Why Everyone Should Split Test Everything
By Richard Parkin
Split testing is a fundamental essential for ecommerce of any kind, giving you an opportunity to understand the messaging, marketing approach and aesthetic design that most appeals to your audience.
Despite the real, objective data provided by split testing, it’s often underused, with many overlooking the full range of possibilities and opportunities presented by split testing. You shouldn’t limit the scope of your testing to a single part of your business. Don’t take anything for granted – split test everything.
That may seem relatively simple, but maximizing the impact of your split testing can be more complex than it appears at first. In this blog, we’re taking a look through some frequently-forgotten approaches for split testing – and how they can benefit your business.
Split Testing Your Imagery
Whether it’s in emails or on-page content, imagery is always a way to make a direct, immediate impression for your visitors. Despite that, it’s fairly common for businesses to just assume that the imagery they’re currently using is guaranteed to be the best choice.
This isn’t necessarily the case. Imagery sets the tone for your entire pitch, so it’s absolutely crucial to ensure that you get this right from the start. If you’re leading with an image of a person, are they an aspirational figure for your demographic? If you’re presenting the reader with natural scenery, have you considered using different views and angles? If you’re frontloading the viewer with product imagery, have you tested different ways to represent the product?
All of these points genuinely make a difference for conversion rates. Take the time to experiment, and ensure that your imagery reinforces the points you’re making for your audience, and you can start making a real difference for your conversion rate.
This isn’t limited to still imagery. While editing the visuals in your visual content is naturally more time-consuming and complex, testing is just as critical here. Work with your CS team to get direct feedback on the imagery in your videos, from both buyers and non-buyers – work to locate specific points within your videos where different imagery may have a serious impact on conversion rates.
Why You Should Split Test Colors
Like your imagery choices, colors are a foundational part of how your site appeals to visitors, but often gets completely overlooked when aiming to optimize conversion rates.
Again, you shouldn’t consider colors to be a set, unchangeable part of your operation. While you may have designed a clear, consistent set of brand guidelines, these shouldn’t be set in stone if there’s any risk of them lowering your overall performance.
Simply adjusting the background color for certain elements on a site can significantly improve conversions. When advertising a medical-related product, for instance, switching the background shade for email header images from a pale red to a more friendly green resulted in a consistent 5-10% conversion rate boost whenever tested.
Of course, there’s many cases where the absolute opposite change could have a similarly positive effect. Different products and services benefit from different colors – split testing is often the only way to gain an objective understanding of what works best for your audience.
Split Testing Email Journeys
While you likely run split tests for the emails in your automatic journeys, the emails themselves aren’t the only thing worth testing.
Vary the amount of emails you send, the times you send them, and the way you follow up on your journeys – not just the content you deliver.
For some companies, a 15-minute delay before sending an abandoned cart email can make all the difference. For others, delaying a followup for a full day can dramatically improve conversion uptake. Again, running split testing is the only way to be absolutely sure what works best for your audience – don’t take anything for granted.
Why Repeating Split Tests Matters
Not taking anything for granted isn’t just about checking and testing the assumptions and understandings behind your eCommerce approach – it includes reviewing and verifying the results from the tests you’ve previously carried out.
Consider the ‘Replication Crisis’ – an ongoing issue wherein scientists find themselves unable to replicate the results of past experiments when conducting experiments in the exact same way.
While being unable to reproduce a result doesn’t mean that the original test was wrong, it does highlight the importance of running multiple, repeated versions of the same test before taking a result for certain.
Just because a split test has produced a clear result in one case doesn’t necessarily mean that it’s worth rolling out immediately. Take the time to retest and verify your results, never just assuming that you’ve found the correct answer.