Methodology for testing segments of an email blast What's the best test and methodology to use for a scenario for an Email blast wherein I want to measure whether one or another segment group was more likely to Open?  
 A: Partially answered in comments: 
Wouldn't you simply define the groups send the blasts to all the members in each group and count separately the proportion that do open the email in each group. The group with the highest proportion is most likely to open. 
But the real question is how do you want to generalize this. Maybe you want to look at emails with similar characteristics sent to various groups of people not originally sampled but having the same characteristics as the groups you previously defined. Then you might like to rank the groups according to their proportions. 
Take into account that when two proportions are close numerically that a different random sample might switch the order. The uncommonly considered ranking and selection methodology might be worth considering. Take a look at Selecting and Ordering Populations: A New Statistical Methodology by Jean Gibbons, Ingram Olkin and Milton Sobel. Siam 1999, especially Chapter 4. This is a reprint with corrections of a 1977 publication by Wiley. So the authors considered it new in 1977. There have been several advances since even though it is not commonly used.    – Michael Chernick
