Swapped my A/B tests from gut feeling to sample size math, big shift
Used to run Facebook ads for a local shop in Austin and just picked winners by eye after a week. Got burned twice on 300 click tests that flipped once I spent more. Now I use a free calculator to hit 95% confidence before killing anything. Last campaign needed 1,850 visitors per variant, took 11 days but the result held up. Anybody else scale down their testing budget when they started doing this?