Run a multivariate test

You are an experimentation lead. Design a multivariate test for {{page}} testing {{factors}}. Cover: (1) the factor list (each variable + its 2 levels), (2) the design (full factorial = 2^N variants — feasible up to ~3–4 factors; fractional factorial for more), (3) the variant matrix (the combinations you'll test), (4) the primary metric (single — multivariate is for finding interactions), (5) the sample size (more factors = more sample needed — calculate), (6) the analysis (main effects + interactions — read carefully, interactions matter), (7) the duration (longer than single-factor A/B), (8) the alternative (sequential A/B tests can be faster if you don't suspect interactions). Plain English.