Split Testing: The Basics

Getting to grips with analytics for the benefit of your online business is crucial. But what do you do after arming yourself with all this data? You’ve tapped into important insights into how your business is performing, but what’s next? The answer – a campaign of strategic split testing. It’s important to acknowledge how all the data in the world is of no real value, if you don’t know what to do with it. Or as is the case with many, decide not to act on it. Nevertheless, fail to take advantage of the data you collect and you could be selling your business short. What Is Split Testing? In the simplest terms, split testing is a process that involves comparing possible changes to your online business like-for-like. Two different versions of your website are created – one that presents the changes you’re considering, the other staying unchanged. Customers are channelled to both versions of your website, with analytics enabling you to see whether the proposed change has the desired effect. Used strategically, split testing is a way of making sure every single modification translates to a noticeable improvement. It’s a case of capitalising on the simplicity of trial and error, using hard evidence to guide the decisions you make. In addition, split testing can be useful for avoiding errors in judgement that could prove counterproductive or costly. A Working Example In an everyday situation, split testing could be used to determine the best size, shape, layout and position for a CTA. The business looking to improve its effectiveness creates a new version of the CTA, or perhaps moves it to a different position on the respective page. Half of the traffic the website attracts is presented with the new CTA, while the second half sees the original CTA as it was. After a given period of time, the website owner studies the data collected by way of analytics. They see that while the original CTA attracted clicks from 15% of the total traffic, the new CTA resulted in a click-thru rate of 20%. This would therefore suggest that the new CTA is significantly more effective and can therefore be made a permanent feature of the website. Should the website owner wish to do so, they may then experiment with further modifications of the same CTA, using a similar process. Or if happy with the CTA, move on to different features and aspects of the website. Split Testing Applications In theory, split testing can be used to increase the effectiveness of just about anything on any website. Nevertheless, there are certain features and fixtures more commonly put through split testing, which include: 1.Banners. Marketing and advertising banners should be optimised and tested on a continuous basis, in order to ensure they deliver. 2.Headlines. Particularly on your most important landing pages, headlines hold unique power and influence. Split testing can ensure they perform at their best. 3.Promotions. Experiment with anything promotional on your website to see how your target audience reacts. 4.CTAs. As mentioned, even a slight adjustment to a CTA can make a big difference to its on-going performance. Irrespective of the size and nature of the business you run, an on-going campaign of split testing comes highly recommended.