What A/B Testing Is

Relying on intuition to improve marketing campaigns, landing pages, checkout processes, emails, and any other part of an online business provides inconsistent results at best. Intuition and experience are useful in optimization strategy, but A/B testing is the best way to improve the performance of a website because it relies on data and provides results that can be statistically proven.

How Testing Works

A/B testing is possible with testing software that samples users on your website and shows them variations of the element or page being tested. The simplest version of an A/B test splits users into two groups, showing the original version of the page to the first group and showing the other group a new version of the page. This method tests new ideas in real-time, directly comparing performance, and allows for statistically valid results that can be relied on for business decisions.

usertestingprocess

A/B testing can be much more complex than a simple 50/50 split though too. Here are some other common types and methods of tests:

  • A/B/n tests use the same concept as a simple A/B test, but they introduce more variations to be tested. It is possible to run as many as 5 or 10 variations in some tests.
  • Multivariate tests test multiple elements on a page interchangeably, effectively creating 10s or 100s of unique versions of a page.
  • Complete page tests/redirect tests send sampled groups to completely different pages via a redirect. A normal test simply switches out the relevant content on the same page, but sometimes entirely new designs are better designs with a redirect.
  • Sampled tests use varying degrees of sampling instead of even splits among the test variations. This is often done to minimize the risk of a new variation underperforming significantly enough that it may harm the business.

More information about the different test types and when to use them can be found in our Test Types article.

Examples of Testing to Increase Conversions

Now that you know the basics of what testing is and how it works, it makes sense to cover some examples of using testing to improve your business.

Ecommerce Sites

Ecommerce sites are the most obvious candidate for testing because the goals and measures are clear. Ecommerce sites sell products and selling more of them, or at least more of the high profit margin products, is the goal of testing.

Examples of focus areas for testing:

  • Product pages where most visitors make the decision to buy or not. Consider testing product images, feature listings, descriptions, pricing areas, and call-to-action buttons.
  • Checkout pages where visitors can often get confused about shipping, creating an account, entering coupon codes, and a whole host of other issues. Consider testing form layouts, adding and removing steps, and reducing complexity.
  • Category pages where visitors sort products and skim listings for the products they’re looking for. Consider testing sorting options, listing information and layouts, and navigational options.
  • Promotional banners/sections where you may be promoting sales. Consider testing placement, emphasis, and messaging.

Lead Generation Sites

If the goal of your site is to collect leads, then optimization testing can be fairly complex depending on what is being done with the leads. In most cases, the leads are being sold to and the ultimate goal is revenue. As such, simply getting more leads may not be best, especially if the leads aren’t as qualified and ultimately don’t end in sales. The correct approach for testing on lead generation sites is to consider both lead quantity and lead quality. The focus for an individual test could be on just one of those aspects, but the impact on the other must always be considered.

Examples of focus areas for testing:

  • Messaging, including headlines and primary sales copy, are the first things that visitors read and consider.
  • Lead collection forms are where visitors make the ultimate decision to submit their information to you.
  • Thank you pages don’t need to simply say thank you and nothing else. After lead information has been collected, you can test different offerings, upsells, and directing visitors to other sales funnels.

Content Sites

Oftentimes content sites have goals besides the primary ad views, engagements, or pageviews. These metrics aren’t as easy to test with, but it can still be done. Things like newsletter signups, social shares, and product sales are easy to test with even if they are secondary goals of the site.

Examples of focus areas for testing:

  • Article headlines have become a focus area for testing on major newspaper and content websites. Consider testing multiple variations of article headlines to increase clicks, shares, and other measures of engagement.
  • Related article sections are key areas that keep visitors engaging further with content. Consider testing layouts, ordering, and the algorithms that determine which articles to show.
  • Homepages remain one of the most trafficked areas for content based websites. Consider testing layouts and article organization.

Next Steps

Now that you have the basics, learn more about each test type and finding what to test.