Sometimes, you know that sure variables, normally these that aren't easily manipulated, have a strong impact on the success metric. For example, maybe cellular users of your website are inclined to click on less on anything, in contrast with desktop users. Randomization may lead to set A containing slightly extra
