Please mind your words. Every kind of multivariate testing requires you to theoretically have data for every combination. In partial factorial or taguchi method

Please mind your words. Every kind of multivariate testing requires you to theoretically have data for every combination. In partial factorial or taguchi methods you simply estimate it from other live combinations. This doesn't mean you don't need data for all combinations. Moreover, this article's intended audience is beginners. If you know about full and partial factorial audience, you probably know main differences between A/B testing and multivariate testing.