There’s a reason why more than half of companies A/B test their website.

A/B testing is how you find out what site elements resonate well with your audience and what tweaks you need to make to increase conversions.

While several articles out there tout a concrete list of A/B testing best practices, here’s the thing to remember:

Websites are contextual and audiences are nuanced.

There is no one-size-fits-all checklist for A/B testing a website. However, here’s a closer look at some proven strategies along with advice from experts that will help you.

1️⃣ Start with research

There are a million things you could A/B test on your website, but you don’t need to test everything.

What you end up testing will depend on your website, your audience, the problems/successes you’re experiencing, etc.

The only way to know what you should A/B test on your site is to dive deep into the research.

Peep Laja, the founder of CXL, says this:

“Deciding what to test comes from research. Every test is a hypothesis to solve an identified issue, not trying out things randomly. Companies should use a conversion research framework such as ResearchXL to identify what and where the problems are.”

Peep goes into depth on what the ResearchXL process looks like here.

2️⃣ Look at your current website, data, and KPIs

Part of your research process should include looking at your current website, data, and KPIs.

Use these resources to ask and answer:

• What is already working well for users? Where are users disengaged?
• Is there anything confusing on the site?
• How are users responding to current content?
• Is the website content eliciting a response? Is it the proper response?
• Is it difficult for users to get from point A to point B?

You can use website data, Google Analytics, sales data, user information, surveys, etc., to capture this information.

When you identify and answer these questions, you’ll know what questions to ask and what to test.

3️⃣ Start with small, single, and specific A/B tests

You may have uncovered many potential problems and written a huge list of things you want to A/B test.

That’s great and means you’re on the right track.

But it’s best to test one thing at a time.

Don’t underestimate the significance of setting up independent and strategic smaller tests as you get started, as Jon MacDonald points out.

4️⃣ Establish measurable KPIs to determine the success of A/B testing

Ant Pace, owner of Split/Wit, shares his main pieces of advice when it comes to A/B testing:

• Trust data over intuition
• Online test one thing at a time (a running theme here)
• Have quantifiable goals

I asked him to elaborate on what it means to have quantifiable goals. He said:

“Having quantifiable goals is all about using key metrics that make a difference to the business's bottom line—like sales or sign-ups. You can also use smaller KPIs along the conversion funnel, like bounce rates. Using the right software as you run an A/B test will make it easier to manage and track meaningful A/B test KPIs.”

When you define your KPIs before running your tests, you’ll have quantifiable results showing you which variation had the most significant effect.

5️⃣ Use an established A/B testing software

The best news for brands is you don’t have to go it alone. Several well-established A/B testing programs will walk you through the process.

Here are some top tools to consider:

• Google Optimizely
• Crazy Egg
• Omniconvert
• Optimizely
• UsabilityHub

All of these tools are worth checking out and will make it easy to A/B test elements on your site. 🤓