This post by Armando Roggio explores the most common problems with A/B testing, as it relates to Ecommerce. The article also shows us why marketers should really wait for the results of their A/B testing tool, before making any drastic changes.
It touches on patience in online marketing. Despite our eagerness to make the new changes to our website as soon as possible, it is still wise to wait one buying cycle before deciding to make a change. This ensures us that you have all the relevant data from the test to make an informed decision. So even if the standard suggestion would say 7 days, if your buying cycle is 40 days, you still have to wait 40 days if you want to measure the impact on revenue, and how to apply the results for changes to your site, whatever standard practices say.
Armando touches three very important topics:
- Random Fluctuation
- Not Examining the Data for Unexpected Differences
- Testing Too Many Variables