A/B or A/B/C?

Seven ways to test the same set of ideas, sized to the same power — and what each one costs.

Same ideas, seven designs

How you arrange the same ideas into experiments decides how much traffic the whole plan costs, what you gain in sample efficiency, and how many decisions you get out of it.

Abs. %
Rel. % per idea
Alpha 5.0% per test
A/B, A/C
Reset
Max
comp.
False
positive
Total
traffic
vs
A/B