Apart from the content on the page, there are a few things on the page that can add SEO value. Getting the right title, meta tags and enough links pointing to a page.

But how do we measure the significance of each of the factors?

In most cases, we'll have more than 1 experiment running on a page. Some experiments are meant for the users and bots alike whereas some are solely for the bots. Isn't it becoming harder and harder to attribute a lift/drop in SEO indicators to one particular change.

The author describes reverse engineering Google's Algorithm in this podcast -

https://podcasts.apple.com/us/podcast/seo-testing-how-to-run-measurable-seo-test-kyle-roof/id1038415449?i=1000440616308

How feasible is this ?

Any thoughts?

Thanks!