Real insights require quantitative and qualitative data. Here at UsabilityHub the Five Second Test is great for uncovering how quickly and easily the messaging on your page is understood. Testing candidate landing pages with five second tests on a random audience will help you find out which words, images and hierarchies best communicate the value you’ve got to offer. This is qualitative feedback, and it will tell you what’s going through a visitors head when they first see your site.
But clarity is only part of the challenge, and once you’ve composed a page that clearly communicates your message in the first five seconds it’s time time to test it in the wild, with real visitors, and get some quantitative validation.
I’m talking about A/B testing.
In an A/B test, your visitors are randomly assigned one of two variations of a page, and the rate at which they complete some desired action (eg. sign up, subscribe to your mailing list, view the pricing page) is measured for both variations. Whichever variation performs better is the winner. It won’t tell you why one variation is better than another — just which one is better.
Thanks to affordable software, both five second tests and A/B tests are well within the budget of pretty much all website owners, and you don’t need to be a programmer to get the job done. Services like Convert.com let you set up A/B tests on your live website with a visual page editor, and push the results directly to Google Analytics. Here at UsabilityHub we’ll set up and run a five second test on your website automatically. It’s 2013, and this stuff is easy now.
But if you’re only running A/B tests, or only running Five Second Tests, you’re really only getting half of the story.
If you’re not running A/B tests you won’t know whether your landing pages (clear as they are) actually influence people to do what you want them to.
If you’re not running Five Second Tests you’ll struggle to find which elements on the page actually need changing, and get stuck at local maximums.
Getting the best results requires both qualitative and quantitative data. The good news is that getting that data has never been easier.