The Complete Guide to A/B Testing


Practice makes better and testing makes perfect!
A/B testing allows every marketer or store owner to try two (or more) versions of a marketing copy and see which one outperforms the other, in order to optimize their marketing efforts for the best results.
In this article, we’ve covered the basics of planning and running an AB test, declaring a winner, and then analyzing the results of splitting a “control” vs. a “variant” and reading between the lines of this data.

What Is A/B Testing ?

How to A/B Test?

Why Should You A/B Test?

How to Plan for an A/B Test?

Tips for a Successful A/B Test

What Is an Inconclusive A/B Test?

What Are the External Validity Threats to an A/B Test?

How to Analyze A/B Test Results?

What Is A/B Testing?

A/B testing, or split testing, is the process of comparing two versions of the same content to determine which one performs better.

This content you’re comparing could be a webpage, an email , a call-to-action button, social media content, colors and designs, or even the layout of certain pages or sections.

AB Splitting helps online store owners and marketers assess how one piece of marketing content performs alongside another one.

Then, the version that gives higher conversions is the winning one, which can help you optimize your e-commerce website for better results.

How to A/B Test?

Basically, you have a marketing content or a copy version in hand, which you’re already using, and it’s called the “control.”

Then, you create another version of the same content, and it’s called the “variant.”

You show 50% of your traffic the first version, the control, and 50% of them the other version, the variant.

The version that results in the highest conversion rate wins, and you would push 100% of traffic or visitors to the “winner.”

Tip ? Do not measure the success of one version against the other in an AB test based on conversion rate . Track the value of a conversion all the way through to a complete sale.

You may like 9 Best Small Retail Business Ideas & Opportunities

Why Should You A/B Test?

As long as you’re running a website with high traffic, you should A/B test continuously to improve the effectiveness of your store and deal with many problems in the e-commerce process.

Increase Traffic to Your Online Store

putting into test different blog posts, webpage titles, visuals, and texts can change the number of people who click on that hyperlinked title to get to your website, which increases traffic to your e-commerce website as a result.

Solve Visitors Pain Points

Visitors come to your e-commerce website with a goal in mind: buy a product, read a review, request a demo, know the prices, compare features, and so on.

Your job is to help users reach their exact goal fast and easy for a good user experience and, hence, a successful conversion.

Some common pain points online shoppers face include not understanding a product description , not finding the CTA button, or struggling with the check out.

AB testing different versions of product description copies (maybe one with text and another with video) or two forms of check out (check out as a guest vs. login to proceed) can yield significant results that lead to a better performance with users and a higher conversion rate.

Tip ? A/B splitting can give you high ROI as sometimes, even the most minor changes can result in a significant increase in conversions

Reduce Bounce Rate

A high bounce rate is every website’s worst nightmare! There are many elements that can make a user bounce away from your online store too fast, like getting lost among too many options or finding that the content doesn’t match their expectations.

A/B testing helps you greatly improve that because you can test multiple variations of an element of your website till you find the best possible version.

Lower Cart Abandonment Rate

“Shopping cart abandonment” is a bitter fact that 70% of e-commerce sites suffer from. A/B test different product photos or videos, check-out page designs, and even where shipping costs are displayed, which can lower this abandonment rate.

How to Plan for an A/B Test?

There are a few steps to plan for an effective AB test that will actually give you accurate results and lead to a better performance for your eCommerce store sales.

1. Identify a Variable to Test

Before you start the AB test, you need to lay down all the variables you want to put to test in front of you, and then pick only one variable to start with.

A/B testing more than one variable in a single email or webpage will not yield accurate results for you to evaluate which one was responsible for changes in performance.

As we agreed above, even simple changes can drive big improvements.

So, pick one element from your marketing resources and put one or two alternatives to it. For example, the image in your social media post linking to a mega sale on your store.

If you decided to change both of the image and the call-to-action button in the same test between the control and the variant and you saw big changes in performance, you won’t know what’s responsible for this change, the visuals or the wordings.

2. Set a Goal

Before you proceed with the A/B test, set a definitive goal for the test – think about which metrics are important to you and how the changes you’re proposing might affect user behavior.

The goal of your test might be examining the results of a hypothesis you stated or a prediction you made.

3. Create the “Variant” to the Existing “Control”

Say you want to test adding a social proof section on your homepage and whether this will return a different outcome or not.

Then your existing homepage is the control, and the other page you’ll create is the variant.

Tip ? ExpandCart allows store owners to create an alternative to the store’s homepage for AB testing means, which is an exclusive feature to ExpandCart for stores built and hosted on an e-commerce platform.

Now you have a homepage with reviews from previous customers to test against your current homepage (with no section for reviews), and the “winner” will be the next “control.”

4.     Use an A/B Testing Tool

The most famous marketing platforms are Google and Facebook, and, fortunately, they both have their own tools that automate the whole process for you.

For Facebook ads for example, let’s say you’re running a campaign to drive traffic to your online store through a Facebook post. Facebook allows you to A/B test different versions of ad creative, variable audience-targeting strategies, or different placements.

Facebook explains it all step by step in this guide.

You can also AB test using Google analytics. They suggest you Test a change to a Call To Action (CTA), change the color of a button, or remove an extraneous form field.

Google analytics explains it all step by step in this guide.

Tips for a Successful AB Test

Before declaring a “winner” from your AB test, don’t be hasty and consider the following points:

  • Make sure you’re running one test at a time on any campaign so that you can decide exactly which change caused the increase in leads. Putting into test more than one thing for a single campaign can complicate your results.
  • Test both variations simultaneously at the same time frame, whether it’s a day, a week, or any business cycle. You need to know exactly whether the performance change was caused by the different design or the different timing.
  • Give the A/B test enough time to produce useful data. Otherwise, it’ll be hard to tell whether there was a statistically significant difference between the two variations.

What Is an Inconclusive AB Test?

Normally, one variation turns out to be statistically better than the other, and you declare it a winner.

But sometimes neither variation is statistically better – the variable you’re splitting didn’t affect the results greatly. This is called an “inconclusive A/B test.”

In this event, you use the failed data to run another test; you might want to re-think the whole planning process from the start.

What Are the External Validity Threats to an A/B Test?

The validity or results of your AB test can be threatened by external factors like

  • Black Friday
  • Launching a major paid campaign
  • The weekend
  • Changing seasons
  • World events

If your AB test coincided with an external major event, you cannot depend on its results or data nor call a winner/loser variation.


Because the results have been impacted by an external factor: sales on Black Friday (some countries call it White Friday) rocket to the sky regardless of any copy version or CTA placement, paid campaigns drive a huge traffic to the website no matter what the design is, sales are usually high during the weekend days or before holidays, and (as we’ve seen this year) events happening in the world can certainly change the performance of your e-commerce website positively or negatively disregarding any test or variation.

What happens here is that the data you based your test decision on was an anomaly. When things settle down afterwards, you might be surprised to find your “winner” losing.

Can you eliminate external validity threats?


But you can definitely mitigate external threats by running tests for a full week, measuring results from organic traffic only, avoiding seasonal and international events, and being mindful of potential threats.

You may like What Is Customer Acquisition and Customer Retention ?

How to Analyze A/B Test Results?

Number 1, you identify the winner based on the results from the AB splitting tool you’re using and declare that the control loses to the variant.

Number 2, do not dismiss the loser all in all; learn from the results and gain insights you can use for future tests and in other areas of your business.

Look for other segments or areas where your loser performed well, maybe driving more visitors from certain places (Chrome vs. Safari visitors, iOS vs. Android visitors, etc.).

So, segmenting your data, even from the losing variation, will help you find hidden insights below the surface.

To Sum Up,,,

Although A/B testing is not that complicated to learn and run yourself, many store owners ignore it as a technique for better results to their online stores, as increasing traffic, solving pain points, reducing bounce rate, lowering cart abandonment rate, and more.

Planning for an AB test includes setting a target, designing a variant to your existing control, and using an A/B splitting tool that fits your needs and skills.

Running an AB test will give you deep insights about your e-commerce website, user behavior, purchase journey, positive and negative elements on your store, and so much more depending on the elements you put to test.

Read the results of the test carefully and segment the data you get so that you can optimize every aspect for a better performance and a higher ROI .

Build your online store with ExpandCart to enjoy marketing consultation, 24/7 customer support, and a full marketplace of apps & services tailored to every eCommerce business.

No credit card required and no hidden fees.


Comment Notes