Skip to content
Let’s Talk

Growth

A/B Testing for Websites: How It Improves Conversions

Guessing which page design works costs you conversions. This guide explains how A/B testing for websites works, what it improves and how to run tests you can trust.

By Lisa3 min read

Side-by-side yellow door number 31 and green door number 33 with matching knockers and letter slotsA/B Testing for Websites: How It Improves Conversions

A/B testing, also called split testing, compares two or more versions of a web page to see which design, content or feature performs better. Instead of guessing what visitors want, you show each version to a share of your traffic and measure the results. Here is how A/B testing for websites works, how it improves conversions and how to run tests you can trust.

How A/B Testing Works

You split visitors at random into groups. One group sees the current page (the control, or version A). Another sees a changed version (the variant, or version B). After enough visitors have taken part, you compare a single main metric, such as purchases, sign-ups or form submissions, and keep the version that wins.

A/B test diagram: half the audience sees version A (27% conversion), half sees version B (8% conversion)

Why Small Changes Can Matter

Testing often reveals that ideas nobody prioritized have the biggest effect. In Harvard Business Review, Ron Kohavi and Stefan Thomke described a Microsoft Bing test in which a simple change to how ad headlines were displayed, one of hundreds of ideas proposed, increased revenue by 12% without hurting user experience measures. Without a test, nobody would have known.

The same article makes the opposite point too: many ideas that seem obviously good make no difference, or make things worse. Testing protects you from shipping those.

What A/B Testing Improves

Conversion Rates

The main goal of most tests is a higher conversion rate: the percentage of visitors who complete a desired action, such as buying, subscribing or filling out a form. Test headlines, calls to action (CTAs), forms, page layouts and offers.

User Experience

Testing lets you base user experience (UX) decisions on evidence. Try variations of navigation, layout and content order to find what helps visitors get where they want to go. Our UX and UI design team often pairs research with testing for exactly this reason.

Content and Messaging

Test headlines, copy and images to find the messages that drive engagement. Testing also works well as part of your lead generation strategy, for example on landing pages and sign-up forms.

Personalization

Create versions of a page for different audiences, such as new and returning visitors, and test which delivers the more relevant experience.

Mobile Experience

Most web traffic is now mobile. StatCounter puts mobile at 58.99% of worldwide web traffic in September 2026, against 39.26% for desktop. Test your mobile pages separately, since what works on a large screen often fails on a small one.

How to Run an A/B Test You Can Trust

  1. Start with a hypothesis. For example: "Showing delivery times on the product page will increase add-to-cart rate, because shoppers worry about delivery."
  2. Pick one primary metric before the test starts, and don't change it halfway through.
  3. Change one thing at a time so you know what caused the result.
  4. Run the test long enough. Decide the sample size in advance and run for at least a full business cycle, usually one to two weeks, so weekday and weekend behavior are both included.
  5. Don't stop early because a result looks good on day two. Early leads often disappear.
  6. Record and share what you learn, including tests that lose. They stop the team repeating the same idea.

Choosing an A/B Testing Tool

If you used Google Optimize, note that Google sunset Optimize and Optimize 360 on September 30, 2023. Teams now use dedicated experimentation tools, many of which integrate with Google Analytics 4, or testing features built into their website platform. Choose a tool that fits your traffic, your team's skills and how easily you can publish the winning version.

Benefits of A/B Testing for Websites

  • Better conversion rates: each winning test lifts the share of visitors who act.
  • More revenue: higher conversion on the same traffic means more sales without more ad spend.
  • Better user experience: you keep the changes visitors actually find easier.
  • Evidence over opinion: decisions rest on results, not on the loudest voice in the room.
  • Continuous improvement: small, repeated wins add up over time.

To act on what your tests show, you need a platform that makes changes easy. Apex® brings publishing, personalization and analytics into one place, so you can update a page and see how the change performs.

#A/B Testing for Websites#DXP Features#Website Testing

FAQs

Questions, Answered

Can’t find what you’re looking for? Our team is happy to help.

Ask Us Directly
What is A/B testing for websites?

A/B testing, also called split testing, compares two or more versions of a web page to see which design, content or feature performs better. Visitors are split at random between the versions, and the results on a chosen metric show which version to keep, replacing guesswork with evidence.

How does an A/B test work?

In an A/B test, visitors are randomly split into groups. One group sees the current page, called the control or version A, and another sees a changed version, the variant or version B. After enough visitors take part, you compare one main metric, such as sign-ups or purchases, and keep the winner.

What should I A/B test on my website first?

Start with elements that directly affect conversion: headlines, calls to action, forms, page layouts and offers. Landing pages and sign-up forms are good early candidates because small changes there can shift lead generation. Base each test on a clear hypothesis about why the change should help.

How long should an A/B test run?

An A/B test should run until it reaches a sample size decided in advance, and for at least a full business cycle, usually one to two weeks, so both weekday and weekend behavior are included. Stopping early because a result looks good on day two often produces false winners.

Why should I change only one thing per A/B test?

Changing only one element per A/B test lets you know exactly what caused the result. If you change the headline, image and button at the same time, a winning variant tells you something worked but not which change to credit or reuse on other pages.

Can small changes really make a big difference?

Yes, sometimes. In Harvard Business Review, Ron Kohavi and Stefan Thomke described a Microsoft Bing test where a simple change to how ad headlines were displayed increased revenue by 12%. The same article notes that many ideas that seem obviously good make no difference or make things worse.

Should I test mobile pages separately?

Yes. Most web traffic is now mobile, and what works on a large screen often fails on a small one. Testing mobile pages separately shows how layout, forms and calls to action perform for phone visitors, rather than letting desktop results hide mobile problems.

What replaced Google Optimize for A/B testing?

Google sunset Optimize and Optimize 360 on September 30, 2023. Teams now use dedicated experimentation tools, many of which integrate with Google Analytics 4, or testing features built into their website platform. Choose a tool that fits your traffic, your team's skills and how easily you can publish the winner.

Should I record A/B tests that lose?

Yes. Recording and sharing every result, including losing tests, builds a library of what does and does not work for your audience. It stops the team from repeating ideas that already failed and helps sharpen the hypotheses behind future tests.

What are the benefits of A/B testing for websites?

A/B testing improves conversion rates, raises revenue from the same traffic without more ad spend, keeps the user experience changes visitors actually find easier, and replaces opinion with evidence. Small, repeated wins add up over time into continuous improvement.

How does Apex® support A/B testing?

Apex® brings publishing, personalization and analytics into one place, so teams can update a page and see how the change performs without moving between tools. That makes it easier to act on test results and publish the winning version quickly.

Keep Reading

More From Cyberdunes

All Articles
Put It Into Practice

Want Help Putting These Ideas to Work in Your Business?

Talk to an Expert Read More Articles