Analysing social media via A/B testing


Lately, I was working on a project for auditing social media presence for a brand “Emirates Canada”. While working on the audit I came to know about several tools that are really useful for any marketer to help their client or organization in measuring its effectiveness on certain social media platforms.


While digging through the tools I came to know about a testing technique known as A/B testing also known to be split testing which is sort of a marketing strategy which goes way back to the days before the internet even existed. For marketers, social media A/B testing is a form of real world research that allows an individual to test small variations in marketing materials in order to determine the most effective platform with respect to your specific audience.


A generic example of A/B testing is to analyse if a tweet will perform better with a video or a picture. Moving forward, if the video works better you one can test variations in the video too – by analysing the variations in the video on the parameters like short vs long, sound vs silent, animated vs human action and so on.



But how is A/B testing performed?

While using A/B testing technique, a marketer separates the audience into 2 random groups and show different variations to each other – afterwards the feedback received is then compared to calculate the most effective variation out of the two.


Nevertheless, its important to understand that split testing should be used only to test one variation at a time. Hence, lets’ say in order to test a couple of headlines, images or may be a different ad format an individual will need to conduct multiple tests. As if you test poles apart variations, you obviously won’t get the desired results and may end up making wrong assumptions regarding responses of different elements working together.