> After we get enough traffic to run through our test And therein lies the trick. Google and Facebook can do multivariate testing easily because they get 100k

> After we get enough traffic to run through our test And therein lies the trick. Google and Facebook can do multivariate testing easily because they get 100k pageviews without blinking an eye. You probably cannot. In fact, the more variables you start to simultaneously manipulate, the more likely it is to guarantee that you will not attain statistical significance for any decision (i.e. your study will be underpowered).