Setting up an A/B test using Adobe Target

Adobe Target lets you efficiently perform A/B tests to evaluate whether the changes you make to your website generate the desired conversion.

Broadly speaking, Adobe Target houses your entire A/B test workflow within what is called a campaign.

A campaign assembles the locations (mbox) where you want to swap content, the various offers, and experiences (rules for what content goes in what locations). In other words, a campaign encompasses all the rules governing who gets to see a particular piece of content, where on the site they see it, what content they see, and the time frame for which they see it.

A campaign life cycle includes the following stages:

  1. Campaign strategy and design: Determine the goals for testing and identify who does what. This is done before making any modifications to the website.
  2. Campaign setup: Set up mboxes on the website (typically the IT team or a web developer does this), create offers, and organize the creative content for your campaign.
  3. Preparing the campaign: Put together the offers, determine what content is displayed where on the website, and perform quality assurance.
  4. Running the campaign: Launch the campaign and monitor it.
  5. Analyzing the results: Examine user traffic to determine how a particular variant fares in comparison with the original.
  6. Promote winner: If test results consistently indicate that your changes significantly improve conversion, declare the variant as the winner.

workflow_update_1

As you can see clearly, stages 2 and 3 may be impediments for non-technical marketers who want to quickly construct, deploy, and test their creative experiences on websites.

The new Visual Experience Composer is especially aimed at these marketers. It provides a guided testing and targeting workflow— a wizard-based framework, with built-in best practices, for implementing A/B testing and targeting effectively. This framework simplifies the testing process, which helps marketers make optimal use of Adobe Target’s capabilities.

Visual Experience Composer lets you create unlimited variations of content directly on a web page, including offer and image swapping, as well as color, text, and design changes.

Because Visual Experience Composer lets you make changes directly to your website in a staging environment, you can implement tests more rapidly. A simplified implementation reduces the overall time required to draw insights, which enables you to implement the right changes to your website early to maximize conversions.

However, if you are a technical person, you can set up an A/B test in the traditional way using Form-Based Experience Composer. This feature enables you to deliver standard Adobe Target A/B testing and targeting activities through emails, mobile apps, kiosks, and other modes that don’t work with Visual Experience Composer.

If you use Form-Based Experience Composer, modify your website code to set up the mboxes at offer locations. Then, design the offers and assemble them together in a campaign.

Setting up A/B test using Visual Experience Composer

  1. Log in to Adobe Marketing Cloud, and launch the new Adobe Target interface.
  2. From the toolbar, click Create Activity and select A/B Test from the list.
  3. From the dialog, choose Visual Experience Composer (default selection) and then click Create Activity. The website you want to test loads in the wizard.

Note: If the website does not load automatically, click the browser security icon on the URL bar and allow your browser to load unsafe scripts.

Unsafe Scripts

Note: To load a different web page, click the Settings icon from the toolbar and choose URL from the menu.

  1. Click the Edit icon beside the activity name field (by default Untitled Activity is displayed) and specify a title for your test.

Untitled Activity

Note: By default, Visual Experience Composer treats the current web page as Experience A (control experience). In addition, it presents a copy of the web page as Experience B for you to modify. If you want to modify Experience A, select it from the left pane.

  1. Click the Overlay icon from the toolbar. From the menu, and choose the type of component you want to modify.

Overlay

Note: You can choose to modify, images, text, locations, and so on. You can move components around the webpage, delete them, or add new ones.

  1. Choose a color to highlight components of a specific type that you want to overlay on the web page and confirm your selection. Based on your choice of color, the wizard highlights all instances of the component on the web page. For example, if you choose to overlay images on the web page and choose the red color to highlight the images, the wizard highlights all images on the web page in red after you confirm your selection.

Confirm Selection

  1. Select the image you want to change. From the menu, you can choose to edit its source code, move the image, delete it, resize it, replace it with a different image, and so on.

Swap Image

  1. To modify another image, select it on the web page and choose the desired menu option. To modify a different component, for example a link, repeat steps 5-7. Click Continue.
  2. Typically, you run equal traffic to all the experiences (recipes). However, to distribute traffic in a particular ratio, click the Edit icon from the All Visitors box and choose an audience segment for the test. Alternatively, click the Add icon to create a custom segment. Specify the ratio in which you want to split the traffic between the recipes.
  3. Click Continue to display the settings page.
  4. Under Activity Settings, specify the following:
  • Objective: Measurable objectives for the test, for example, conversion, revenue per user, and so on. In other words, express the objectives in terms of a conversion metric. For example, an objective could be to measure conversion against the new layout.
  • Priority: Priority you want to attach to the campaign.
  • Duration: Start and end dates. Choose specific dates or use the default (activation and deactivation dates).
  1. Under Reporting Settings, enter the following details, which you want displayed in reports:
  • Name: Name of metric that you want displayed in reports
  • Success metric: Select a conversion metric, for example conversion, revenue, engagement, and so on.
  • Advanced Settings: Specify how a conversion is counted.
    • Specify whether a converting user is considered for subsequent conversions or allowed to re-enter the campaign as a new user.
    • Specify whether a conversion count is incremented per entrant or based on user impressions

You can add additional metrics and specify the audience for the reports.

  1. Under Other Metadata, include your notes and observations as you progress through the test. You can have checkpoints for the entire duration of the test and record results when a particular checkpoint is reached. The advantage of having checkpoints is that as the campaign progresses and gathers traffic, the winning experience is clearly highlighted across the checkpoints.
  2. Save and exit the campaign.
  3. Launch the campaign from the Launch pad or the Campaigns list after performing the necessary quality assurance.

Setting up A/B test in the traditional way

For more information on the traditional (non-visual) method of setting up an A/B test, see Form-based experience composer.