Your ultimate introduction to split testing

Posted on Educational, Marketing, Strategy

There are some things in business that you just can’t trust to your instincts alone. The world of social media is one of those things. Digital trends are often unreliable, and you can’t use what is helping your competitors succeed as an inspiration point. To make decisions you need data and numbers in black and white. The best way to do this, particularly online, is to use split testing.


What is split testing?


At its most basic, a split test is two different versions of the same message at the same time that are sent out into the world of the internet to see which one performs better, based on a combination of key performance indicators (KPIs) and conversions. The beauty of this method of gathering information lies in its simplicity. Anything can be split tested. You could compare the results of anything from colours and the use of emojis, to button sizes and your tone and style of language.


While we mostly use it digitally these days, the concept is actually over 100 years old, and has evolved exponentially. In the 1920s, split tests were used scientifically and mathematically for the first time in agriculture, where biologist Ronald Fisher performed split test experiments on his plants. Later, in the 1950s, the medical field began adopting the process. Then, in the ‘60s and ‘70s, we got split tests that resemble the ones we create today. Of course, they didn’t have social media back then, so their split tests would look something like an experiment comparing the results between sending out marketing letters versus post cards. 


Setting up a split test


Before you begin to design your split test experiment, you need to define your current business goal. Is there a specific product you want to sell? Do you want people to sign up for your newsletter? Do you want your customers to review your services online? Whatever your objective, your split test needs to mirror it, with you designing new experiments every time your goals change or you want to shift focus.


Let’s say, for example, you’d like potential customers to visit your online store. You could experiment with using a single, static creative as well as a carousel ad. From there, you would examine which one achieved the greatest click-through rate. Once you have acquired sufficient learnings, you can hone in this A/B process even further by testing whether shorter or longer ad copy gets more clicks.


It’s also important to remember to test how different elements interact with one another. For example, you may want to cross-examine how various age groups react to different iterations of your use of language in your CTA (call to action) buttons. Upon cross-referencing results, you may find that several CTAs will have to be segmented across your ad sets where audience targeting settings differ.


This continual process of gathering learnings will help optimise your ads to align better with your audience, subsequently achieving upgraded KPI metrics and great results.


What can you get out of it?


Aside from the obvious benefit of getting to know your customers and how they prefer to be communicated with and sold to, split testing is a great way to optimise your paid media strategy. This will result in a more focused customer journey for users, better return on ad spend and, of course, more conversions!


If you’d like to know more about how split tests work or need help getting started, contact us here.


READ MORE: Learn a lesson in paid media advertising