What is Split Testing?
Split testing, also referred to as A/B testing or multivariate testing, is a method of delivering multiple versions of a web page in order to gather data about the efficacy of particular content (images, text, forms or video). Incoming traffic is distributed between the original (control) version and different variations without the user knowing that they are actually part of an experiment.
A/B testing involves testing two variations of a web page against each other to distinguish which version of the web page is the most effective.
Multivariate testing involves comparing multiple variations of a web page in order to distinguish which combination of variations is the most effective.
A/B testing is excellent if you are after quick insights regarding isolated page elements, and is best for websites with lower traffic volume. Multivariate testing tends to be a more intricate process. Given the multiple variations and possible combinations, multivariate testing requires more testing time and is better suited to high-traffic websites, in order to yield statistically significant results.
Why Split Test?
Perhaps your website is ranking really well in the major search engines, or maybe your search engine marketing and social media marketing efforts are generating traffic through to your landing pages, but that traffic that you have worked so hard for isn’t turning into conversions (clicks, enquiries, sales, signup…).
Conversion rate optimisation (CRO) is all about making the most out of your website visits, gathering data about your visitors’ interaction with your website to ascertain how your conversion rate can be improved. Split testing is an important part of CRO.
What can be tested?
Any element that can be moved on your web page can be included in your split tests, however before testing, like any good science experiment we start with an hypothesis formulated after analysing the behaviour of your current traffic perhaps using heat map analysis.
The most important on-page elements on a web page are likely to be:
- Call-to-action buttons / text / links
- Introductory headline
- Web page graphics
- Text content such as sales copy and product descriptions
How is it implemented?
Once an hypothesis has been agreed, the alternate content needs to be written or designed, and programmatically added to the target web page using special script that will deliver alternate versions of the content to different visitors. A returning visitor will always see the same content, as long as they are using the same PC & browser.
After a statistically significant set of test results have been gathered, the outcome of the test can then be analysed and if conclusive, the winning content can be permanently applied to your web page, for all visitors.
Running a simple split test on the position of a call-to-action, or the text used in your main headline can make a big difference to your bottom line. Repeating the process and testing another change, and another change, will ultimately lead to optimal content for the desired visitor action.
Split Testing is a tool used by the most savvy websites, and is now available to Contact Point clients – please get in touch if you have a web page that isn’t delivering the conversions you are seeking!