Correlation in SEO is aimed at defining what ranking factors are more or less important during pages ranking.
There are correlation studies that analyze thousands of search results to understand what the ranking factors were and what mattered more.
These studies exist due to the fact that search engine companies keep their algorithms secret and sometimes drop hints about what ranking factors are but not their importance. Therefore, we already got to know about keywords and that keyword stuffing does not work anymore, about website loading time, bounce rates, backlink, etc.
However, correlation studies are not foolproof. The algorithm always changes as machine learning grasps more information and there are too many factors to consider and actually to discover.
At the end of the day, it is also not that productive to work towards algorithm pleasing. The best tactic, as repeatedly stated by search engine companies, is to create great content and serve the best user experience. You should be pleasing the human audience that will send all the traffic you need after you have won their attention.
