In web development, change can often seem daunting, but a few regular tweaks to design and layout could have major implications. Aside from improving user experience, your conversions could increase as well as customer referrals. This process of optimising a website through incremental changes is possible through A/B testing.
For those unfamiliar with this scientific approach, it’s fairly straightforward and yields impressive results. Also known as split-testing, this experiment allows you to test two versions of a webpage simultaneously. These metrics can then be analysed to discover which version – A or B – offers the best chance of success in terms of conversion.
This performance metric is one of the few ways to understand your visitors and what they like or dislike. As an essential tool, A/B tests should be run frequently on landing pages. Here are some top tips on performing the tests without hiccups.
Change one element only
While your control page (A) always remains the same, version B allows you to be creative and change a wide array of elements. Many dive in head first making several alterations from editing font to switching out images and replacing the call to action button.
Although it can be tempting to get carried away, this will undermine the entire test. If conversion rates do increase, you won’t be sure which element was the causing factor. This is important, as you may want to use the information for other pages on the site. Equally, conversions may slip and it’s virtually impossible to pin point the exact reason. Always start by testing one element per test for identifiable results.
Start off big
You may only have one element per test to change but that doesn’t mean you’re limited to what you can choose. The flexibility of an A/B test allows you to experiment with virtually anything. From colour scheme to content layout, headlines to multimedia, the options are limitless. Avoid getting bogged down by starting with the largest or most visually striking elements.
If you’ve placed emphasis on a controversial headline, it may be worth testing different variations of size, colour and even wording. Alternatively, you may want visitors to watch a video introduction. Try testing its position on the page or the borders that surround it. After testing the main elements, you can always move onto the finer details if you feel it may add value.
Get the timing right
Be sure to run your A/B test long enough to provide valid results. Although the exact timeframe is largely dependent on your current traffic, it should usually run for at least seven days. Finish the experiment earlier and your results could provide false positives due to anomalies. Any changes that were shown to improve conversions could have been due to an external factor.
On the other hand, you won’t want to test for too long either. A poor performing B page could be impacting user engagement, turning away potential customers. You’ll also want to implement any improvements discovered as soon as possible.
A/B testing provides an invaluable insight into the user experience. Follow this guide and you’ll get great results and plenty to work with.