Website: analysis of visitor preferences

My idea is not new in the global sense of the word. It is also not universal or complete. But, in my opinion, in some cases and on some types of sites, it can be very useful.

There are some types of sites where content is located on one page, more or less different in topics (blogs, ads). You can determine the preferences of a site visitor using a simple client script. To do this, the page script can analyze:

• moments of the “stop” of the user to read certain parts of the page (scrolling - stopping - scrolling)
• in the moments of “stop” - analysis of the mouse positions (the cursor is often located inside or near the object of interest). This factor should be taken into account with a set of reservations, as well as a lesser degree of trust.

The introduction of sequential refinement and heuristics, it seems to me, will allow you to quickly identify user preferences and dynamically provide target page elements to it by loading them via AJAX (sent the URL, context and scroller position - received a set of elements). These can be either more suitable notes / postings / announcements, links to contextual internal pages, or advertising, or simply useful links.