STAT 101 Lecture 22: AB Testing
Document Summary
Stat 101 - introduction to business statistics - lecture 22: a/b testing. Making decisions based on what you know, not just a hunch. It has been around for a long time, but popularized in web development. Two versions of a web page with different loading times. Outcome metric: customer satisfaction (0 to 10, 10 is better). The faster loading version takes much more server resources so we want to make sure it really is appreciated. Each test participant sees both versions of the page (so it is a paired test) Do a one-sample t test on the column of differences to see if the mean is significantly different from 0. Ex) two versions of the same web page - version a waits 10 seconds before popping up the signup box. Version b gives an immediate sign up box. 1001 independently selected individuals saw each page versions. The 95% ci for the difference in conversion rates is (2. 7% , 6. 5%)