How can you increase the conversion of an online store through analysis?

Question

Analytics

Good afternoon, dear users and experts. I ask you not to kick your feet hard, the question is more theoretical.
There is an online store, we collect every step of the user
* Where did he come from (what UTM tags, and from which domain)
* Where did he go
* What did he add to the cart
All information is stored for 5 years, so you can track the entire purchase path. Question!)
How to detect the common intersections of all buyers, and the intersections of those who lost their way (abandoned cart).

Answer 1

ML - will give you an answer (guess) if the customer is going to leave the page or abandon the cart
https://www.quora.com/What-are-the-applications-of...
https://cloud.google.com/ solutions/commerce/
But you will have to find out how to force him to make a purchase using tests on such clients
. If you have few entries in the logs or the entries are dirty, then ML will not work
For a simple "user" there is too much "magic"