Built the customer research system that informed every messaging decision for the You Ai launch

The methodology

The product was new. There was no existing playbook for who would buy it or what they would respond to. The sales team was running discovery and sales calls with target customers as part of the early go-to-market, and those calls were the richest source of voice-of-customer data we had. I sat in on the recordings, took transcripts, and turned them into a research system the rest of the team could draw from.

• Sales-call review. Twenty-plus discovery and sales calls. Most were led by one of the co-founders or someone from the sales team. I watched the recordings, took transcripts, and tagged language by category.
• Three-layer signal extraction. First layer: facial cues. The expressions that lit up the prospect's face versus the moments their tone dropped. Second layer: vocal patterns. Where they paused, where they sped up, where they got specific versus vague. Third layer: the actual words. Pains, desired outcomes, objections, and the limiting beliefs underneath each.
• Sample expansion. Took the language I extracted from the calls and cross-checked it against Reddit threads, niche forums, and creator-economy social channels. The point was to verify that what twenty prospects said was representative of the broader market, not the quirks of the people who happened to book a sales call.
• First-principles analysis. For customer research I worked from first principles. For competitive analysis, the model was different (analogical reasoning). The two-track approach kept the persona work grounded in primary data while the angle work could borrow from what was working in adjacent markets.

What the system produced

• Three named customer personas. Each had its primary pain, its motivation, its buying stage, its top three objections, and the content type that resonated. The objections were quoted directly from sales-call transcripts so the team could write against the customer's actual language, not a brand-doc paraphrase.
• The You Ai Messaging Rules document. An ever-expanding page that codified the do's and don'ts for every copywriter and marketer working on You Ai. Every time we found a new angle that worked or a phrase that pushed the wrong button, the document got updated.
• Audience expansion via persona refinement. Built an "audience expansion" workflow inside the team's Claude account that ingested the persona research and generated three or four solid sub-avatars per primary persona. The output kept ad creative and email angles fresh, which materially reduced creative fatigue across paid media.
• Cross-feed into BFCM and platform work. The personas and angles I built for You Ai fed downstream into the Black Friday campaign segmentation and the platform's own messaging. Same methodology, different product.

One specific moment the research changed a decision

The most consequential research finding became the $100K-to-$10M repositioning. The transcripts surfaced that buyers under $300K could not picture the path the You Ai promise was selling, and buyers over $1M wanted scale specifics, not foundation messaging. That observation drove the team to subdivide the ideal customer profile into revenue brackets.

That decision happened because the research was specific enough to be cited, not because someone had a hunch.

Why this counts as product marketing work

The work was upstream of copy. It defined who You Ai was for, what they actually believed about their problem, and what language would and would not move them. The messaging the team produced after this point was a direct application of the research, not a search for it.