ICP tests — every assumption about who buys gets a funnel and a verdict
“AS9100 shops that advertise CNC roles themselves convert best — they already feel the pain.”
4 signed terms from 6 meetings — this is the desk's bread. Double down.
“Going site-level (the manager with the empty ramp) beats group procurement.”
Early but promising — reply rate 44% vs 18% going via group HQ.
“Bought-list cold email to yard managers would open the marine patch.”
2 replies in 40, zero meetings. Killed after 3 weeks — the patch wants referrals, not lists. The kill is the lesson.
Example tests on the demo book · with real data every funnel number is measured, never typed — and a killed test is a result, not a failure
Work the patch
Who are we hunting? — the signals builder
No list bought, no site scraped — this is your own book plus the signal engine. Phase 2 plugs in the outside world (Companies House, job-board monitors, planning applications) through the same builder.
Why this beats buying a list: a list is 5,000 strangers and a 2% reply rate. This is YOUR BOOK × LIVE SIGNALS × A TESTED ICP — fewer targets, warmer openers, and every campaign teaches the next one who your real clients are.