joined:Mar 27, 2013
I've been having a debate with a few people about the appropriate measurement for finding a winning ad. I think it should be IUC or Impressions Under Conversions where you run an AB split test and compare conversions with how many impressions the 2 ads have and run your statistical analysis. I hear others telling me to measure CTR. The problem with CTR is that you can have 2 ads with 1000 impressions each ad A has 500 clicks and zero conversions and ad B has 250 clicks with 100 conversions. Clearly ad B is the winner but it is the loser when you measure by CTR. How do you optimize towards conversions without sacrificing CTR?
Any thoughts would be appreciated.