Configure A/B Tests in Adobe Target for Advertising DSP Ads

Last update: 2023-12-16
  • Created for:
  • User
    Developer

Advertisers with Advertising DSP only

Adobe Advertising and Adobe Target make it even easier for marketers to deliver a personalized and connected experience across paid media and on-site messaging. By sharing signals between the products, you can:

  • Decrease site fall-through rates by linking customers’ ad exposure from DSP campaigns to their on-site experiences.

  • Establish A/B testing by mirroring on-site experiences with advertising messaging using Adobe Audience Manager exposure data and click-to-feed Target audiences.

  • Measure the impact of unified messaging on an on-site objective lift with simple visualizations in Adobe Analytics for Target.

See the following sections for the prerequisites and for instructions to set up click-through and view-through tracking, implement signal sharing between DSP and Target and set up an A/B test activity, and set up Analytics Analysis Workspace to view the test data.

If you have additional questions, contact adcloud_support@adobe.com.

Prerequisites

This use case requires the following products and integrations:

Step 1: Set up the Click-through Framework

Click-through framework

When you add DSP macros to a click-through URL (the URL displayed when a user clicks an ad and reaches the landing page), DSP automatically captures the placement key by including ${TM_PLACEMENT_ID} in the click-through URL. This macro captures the alphanumeric placement key and not the numeric placement ID.

Click-through URL appended to the landing page URL

(DSP Only) Add DSP Macros to Your Click-through URLS

Within Flash talking or Google Campaign Manager 360, manually update the click-through URL for each ad to include the macros required to capture AMO ID variables. The AMO ID variables are used to send click data to Adobe Analytics and to share placement keys for A/B testing. See the following pages for instructions:

Contact your Adobe Account Team and the Advertising Solutions Group (aac-advertising-solutions-group@adobe.com) to retrieve the required placement key and finalize the setup, and to make sure that each click-through URL is populated with the placement key.

Step 2: Set up the View-through Framework Using Audience Manager

View-through framework

By adding an Audience Manager impression event pixel in your ad tags and placement settings, you can create a test segment for additional view-through testing opportunities.

  1. Implement an Audience Manager impression event pixel in your ad tags and DSP placement settings.

    For instructions, see “Collect Media Exposure Data from Advertising DSP Campaigns.”

    Make sure you add DSP macros to capture all data you want the impression event pixel to pass back, including ${TM_PLACEMENT_ID_NUM} for the numeric placement ID.

    NOTE

    Click-tracking URLs include the ${TM_PLACEMENT_ID} macro for the alphanumeric placement key, instead of ${TM_PLACEMENT_ID_NUM} for the numeric placement ID.

  2. Configure an Audience Manager segment from the DSP impression data:

    1. Verify that segment data is available:

      1. Search for the signal for the key-value pair that determines at what level the segment users are grouped.

        Use a supported key with a value that corresponds to a macro that you added to the Audience Manager impression event pixel.

        For example, to group users for a particular placement, use the d_placement key. For the value, use an actual numeric placement ID (such as 2501853) that’s captured by the DSP macro ${TM_PLACEMENT_ID_NUM}.

        If the search results show user counts for the key-value pair, which indicates that the pixel was placed correctly and data is flowing, then continue to the next step.

    2. Create a rule-based trait for segment creation in Audience Manager.

      • Name the trait so that it’s easily identifiable within test activities. Store the trait in whichever folder you prefer.

      • Select Ad Cloud as the Data Source.

      • For the trait expression, use d_event as the Key and imp as the Value.

    3. Set up a test segment for the new trait in Audience Manager, selecting Ad Cloud as the Data Source.

      Audience Manager automatically splits the segment into a control group that receives the standard landing page experience and a test group that received a personalized onsite experience.

Step 3: Set Up an A/B Test Activity in Target for DSP

The following instructions highlight information pertaining to the DSP use case.

  1. Log in to Adobe Target.

  2. Create an A/B test:

    1. In the Enter Activity URL field, enter the landing page URL for the test.

      NOTE

      You can use multiple URLs to test view-through site entry. For more information, see “Multipage Activity.” You can easily identify top entries by page URL by creating a Site Entry report in Analytics.

    2. In the Goal field, enter the success metric for the test.

      NOTE

      Make sure that Analytics is enabled as a data source within Target, and that the correct report suite is selected.

    3. Set the Priority to High or 999 to prevent conflicts when users in the test segment receive an incorrect on-site experience.

    4. Within Reporting Settings, select the Company Name and Report Suite connected to your DSP account.

      For additional reporting tips, see “Reporting best practices and troubleshooting.”

    5. In the Date Range field, enter the appropriate start and end dates for the test.

    6. Add audiences to the activity:

      1. Choose the segment that you previously created in Audience Manager to test view-through audiences.

      2. Select Site Pages > Landing Page > Query, and enter the DSP placement key in the Value field to use the Target query string parameters for click-through audiences.

    7. For the Traffic Allocation Method, select Manual (Default) and split the audience 50/50.

    8. Save the activity.

  3. Use Target Visual Experience Composer to make design changes to the A/B test landing page template.

    • Experience A: Don’t edit because it’s the default/control landing page experience without personalization.

    • Experience B: Use the Target user interface to customize the landing page template based on the assets included in the test (such as headlines, copy, button placement, and creatives).

    NOTE

    For example creative test use cases, contact your Adobe Account Team.

Step 4: Set up Your Analytics for Target Analysis Workspace in Analytics

Analytics for Target (A4T) is a cross-solution integration that lets advertisers create Target activities based on Analytics conversion metrics and audience segments and then measure the results using Analytics as the reporting source. All reporting and segmentation for that activity is based on Analytics data collection.

For more information about Analytics for Target, including a link to implementation instructions, see “Adobe Analytics as the reporting source for Adobe Target (A4T)”.

Set up the Analytics for Target Panel

In Analysis Workspace, configure the Analytics for Target panel to analyze your Target activities and experiences. Keep in mind the following important pointers and information about your reports.

Metrics

  • Create a panel within the workspace specific to the Adobe Advertising campaign, package, or placement for which the test was run. Use summary visualizations to show Adobe Advertising metrics in the same report as the Target test performance.

  • Prioritize using on-site metrics (such as visits and conversions) to measure performance.

  • Understand that aggregated media metrics from Adobe Advertising (such as impressions, clicks, and costs) can’t be matched to Target metrics.

Dimensions

The following dimensions pertain to Analytics for Target:

  • Target Activities: Name of the A/B Test

  • Target Experiences: Names of landing page experiences used within the activity

  • Target Activity > Experience: The activity name and experience name in the same row

Troubleshooting Analytics for Target Data

Within Analysis Workspace, if you notice that activity and experiences data is minimal or not populating, then do the following:

Further Reading

On this page