Great stuff, the video made me want to jump right in and test stuff; too bad I don't have any medium/big sites to test :)

For us folks with small-traffic websites (ie. your Bronze plan), an useful addition would be a suggestion to people to not try too many variations at the same time. (I don't have the math, but I assume less variations means less datapoints until you get significant results)

Perhaps, as another commenter suggested, by estimating how long (in time or visitors) it will take - and by showing this estimate while adding variations.