The How, Why and When of Website A/B Testing

Nikole Haiar Nikole Haiar - March 26, 2015

A/B test can show how end users react to subtle or large-scale changes to a Web page. In today’s marketplace, companies cannot afford to overlook their online presence, including the brand’s website. Whenever new content is added or adjusted on this platform, it is critical to ensure that new features work appropriately and information is displayed as planned. One way to help guarantee the functionality of a webpage is with A/B testing.

What is A/B testing and how does it work?
According to testing software provider Optimizely, A/B testing involves performing trials on two versions of a page, including the control, or A, version, and the version that contains the changes, or the B version. These tests hinge on the measurement of live traffic to gauge how the different versions impact the website’s conversion rate.

The process begins by determining a goal that the firm will try to reach through specific changes to its website. Via A/B testing, the company can see how even a small adjustment impacts the way customers interact with the site, including if a sales increase results from the change.

This is just one example of how A/B testing can be used on a website. Teams can test a whole host of variations with this strategy, including everything from subtle adjustments in wording, to large-scale design changes.

Why carry out A/B tests?
At this point, many business leaders may be asking themselves, “Why take part in this process?” Often, it is a simple matter of providing evidence backing proposed changes. If users react favorably to the difference, it is proof that it was a change for the better.

For instance, if an e-commerce company decides to adjust its shopping cart process, it would be in the business’s best interest to A/B test the new checkout processes versus the old approach. Quicksprout contributor Neil Patel did just this on his website, where he ran trials of a two-step checkout process and compared it to a three-step process. While the information collected was the same in both cases, Patel broke down the fields differently, separating the checkout operation into three pages with the B version as opposed to two pages in the A version.

The result, which included 817 conversions, saw a 10 percent increase with the three-step process. This shows that although the change might be small and subtle, users did have a response to the variation.

When should A/B tests be carried out?
When building your own website, it is essential to test items like fonts, colors, locations of features and other elements to see which versions users prefer. This can help pinpoint the best options for attracting customers and raising conversion rates. With an existing website, it is important to test any changes made by the design team to gauge how the change impacts customers.

KISSmetrics suggested performing A/B tests on the following items:

  • Typography: The font style and size, colors and typefaces.
  • Calls to action: The placement, color and text.
  • Pricing schemes: Free trials vs. freemium vs. money back guarantees, the length of the trial, and pricing of each plan.
  • Landing page content : Short-form vs. long-form, video vs. text, etc.
  • Other general test: The number of columns on each page, background images and patterns, links and number of steps in the checkout process.

Overall, A/B testing can offer a range of valuable insights, including support for the changes being made. By gaining a better picture of how variations impact end users, decision-makers can craft the best website possible for their brand.

The Hostway Difference

Trusted
Relationships

Trusted
Expertise

Trusted
Infrastructure

Shares