What is A/B Testing? - Chapter 3

What is the A/B Testing Process?

Last updated on December 21, 2016

A/B testing allows you first to create — and then measure — two variations of your marketing message. So, it’s easy for you to determine which message is winning you more conversions with your visitors.

The winning variation is calculated with the help of data and statistics, helping validate changes you’ve made on your website or your landing pages and takes the guesswork out of conversion optimization.

A/B testing can work for all marketing materials. You can test your website, marketing emails, and your landing pages easily with the help of A/B testing.

The A/B testing process

All A/B tests begin with the original version of your landing page (also referred to as the “control variation”). There are four general steps to A/B testing:

Step 1: Set conversion goals

Your conversion goal depends on the purpose of your landing page. Your goal could be anything from ebook downloads to email signups to free consultation requests. Goals are what determine the success of one variation over the other.

Step 2: Create variations

Variations are changes to the original landing page you make to see which change your visitors respond to the most. The process of creating variations varies depending on the testing software you choose.

Step 3: Start testing

When you begin your test, visitors are randomly sent to one of the two variations created. The conversion goal is measured every time a visitor lands on each variation. Some testing tools, like Instapage, give you the opportunity to duplicate, pause, transfer, or delete any variation at any time.

Step 4: Analyze results

Once your test reaches statistical significance, it’s time to analyze your results. Instapage, for example, offers users an easy-to-understand testing analytics dashboard complete with four primary metrics: visitors, conversions, conversion rate, and improvement.

A/B Testing Analytics Dashboard

Suppose you’re A/B testing your landing page CTA’s button color from orange to red…

In this particular case, the control variation will be the landing page with the orange CTA button. Next, you’ll create variation B of your landing page, with a red CTA button.

After you create both variations and start the test, the traffic coming to your landing page is divided evenly amongst the two variations. As soon as the end point is reached, one variation will “win” having scored the most conversions or received the most email sign-ups (or whatever else your testing goal is).