Analysis of an email experiment testing two subject messages and 4 content messages

by Marie-Eve   Last Updated January 20, 2018 18:19 PM

I'd love to hear how I would analyze aggregated data from a email experiment.

My design is very straightforward. I wanted to test two versions of the subject message on the email opening rate, and four versions of content messages. But, I've not used and interaction between content and subject message.

For this, I've used two types of messages randomly assigning them (about 50% each) to 1600 email accounts. Then, with another batch of 1800 accounts I tested the content message by assigning four versions of the message at a rate of 25% each experimental message.

What would be the best strategy for acquiring some extra insight from data other than the simple percentage figures? Is there any test that could provide extra explanation?

The actual outcome data

  data <- data.frame(
    subject=c("A","B","A","A","A","A"), 
    content=c("A","A","B","C","D","E"), 
    opens=c(475,488,321,312,299,300),
    clicks=c(158,168,189,174,159,165), 
    total=c(798,802,449,452,448,451)
)


Related Questions


3 + 3 Phase 1 clinical trial design

Updated February 26, 2018 10:19 AM


Determining Probability of Smallest Departure Delay

Updated December 19, 2017 23:19 PM


How to test the probability analysis result?

Updated May 12, 2018 12:19 PM