Build one 8-slide carousel, generate four different cover openers with Claude, A/B test them for a month, and let the data name the winner — a step-by-step hook-test protocol I now run on every carousel.
Stop A/B testing two subject lines and hoping. Pre-write 15 falsifiable hypotheses (length, emoji, personalization, urgency, numbers, case, brackets, sender, preheader...) in a 15-row spreadsheet, design 2-4 cells per hypothesis, then let the data pick the winners — one row per send, not a multivariate stew. The 6 starter rows, the Klaviyo test setup, and why one row per send is the rule.
How I turn a single 30-minute prospect call into three landing-page variants — each one built around a real objection — and split the traffic to find the winner. The exact prompts, the variant matrix, and the A/B math I use.
