[joosters]

I wonder why companies who host their own analytics choose to use a separate request to track users?

Surely it is better just to get the data from your own content web server logs ? Wouldn't you get the same information while saving on the extra HTTP request? It would make the site loading times slightly faster as well.

[strommen]

Server logs will track all requests from crawlers, accelerators, and cancelled navigations. An AJAX callback will only happen for real human visitors that actually view the page.