How I actually test CTR vs RPM (not vibes)

How I actually test CTR vs RPM (not vibes)
Everyone chases CTR. Here's the playbook that finds real money:
— Pick ONE high-traffic page, lock everything else
— Variant A: 3 units. Variant B: 5 units. 50/50 split if you can, else 4 days each
— Record page RPM AND CTR both runs
— More units usually = lower CTR but HIGHER page RPM (more inventory)
— Stop when added unit drops page RPM, that's your ceiling
My ceiling was 5 units: CTR dropped 2.1→1.7 but page RPM went +18%.
Steal it.