|
Abstract:
|
Online experiments are conducted on e-commerce web sites to optimize site design to increase conversion rates. There are many interesting and challenging statistical issues involved in the design and analysis of e-commerce online experiments, including sample sizes, sample assignment to treatment groups, and the statistical analysis of the large data sets obtained, data that is often nonnormal and containing outliers. This paper will review the key challenges, present examples of data obtained from online experiments, and review how two major e-commerce sites have approached these challenges. The challenges involved in e-commerce online experiments have had little coverage in the statistical literature. It is hoped a review of the challenges and examples of actual experiment results will increase interest in developing robust approaches to address these challenges.
|
