DO NOT DO A/B TESTS IF YOUR SAMPLE IS UNDER 100K. This is a complicated hack, so even if you have 100k plus sample think twice, to make up your mind look at this analysis of the takeaways of A/B testing, and this analysis of what works with the on-page design. Additionally, read this guide and if you decide to go for it, Optimizely and VWO are the tools you’ll need.