Cómo hacer A/B Testing en tu Ecommerce & Shopify

How to do A/B Testing in your Ecommerce & Shopify

Hitting the key point in an e-commerce in any online store is not an easy task. Nowadays, competition is stronger than ever, since there are many options to cover a need. So, which of all the options that exist on the market will the user opt for?

The answer seems obvious, whichever best meets your needs, but... what if not?

As we have said today, with the level of competition that exists, it is no longer enough to satisfy the user's need, since that is what everyone (or the majority) does.

Nowadays the importance is knowing how to capture the user's attention, making the purchasing process an experience. The key is to optimize.

What is A/B testing or A/B testing?

A/B tests are a digital marketing strategy that, when taking two different variants of our online store, tests are carried out by sending variant A to one part of our audience and the other variant B to another.

In this way, the data that we collect from this testing will help us to know which of those variants subjected to testing have obtained better performance and which is the best option to permanently implement in our e-commerce.

Optimizing the website based on testing is a good strategy used in digital marketing to get to know the user better. We will know better what your behavior is like on the web, what your preferences are and what motivates you to continue browsing our online store and the emotions by which you are guided on this journey. This is one of the most used techniques in Growth Marketing , in order to obtain solid results based on the hypotheses that are formulated.

However, an A/B test is not an easy task since we need a large enough sample for the results to be more precise or more revealing. That is, testing a group of 50 users will not be the same as testing a group of 50,000. That is why we will need a large database to carry it out and get greater performance from it.

How to develop an A/B test

Why use an A/B test

A/B tests greatly help to optimize a website. This type of testing is a fundamental part of the CRO or Conversion Rate Optimization methodology , where we analyze the current performance of an ecommerce measured by its conversion ratio. Once analyzed, corrections are detected that should be carried out on the website and also opportunities for improvement. Thus we move on to the brainstorming stage on which different hypotheses will be established according to the previous analysis. And once we have the list of hypotheses we will move on to develop the A/B tests and start measuring results to know which variations or approaches obtain better performance.

What is CRO and how it can enhance your online business

A/B testing is also used for email marketing. The procedure is very similar. We take the same email and propose different variations that we send to our audience. We will begin to collect information on KPIs such as the opening rate, access to the website from the email, which part they clicked on... In the emails we can test different copy, CTAs, Anchor text, images...

Finally, as we have seen, A/B tests aim to optimize the user's purchasing process and ensure that the user reaches the final objective, which is conversion.

Types of Testing

Until now we have talked about A/B tests since they are the most famous and most used, however, these are not the only ones that can be carried out to improve an ecommerce.

A/B tests are two versions of the same product that are launched to the public to analyze which version is the most appropriate to improve the conversion rate .”

Now, depending on how we execute these versions we can classify them into:

  • A/B or univariate test . In this type of testing, 2 practically identical versions of a website are presented, but with a single specific variable, be it a CTA, a copy, a button to click, a banner...

ab test

  • Multivariate test . This type of test tries to place multiple variables in our ecommerce at the same time and thus study which one offers us the best performance and begin to improve the conversion rate. Just like in the A/B test we can measure copy, banners, image layout... only now several variants at the same time.

multivariate test

  • Design Test . In this practice the versions are almost totally different. Different designs are tested at the same time to see which one obtains the best results in terms of user behavior. When we have results, we choose the one that has obtained the best performance and we will continue working on the optimization of that design.

design test

Steps to perform A/B testing 

To carry out A/B testing, we are going to pay close attention to the emotions that users have when browsing the web. In this way, detecting these emotions and working on them will take us on the right path to capture the attention of potential clients. We can divide the steps to carry out effective A/B tests into 4:

Step 1: Analysis of emotional competence

Unlike a regular competitor analysis where we focus on the product and features of your competitors, with an emotional competitor analysis the goal is to focus on the emotional side of things. With this we will analyze where the market is emotionally. To do this we can follow the following process:

We choose a series of competitors, they can be both direct and indirect competitors. And we score them according to four different parameters.

  • The first would be the type of message. What is your main message and strategy? What are they saying on their landing page?
  • The second is the color. Although it may not seem like it, color has an important emotional effect on us and we can use it to say or express a lot without actually saying it, we want to understand what type of colors our competitors are using.
  • Third is the image. They are one of the first things our clients and users notice. Many times, based on the first images that the user sees, they will quickly decide whether to continue browsing the web or not. This is why you really have to understand and analyze what type of image they are using and why.
  • The fourth is emotional triggers. How do our competitors try to make their customers feel? What emotions do they attack?


Step 2: Emotional SWOT

The emotional SWOT will be used to help us understand how our customers feel in relation to our brand and how they feel in relation to the industry.

Carrying out an emotional SWOT exercise can be done using polls, forms or surveys, for example, on your website to find out all this information that we want to collect.

Just as we use this to know the opinion of users, we can also use them to know the opinion of our internal staff.

Finally, with all the information we collect we will be able to detect the weaknesses, problems, opportunities and strengths that they perceive about our brand. In this way we can obtain ideas on how to carry out A/B tests in our ecommerce.

Step 3: Emotional Content Strategy

Once we have obtained information generated by the previous step, we take the weaknesses and we take the threats, and we see what the biggest problems that people have and how we can address them within our emotional content.

Therefore, once we have defined what the public expects from both us and the market, we will already know the emotional content on which we want to start developing our strategy.

Step 4: Test

After having carried out an exhaustive analysis of the competition, carrying out a SWOT test and having developed a clear content strategy, there would only be one thing left: take action. Now it's time to run the tests.

We can set in a table all the ideas that we want to include in the tests. This way we will have everything better organized and it can be more effective when it comes to obtaining results.

The first thing we will take into account will be the different types of emotional triggers on which we want to do the test.

Then, we will define the elements we need to use to represent these emotions.

We will define the words we can use to represent these emotions, the images and the colors.

Once all this information is defined, we will define all the variants that we are going to test and we will launch them to the user groups on which we will do the test.

Depending on what we want to test and the type of test we carry out, we will apply the previous steps to a greater or lesser extent.

It may interest you:

How to improve the sales funnel in your ecommerce or shopify store

Keys to improve the performance of your online store

The best 5 free templates for your shopify store

Consultant specialized in ecommerce