Hi Rock
We've already built something like Google Analytics scripts, we call wrappers, for popular behaviors on usual learning platforms. The wrapper is designed according to
AcrossX vocab/profiles (
Chinese and English doc. links here) and has dataviz available for visualizing those xAPI statements. With the complication xAPI statements, if without pre-designed dataviz, the analytics will be pretty primary.
If you can identify all your web behaviors, we can check if any more behaviors needed to be added to our profiles.
From your last name, I am hoping you can read Chinese. There is a case in our
Chinese CoP that leverages xAPI + machine learning for online marketing, obvious improvement has been obtained. The online behaviors include events across the web and mobile browsers. The write-up is here:
http://xapi-cop.net/zh/2017/10/28/xapi-online-marketing/e.g. bounce rate and exit rate down ~40%, click rate on pushed content up from 8% to 34%, call-to-deal rate up from 33% to 53%.