
the deciding attribute85
clean product feeds70
recency58
your own website28
Latest
Teardown
Teardown
How ChatGPT recommends business software: a default with winnable cracks
HubSpot leads almost every CRM question — but specialists steal the specific ones, and Salesforce is the only brand ChatGPT actively warns buyers away from.
Teardown
How ChatGPT recommends cold brew: the most wide-open category we’ve measured
Ask the same cold-brew question three times and get three different winners. Niche brands lead, the giants are named but never championed, and one brand’s own blog became a source.
The GEO brief, every other week.
One sharp idea on getting cited by AI engines — plus a fresh teardown. No fluff.