Test A / B is a method by which users are offered different versions of a product (web, app, software, …) to measure which one is more optimal and effective.

Tests A / B (A / B Testing) consist of making changes or variations in an element that may affect the behaviour of users.

The different versions will be offered randomly to the users, will collect the usage data of each of them and will be analysed statistically to evaluate which version is most effective.

This test method allows us to evaluate the effectiveness of many elements such as:

• The location of items (buttons, options, banners, …)
• Colours
• Words
• The images
• The visual structure
• The navigability

How should you use an A / B test?

To perform an A / B test we will create the segments that we want to analyse (elements that will be modified and analysed) in the original product.

Subsequently we will make the modifications of the segments by creating one or more alternative versions of the product and we will launch them so that they are randomly shown to the users.

A / B Testing software will collect the usage data of each of the segments created and will be statistically shown in a comparative way so that we can evaluate which version is most optimal for our objectives.

Let’s see an example:

1º DEFINE SEGMENTS AND PARAMETERS:

Segment 1: …

Segment 2 : …

Segment 3: …

Segment 4: …

Segment 5: …

2ª MODIFY SEGMENTS:

Segment 1 = Change text, size, color and position

Segment 2 = Increases size, changes color and position

Segment 3 = Increases size, changes color and position

Segment 4 = Change position

Segment 5 = Change format (from drop-down list to checklist)

3º LAUNCH THE TWO VERSIONS AND COLLECT USE DATA

4º ANALYSE THE DATA

The importance of design alternatives

To realise an alternative design of our product can be highly beneficial and reveal us key point and aspects that modifying them will increase its success.

When we design a digital product as a web page for example, we do it based on studies, traditions and customs and we design and implement the elements as we think they will be more effective.

But the behaviour of users is unpredictable and changing and what we considered as the best option may not really be the most effective.

An alternative design with variations in the elements can help us to verify that our web is designed, structured and implemented optimally or can be improved in some way.

We have a web page and we think it is really good but … what would happen if we changed the colours ?, and if we change a button? Would it improve our website ?, would increase the success in its objective ?.

An alternative design can give us an answer to these questions because perhaps what we believe is better is not the same as the users believe and, after all, we design and program digital products with the sole objective of being used by the Users.

How they affect results

The results of the A / B tests can reveal data of great importance for the optimisation of our web.

They can show us how the variation of an element increases its conversion rate.

With the results of the AB Tests we can make the necessary modifications to make our product more optimum.

Sometimes a small change generates great results:

The performance of an A / B test of a well-known electronic commerce helped to increase the registration of new users by 68% with the simple modification of a word:

The main page of the Ecommerce offered a button to register under the title “Register”.

They did the test to change the text of the button by “Create Account” and observed how the registry of new users was increased by 68%.

Take action with Checkealos:

• Blocked in the design of your eCommerce? Discover how can we help you.
• Know the needs of the users of your eCommerce from its initial phase and improve its usability before launching it to the market. Usability testing