Religiously testing will only lead to a local maximum and ignores the possibility that the design may have not been great to begin with.
Google's homepage is not even remotely close to the standard a highly competent minimalist designer would offer. Little attention to detail, remedial typographical choices, a 90's looking poorly executed Logo, random UI elements with no clear purpose. It's very much like an engineers interpretation of what a minimalist designer would do.. which is exactly what it is.
All of their products are on the utilitarian end of minimalist design and could stand to be improved significantly from an aesthetic perspective without actually changing the interface noticeably.
You know what? I liked this redesign a lot. I like how white it is. I like the smaller logo. I like the effort the author put into making each logo look really crisp, the Wikipedia one in particular. I like that there are only four words on the entire page. I like the blue bar around the search box. So seeing this page made my day a little more enjoyable, and I'm glad it was posted here. No, I don't think it would be a good idea to implement this, but maybe that wasn't the point.
Compare what a competant designer comes up with to Google's home page and start asking "why did Google do it this way", and you might learn something new.
I've actually always been against that mindset myself, and I've argued against it before here. I don't necessarily think it's a good thing to sacrifice everything just to boost numbers. If we had an Internet where there were half a dozen really good search engines where each one handled search a little bit differently, I'd much prefer that to the current situation. But I'm in the minority there.