There is no real A/B testing for SEO, sorry. As you say, there is only one Google. You can try to simulate a 'test' with two equal throwaway sites, but this wi

There is no real A/B testing for SEO, sorry. As you say, there is only one Google. You can try to simulate a 'test' with two equal throwaway sites, but this will never give you the same insight as using it on a big site with real content and real visitors. But it can help you find problems in linking structures, tweak the (really small) benefits of using one or another html tag. The success of SEO can't be isolated. But, increased crawling of the page and disproportionate visitor growth from Google are good signs.