Design of Experiments

DOE is the statistical craft of structuring a test so a handful of runs reveal which factors matter and how they interact — instead of the slow, naive ritual of changing one thing at a time. I learned it formally as a Six Sigma Green Belt, running it on manufacturing processes.

It’s exactly what good product experimentation should be: decide what you’re trying to detect before you start, vary deliberately, and design the test so the answer is unambiguous when it arrives.