It’s amazing to think about how we used to make web site decisions. What looks good? What colors do you think the audience will best respond to? And eventually, after launch, we would dig through the analytics to see how visitors were engaging and converting on site.
NOW, we make intelligent decisions based on A/B tests. We recently launched an A/B test for a new web design for one of our e-commerce clients. We labeled the original design “A” and labeled the new site design “B”. We asked 20+ people and they all thought that design “B” was more visually appealing and that site visitors would convert at a higher rate. After a two-week A/B test, sending 10% of the site traffic to the new design, the results came in. And to our surprise, the old-school, very simple design outperformed the new, fresh design. The numbers speak for themselves.
So what’s next? We continue to tweak design and retest to deliver optimal results.
Eliminate the guesswork. Bring in the science.
Testing and tallying,
Robin
{ 2 comments… read them below or add one }
Now, that’s an interesting way of testing a new design. There’s only so much a control group will tell you, and when you do launch the new design, you always run into things you never thought you would. So this A/B method gives you the best of both worlds, or at least a little taste of it. Congratulations on your innovation.
I tried some simple A/B testing on my personal website but the results were inconclusive (mostly because I decided “Hey, let’s try A/B testing” at 10:30 some night). However, I think the concept is one of great potential. It can reveal a lot about a website, especially if you make small, but conscientious, changes to the design or content and measure that response.