Controversial data[ Edited ]
March 2017 - last edited March 2017
I'm testing Google Optimize on a sample site before implementing it on public site. I run a multiversional test and get the following data: https://drive.google.com/file/d/0B_V8Wqsf_zLhZE4yZU5DamN1ajg/view
So I noticed, that Optimize started to show the variant with the best conversion rate more rare during the experiment. Also you can see in the table, that the last combination received the least amount of clicks. Also it is pretty obvious that the last combination is the best so far, but its probability to be best is only 1% greater of the third variant and estimated conversion rate is lower than second variant.
I realize that there is not too much data, but even now I think that 4th variant's metrics should be higher. I will be very pleased, if someone could explain it to me.