Picture the scenario: as a start-up, you’ve figured out a way to create a great sales funnel which is generating you a ton of leads. Your subscriber churn rate is low, you’re seeing increasing amounts of consistent traffic every week – what could possibly be improved? The answer is an ostensibly counter-intuitive one: sales.
Lead to sale conversion is one of the most mysterious elements of running a business, shrouded in a dense fog of rumor and hearsay: everybody seems to have an opinion, but nobody can offer a concrete answer. After all, what exactly can you do to persuade your visitor to go from browsing at their own convenience to actually spending money on your website?
Split testing (also known by more technical names such as multivariate or A/B testing) is essentially a method of conducting experiments on your website with a view to improving some metric or other. Once the test has been set up, new visitors arriving to your site will be directed to one of two types of variations: a control, which was the original set-up, and one or more different versions of your existing page. The idea is to wait and see if any of the changes result in different behavior by the visitors – to a statistically significant degree. Once a conclusion has been drawn, you can modify your website’s content, confident in the knowledge that the results of the experiment will apply to the larger scale of 100% of your traffic.
Sound simple? That’s because, at the heart of it, it is – split testing has been used successfully by researchers in a whole host of different fields for decades. Don’t be fooled into thinking that this is a one-size-fits-all solution, however; with so many definable parameters, it’s possible to create experiments tailored precisely to your specific needs. Examples of variations on existing web content could include:
- reducing the number of elements in a page, minimizing clutter and keeping the visitor’s attention focused;
- changing the flow of the user experience – for instance, fiddling around with the point at which the call to action appears in the user’s typical journey through your website;
- or rewording different aspects of your website’s content in order to observe the way changes in copy affect different audiences.
Split testing is a powerful weapon in any business owner’s arsenal, but in order to maximize the benefits it’s important to make it a habit – scheduling regular trials of new split testing variations is an easy way to improve the overall statistical significance of the conclusions drawn. Ensuring that you spread the experiments throughout different sections of your website is important, too; you never know what might leap out of the data, and knowing that you’ve thoroughly examined your business can only make it easier to trust the resulting analysis.
Do YOU have any experiences with split testing you’d like to share? Let me know in the comments below – I’d love to hear your opinion.
