Graphing trends can make you better at calling A/B tests

by Brian Massey
An A/B test is designed to be binary. We want to know if one of our ideas results in an improvement in performance and advances our fortunes, or if it is a dog that lowers results. A purist will tell you that there is no reason to look at test results until statistical significance has been achieved. Thank you, Purist. However, there are other concerns to consider when a test is running.Read the full article