Ross Hudgens sits down with Eric Wu, who leads SEO, GEO, and CRO at Ramp, to unpack one of the most experiment-driven search programs in B2B.
Eric walks through the test that proved a simple markdown file could change ChatGPT’s answers, a $162 “AI-exclusive” demo offer that real prospects asked sales about by name, and why Ramp leaves experiments live for months to track how answer engines shift over time.
Show Notes
0:00 – The markdown experiment: changing ChatGPT’s answer with a $162 offer
3:05 – Marketing to agents without hurting the user experience
4:37 – Personalized AI content at scale
8:24 – Schema tests and why answer engines look like early-2000s search
10:17 – Long-running experiments and getting flagged as manipulation
12:47 – Prioritizing experiments vs. business outcomes
14:42 – Dark search: why AI traffic shows up as direct
17:37 – How Ramp found 30–40% of direct traffic is answer engine traffic
22:00 – Why Google won’t share AI Mode clicks
24:15 – The Search Console vs. GA4 flip after October 2025
25:29 – Why Ramp ranks for the word “yes”
28:29 – Ramp’s Reddit experiments
34:49 – Repurposing blog content on Reddit
38:55 – Vendor comparison pages built for procurement
42:40 – LLM visibility tracking: vanity metrics vs. what correlates with sales
45:15 – Using “How did you hear about us?” to attribute dark traffic
