What is it and How to Create A/B Tests (Premium Resource)

With A/B testing you can expand your business and maximize your success comparing different versions of your website or app to detect the most successful version that increases your sales, revenue, conversions, page views and more.

What is an A/B Test?

An A/B test is a method of comparing different versions of your websites or apps and seeing which variation makes it more successful. A/B tests are also known as experiments or split tests. In an A/B test, you show two or more variations different for your users (visitors) and the variation with the best performance wins. When a user joins the experiment, a variation will be chosen at random and the user will see that variation on all subsequent visits. Experimenting this way allows you to maximize your success.

How to Create A/B Tests

To access the "Manage A/B Tests" screen, click on the Administration icon at the top right corner. Then click on "Manage A/B tests" in the left menu and, finally, click on the Create an A/B test now button. For each website or application, different tests can be created.

Throughout the process of creating and running your tests, MVMCloud Analytics will show a help text explaining what each element means. This is especially useful when you haven't run A/B tests before.

Abaixo você pode ver a descrição de cada campo a ser preenchido e um exemplo de prenchimento:

Field Name	Description
Name: landing-page-test	The name is a unique name for this A/B test. The name you choose can be visible during A/B testing to your users in the source code or URL. Use only alphanumeric characters without any spaces or special characters. A maximum of 50 characters are allowed.
Hypothesis: If I change the color of the buy now button, then I hope to sell more products because the button will be more visible.	The hypothesis explains what you predict will happen when you run the A/B test, what the result will be, and why it will happen. The hypothesis is a necessary step in defining your A/B test and we recommend that you spend some time thinking about it.
Description: Compare blue and red color in buy now button	This field is used to describe what things will be compared.
Variations: blue-color	Variations are the name for any new version (variation) that will be tested against the original (current) version. For example, if you want to test how the difference in button colors behaves with each other, you can create a variation for each color you want to compare. Variation names will be visible to your users in the source code or address. You can only use alphanumeric characters without any spaces or special characters. A maximum of 50 characters can be used per variation name.
Target-page: Visitors will participate in A/B testing on any page	By default, an A/B test will run on all your pages. Alternatively, you can choose to run the A/B test on just a specific page. If you specify a domain, including the optional path, the A/B test will only run on that page. Once the A/B test has been created, you can include or exclude more pages.

Comparing the performance of different page URLs

If you want to experiment and compare the performance of different page URLs, you can create your experiment by specifying a redirect page URL for each variation.

When creating your experiment, in the “Redirects” section, for each variation (including the “Original” variation), you can enter the URL of the page to redirect and test.

Here's how it works: When users access the original page URL, they will enter the experiment (and receive one of the redirect URL variations). Then they:

1. They will remain on the current page (if the original variation was chosen) or;
2. They will automatically be redirected to the redirect URL assigned to them.

When creating an experiment using “Redirect URLs”, be sure to include the MVMCloud Analytics tracking code on all your pages, including on redirect URL pages.

Starting your experiment

To start your experiment, click “Embed Code” and embed the displayed code into your website, app, server or campaign. The experience It will start running automatically as soon as the first user enters the experience. Alternatively, you can schedule a “Start Date” by clicking “Schedule”. To learn more about how to implement it, see the A/B Testing Guide - Experiments.

Finishing your experiment

After running your experiment for a while and coming to a conclusion about your results, it's time to end your experiment. An experiment can be completed by anyone who has at least admin access to a specific website. To mark an A/B test as complete, select the "Completed" status at the top of the Manage A/B Test page. or click "Complete Experience" when editing the experience in the "Manage Experiences" section.

After you complete your experiment, you must remove all code related to this experiment from your website or app.

You will still be able to view the report of your completed experiment as long as you have not archived it.