Skip to content

What Is A/B Split Test

You plan to make ice cream for your kid’s birthday party this weekend and you have an option to make either Vanilla or Strawberry, which one would you make?

You can make the assumption that Strawberry flavors taste better, therefore Strawberry could be the better choice. When children are happy, you will feel happy too.

But wait! That’s your assumption.

The better option would be to ask the children themselves, or just make both Vanilla and Strawberry and see which one of the flavors the children would pick.

In online marketing, the same principle applies. You will need to perform split tests to optimize your sales or opt-in pages. Optimization is important to maximize your opt-in or conversion rates. It helps you to generate more leads and sales and it increases your bottom line.

In the online marketing world, this is called A/B Split Test.

The letter A and B basically represents the 2 variations of the page that you use to optimize your conversions.

One can generate better results than the other or both can produce similar results. Split testing is comparing two versions of a web page to figure out the better performing variation.

When you run a split test, you want to test 1 thing at a time, so that you can get a concrete answer as to what changes influence the result that you get.

What Do You Split Test?

Typically, the more important elements you would to improve are usually:

  • Headline
  • Sub-Headline
  • Copy
  • Image
  • Font/Color

I rarely split test font or color as I am focusing on more important elements such as the headline, but if you have the time, you can always split test every element to get the best results.

How To Perform Split Test

When you first perform a split test on a page, start with a page-level split test then zoom down to the elements on the page.

For example,

Instead of split testing each and every element on your pages such as an image or headline, you would want to start with split testing pages with totally different layouts or designs. At this stage, the headline, or copy would not be split tested.

Variation A
Variation A
Variation B

Once you have a winner page, continue to split test different elements within the page such as headline, sub-headline, etc.

This allows you to work only with the winning page, and that saves you time and effort.

How Much Traffic To Get To Decide on Split Test Results?

As a general rule of thumb, you want to get at least 100 views or impressions on your page before picking the winner page. If your site gets a lot of traffic, you can set the target higher such as 500 or 1000 views.

The bigger the sample size gives higher the accuracy but you wouldn’t want to wait too long as your goal is to optimize as much and as quickly as possible, so 100 views and above is my preference.

Also, learn how to run split tests in Systeme io