One Card Field Doubled His Signup-to-Paid Rate to 9.2%

98% of his signups started the free trial. 5% of them ever paid. Every one of the other 95% burned inference money on the way out. In August 2025 he put a credit card field in front of the trial, changed nothing else, and watched three-quarters of his signups walk. The ones who stayed paid at 42%. This is a case about deleting customers on purpose.