How to A/B Test Your Website Design
You may have used A/B testing for Google AdWords or other marketing campaigns, but you should also consider using this powerful tool to test the effectiveness of your Web design. A/B testing ensures that your design not only looks great but also works well for your customers.
What is A/B Testing?
Your website is useless if it’s not working for you. With testing, you can figure out which design elements get visitors to stay longer, interact with the site, sign up for free materials or make a purchase. You can also discover how to get more conversions for your advertising dollar.
When you A/B test your website design, your customers are directed randomly to two different versions of your website. During the test, you collect data to see which website does a better job of accomplishing your goals. For example, placing a shopping cart button in a different location could increase the number of customers that make purchases on your website. Because you A/B tested the design, you can make a tweak that will increase your conversions.
“Do’s” of A/B Testing
To ensure that A/B testing is most effective, you should keep these key elements in mind:
- Only test one element at a time. If you want to test multiple design elements at once, then you can try multivariate testing. A/B testing can give you a lot of information in a short period of time. Multivariate testing will require your site to draw a large number of visitors before you can gather enough useful data.
- Start testing as soon as your website launches. Making changes in the beginning is much easier than reviewing the site after it’s been up for a while. However, even if you’re getting a late start, you can still learn a lot from A/B testing.
- Start by testing the elements that have the highest impact. The elements that have the biggest effect on conversion rates include headings, subheadings, opening paragraph, images, calls to action and offers.
- Get the right sample size. If you’re a high-volume website with a low average transaction, then you should get a larger sample size of about 100 customers. Alternatively, if you’re a low-volume website with a high average transaction, then you may have plenty of data after only 20 conversions.
- Choose a good A/B testing tool. Some popular choices include Google Analytics, Optimizely and Visual Website Optimizer.
“Don’ts” of A/B Testing
Avoid these common mistakes when you’re testing your design:
- Looking at only one metric. Many different metrics can signal success for your design. Avoid focusing on just one metric, like click-through rate, and look at additional statistics like sign-ups. If your click-through rate drops while your sign-up rate increases, then your design is achieving your goals in spite of the lower click-through rate.
- Forgetting to segment your tests. Some changes may affect new visitors to your website while having no significant effect on returning customers. For example, your returning customers already know where your shopping cart button is located, so you shouldn’t count them in your A/B test results. Instead, you would learn more by focusing your test on new customers only.
- Choosing beauty over function. You may A/B test two different versions of your home page: one with a nice photo and one with only text. Your results may show that the text-only home page gets more results, but you still choose the photo version because it looks better. Give up your aesthetic sensibilities and choose the version that works.
Infintech can create great designs that both increase your conversion rates and boost your sales. Give us a call today for a free quote.