Re:Optimization at the Obama campaign: a/b testing
by karolisd on Wednesday December 12, 2012 @11:06AM
(#4910725)
You're always testing against a control and have something to compare to.
If they really wanted to, they could've kept a small amount of traffic going to the original control from before they started testing.
How did they measure this? Did they isolate the effects of the a/b testing? Donation and sign up conversion rates will increase anyways as the day of the election approaches.