A/B Testing: Optimize Your Website, Beat the Competition

Let’s face it: the web is getting old; it’s growing up. Plain HTML in a text editor has been replaced by WYSIWYG web development environments. Frameworks offer the pre-fabricated building blocks of complex websites. Even Great Aunt Betty can set up her own blog about vintage cookie cutters (okay, that might be stretching it).

It’s seemingly small changes that can optimize your site and allow it to reach its potential. Let’s take the case of a single button by way of example. Google recently changed the name of a button in their Google Drive web app from “Incoming” to “Shared With Me.” This may seem subtle, but it was enough to set the internet astir. It’s changes like this – making your site simpler, more user-friendly, easier to navigate – that can lead to a real boost in clickthrough rates, sales, and other critical site metrics.

What is A/B Testing?

A/B testing is a fancy term for experimenting with a website change or new feature. In a scientific experiment, we have a control and a variable that we are testing. Let’s call ‘A’ the control and ‘B’ the test case. ‘A’ represents your website as it looks right now, and ‘B’ represents a test website, meaning your website with a variable you’ve changed: text, color, image, or anything else you can imagine. Voila – we’re doing science on the web!

Splitting web traffic is the most common mechanism for A/B testing. This means that some of your visitors (maybe half) see your current site, while the other half sees your test site over the course of the testing period.

When you’re done running your A/B test you should have data on clickthrough rates, sales, number of visitors (and any other data you gather) for the following data points:

• Statistics for ‘A’ (current site) prior to testing
• Statistics for ‘A’ (current site) during split testing
• Statistics for ‘B’ (test site) during split testing

In this way you can account for variations over time and back up your web design decisions with some fancy – and useful! – data.

What Should You Test?

The short answer is anything you like! Practically speaking, though, it’s good to focus on A/B testing of specific elements of your website that may directly affect your conversion rate. The following are some key website components that you should consider testing:

• Headlines
• Title of your website
• Images (especially on the landing page)
• Logo changes
• Button text (for example “Free Trial” versus “Increase Sales”)
• Sales funnel (how customers make eCommerce purchases)
• Placement (where site elements are placed spatially)
• Calls to action

Well-designed A/B tests are worth the time investment up-front. A/B tests can reveal problem areas of your site – and show you what’s working well. Data-driven decisions are key to a professional web design strategy.

You’re already spending money to develop your website – take advantage of A/B testing to get the most bang for your buck!

Why Use an A/B Testing Tool?

With A/B testing, as with virtually everything else in web design and development, we have a variety of tools to work with. Many A/B testing tools are available to simplify the task and help you optimize your site without killing your web developers (I’m pretty sure they’ll thank you).

Testing tools can help you create a test website (often with simple WYSIWYG, or ‘What You See Is What You Get’ web design tools), collect and interpret data, help you figure out the duration of your testing, and even provide heat maps and other insightful data about your website performance.

Six Popular A/B Testing Tools

Here are six popular A/B testing tools to consider:

1. Unbounce
Unbounce features a “Drag ‘n Drop WYSIWYG Page Builder” to help you design your test pages in a snap.

2. KISSmetrics
More than just your basic A/B testing, KISSmetrics provides other useful information such as revenue reports, cohort reports to track changes over time, and funnel reports to show where customers get stuck in the sales funnel.

3. Optimizely
Optimizely offers multivariate testing (a little more sophisticated than basic A/B testing), and claims some big names on its list of clients.

4. VWO (Visual Website Optimizer)
As its name suggests, Visual Website Optimizer is all about simplicity with an easy-to-use WYSIWYG editor. VWO offers website heatmaps, promises easy setup, and touts 24/7 customer support.

5. AB Tasty
AB Tasty offers a range of tests from basic A/B to multi-page and multivariate, and allows for precise control over what pages and site visitors are included in your testing.

6. Google Analytics Content Experiments
Google Analytics is a powerful (and free!) platform that can support A/B testing, though you won’t have the customer support and personal feedback to be expected from a paid A/B testing service.

Final Thoughts

The principles behind A/B testing are as old as modern science. They’re also incredibly important for succeeding on the modern World Wide Web. So go head: ask a question, form a hypothesis, and get testing with an A/B testing tool!