The Golden Rule of  Website Optimization: Conduct A/B Testing

by    |    Feb 15, 2019    |       5 min read

Want to know the best strategy for increasing conversion rates? Anxious to determine which website design is better for business? Testing used to be a high labor, high-cost enterprise, but in the world of digital, you can test different options online without a heavy investment of time or money. A/B testing is a basic strategy for website optimization that provides quantitative data you can use to improve your marketing campaigns – and increase ROI.

What is website optimization and how can A/B testing help?

Done effectively, A/B testing can give you important insights into customer behavior you can leverage to increase sales. A/B testing involves exploring how small differences in a campaign have an impact on the behavior, decision-making, browsing patterns, or conversion rates of your customers.

You can utilize A/B testing, for example, to figure out which of two different texts is better on a banner, or which one of two different titles is more effective on a promotional email. By testing both options with a controlled group of customers, you can determine which version is more effective.

Setting up an A/B Test

For A/B testing to be successful, make sure to consider the following factors:

  • Choose a Variable: It’s essential that you test one variable at a time. You can use your existing website version as the control and set up a second website for the purposes of the test. In the second site, make the change that you are considering – and then split your traffic between the two.
  • Define a Metric: You’ll also need to give some thought to your definition of success. For some businesses, success is measured in a number of downloads, while for others, the goal is to increase sales.
  • Select a Test Group: If you’re sending out an email or a newsletter, it’s important to prepare the test groups carefully. The goal is to make the two groups as similar as possible. Make sure each one has the same number of participants, and try to make the groups demographically similar taking into consideration gender, age, geographic location, and other factors that you consider relevant.

Setting up an A/B test is relatively easy, and there are free tools available that can help you such as Google Analytics’ Content Experiments. If you feel that you do not have the technical skills to set up the test yourself, consider hiring a CRO (conversion rate optimization) consultant to run the test and provide you with professional recommendations.


When you’re doing A/B testing, make sure to run your test over a defined period of time. Deliver each version of the ad in equal numbers and at the same time of the day. You might also consider exploring whether you can improve your ROI by testing which day or the week, or which time of the day, works best for your consumers.

The Why & the When: Some Concrete Examples

A/B testing can be applied to a wide variety of marketing scenarios:

  • The title of a newsletter or emails: With A/B testing, you can try out a title on 10% of your mailing list and establish which of two options gives you a better click-through rate – then use the “winning” title when mailing your newsletter or email to rest of your database.
  • The sign-up page on a website: You can play around with different versions of your sign-up page to determine which design increases the number of sign-ups. You may find that changing the color, size, or placement on the page of your call-to-action button makes a significant difference.
  • The wording of Internet advertisements: You can test different versions of advertisements (including banners, PPC, and AdWords), and you can explore the best phrasing and design of the landing pages on your website.

Some of the many variables you can test include titles, product descriptions, length & style of the text, offers, price, images, the design of Call to Action buttons, color schemes, the length & questions on forms, and page layouts.

And of course, don’t ignore the possibility of segmentation. This means that perhaps one version of an ad appeals to an older segment of your target audience, and a second version appeals to a younger group. You can leverage this kind of information to sharpen your marketing strategy and match customers to exactly the right products or service