ICP Profiler

Generic ICPs fail in the same way: they describe demographics instead of triggers. WHO they are matters less than WHAT happened to make them buy this quarter. A perfect demographic fit with no acute pain doesn't close — and the demographic-only ICP can't tell you the difference. This skill builds a profile that names both, in five layers from firmographics through to anti-ICP.

The process starts from your actual best customers if you have them, or from a structured hypothesis if you don't. The questions push past the obvious: "who are your three best clients ever?" usually surfaces a pattern in under ten minutes that the user hadn't consciously noticed. The pain-trigger layer is the hardest and the most valuable — chronic pains are background noise, but the acute trigger (a funding round, a board mandate, a competitor move, a key hire leaving) is what makes someone reach out this week instead of next year.

The output includes an "in their own words" section pulled from real conversations, reviews, and forum posts — not your marketing copy. This becomes the language for outreach, ads, and landing-page headlines. It also includes an anti-ICP: the red flags, the wrong company sizes, the role mismatches that signal a bad-fit prospect. The anti-ICP usually saves more money than the ICP makes, because most wasted pipeline isn't about chasing too few of the right people — it's about chasing the wrong people for too long.

Confidence is labelled honestly. An ICP built from ten-plus actual customers gets "high confidence — drive targeting now." Three to nine customers gets "emerging patterns — validate in 90 days." Pure hypothesis gets "treat as a bet, set a validation deadline." If the user says "we sell to everyone," the skill pushes back constructively rather than agreeing — usually with "if you could only keep 10 of your current customers, which 10?" That question alone breaks most "everyone" claims inside five minutes.