Multivariate testing (MVT) tests multiple page elements simultaneously to find the optimal combination. Instead of testing headline A vs headline B (that's A/B testing), you test headline A + image 1 + green CTA vs headline A + image 2 + blue CTA vs headline B + image 1 + blue CTA... and every other combination. A test with 3 headlines, 3 images, and 3 CTA colors has 27 variations.
The advantage: you find interaction effects. Maybe headline A works best with image 2, but headline B works best with image 1. A/B testing would never reveal this because it tests elements in isolation. The disadvantage: 27 variations need roughly 27x the traffic to reach statistical significance. That's why MVT is reserved for high-traffic pages.
When to use MVT (and when not to)
Use multivariate testing when you have at least 10,000+ conversions per month, you've already exhausted the obvious A/B tests, and you suspect element interactions matter. A good candidate: testing headline copy, hero image, and CTA copy simultaneously on a high-traffic landing page.
Don't use MVT when you have moderate traffic (under 50,000 monthly visitors), when you're testing a radical redesign (A/B test that), or when you haven't done basic A/B testing yet. Most companies would get better results from running 10 sequential A/B tests than one multivariate test. MVT is a precision tool for mature optimization programs, not a starting point.