Skip to main content

Creating content tests with Bandito

Use the Content testing with Bandito feature in PageBuilder Editor to test different content variants and determine which performs better with your audience.

Content testing is an add-on feature that costs extra. Contact Arc XP Customer Support if you are interested in using this functionality.

Tests in PageBuilder apply to blocks and chained blocks, not article content itself. While headline testing is common, you can test any dynamic block, including photos, bylines, headlines, or story blurbs.

A successful content test depends on three factors:

  1. The number of variants to test.

  2. The traffic or total click count your story could generate.

  3. The time during which your test is active.

Ensure that the page where you perform the content testing has high traffic; otherwise, the test might fail or not perform correctly.

We recommend defining the maximum number of variants based on the expected total click count:

  • Less than 100 clicks - we don't recommend using content testing.

  • 100 to 500 clicks - use a maximum of two variants.

  • 500 to 2,000 clicks - use a maximum of four variants.

  • 2,000 to 5,000 clicks - use a maximum of eight variants.

  • 5,000 or more - use a maximum of 16 variants.

These guidelines help ensure your tests gather enough data per variant to produce reliable, actionable results.

Important

Content testing with Bandito works only on the default PageBuilder Editor experience. Secondary Micro Experiences are not compatible with Bandito content testing.

As a sports editor for a news organization, you want to improve user engagement with your latest football match coverage. You enter PageBuilder Editor, locate the article and its headline block "Team A Wins Event!", and create a variant: "Team B Narrowly Defeated!".

You publish the page, and Bandito begins testing both versions with your audience. It directs traffic between your original headline and the new variant and continuously monitors performance metrics. The system adjusts traffic distribution based on real-time engagement data, identifying which headline resonates more with your readers.

Within five minutes, Bandito detects that "Team B Narrowly Defeated!" outperforms your original headline, with higher click-through rates. The system automatically routes most traffic to the winning variant.

Later, you return to PageBuilder Editor, set the successful headline as the default, and end the experiment.

Procedures

You must create a content variant to start a content testing process. To create a content variant, complete the following:

  1. Navigate to PageBuilder Editor > Pages.

  2. Click the page name of the published page where you want to start your content testing. The Curate workspace opens.

  3. In the Blocks card, locate the block you want to test and click the ellipsis icon ellipsis_icon_pbe.png > Create Variant.

    A new variant-1 appears under the original block. The new variant has the same values as the original.

    Tip

    The original block is the Control Variant, and any subsequent variant has a suffix number, for example, variant-1, variant-2. You must have a Control Variant and at least one variant to create a test.

  4. Click the variant-1 name to open the Block Details pane.

  5. Update the variant details, including the following:

    • Display Name - a customizable name that allows you to rename blocks for better organization and identification. Changing this name does not affect the block functionality or how it is displayed.

    • Custom Fields - a set of values, depending on the block, defined through the PageBuilder Admin interface.

  6. Create and update more variants, if needed, following the previous steps.

    To preview the variant's look, click the variant name, and its preview displays in the Editor preview zone.

  7. Click the Publish workspace icon icon_publish_workspace.png. The Publish workspace opens.

  8. In the Draft Options section, click Share & Publish. The page publishes on your site.

Content testing starts after you publish the page. A green confirmation message appears at the bottom of the page.

You can update content tests after creating them, whether they have been initiated or not. To update a content test, complete the following:

  1. Navigate to PageBuilder Editor > Pages.

  2. Click the page name of the published page where you want to start your content testing. The Curate workspace opens.

  3. In the Blocks card, locate the block you want to test and click the ellipsis icon ellipsis_icon_pbe.png > Create Variant.

    A new variant appears.

  4. Update the variant details.

  5. Click the Publish workspace icon icon_publish_workspace.png. The Publish workspace opens.

  6. In the Draft Options section, click Share & Publish.

Each visitor sees the appropriate variants as long as the content test is active.

PageBuilder visually marks when a content test is not yet initiated, is ongoing, and has a winner. The following icons help you track the process of your content tests:

  • The flask icon icon_flask.png marks that a test has been created but has not yet been initiated. Hovering over it shows a tooltip reminding you to publish the page to initiate the test.

  • The branch icon icon_branching.png marks that the test is in progress. Hovering over it displays a tooltip detailing the test length and leading variant rates.

  • The trophy icon icon_trophy.png marks the winning variant. Hovering over it displays a tooltip detailing the winner's CTR (Click-Through Rate) percentage.

When the content test determines a winning variant, you must mark the test as finished. To finish the content test, complete the following:

  1. Navigate to PageBuilder Editor > Pages.

  2. Click the page name of the published page where you started the content test. The Curate workspace opens.

  3. In the Blocks card, locate and expand the block where you performed the content test.

  4. Click the winning variant. A modal opens.

  5. Click Yes, Close Test to automatically close the test, or Cancel to close the test manually.

    Depending on your choice, the system performs the following actions:

    • Close Test - when you choose to automatically close the test, the system:

      • renders and retains the winner variant as the default block.

      • removes the trophy and branch icons from the block.

      • removes the non-winning variants from the Blocks card.

    • Cancel - when you manually close the test, you have two options:

      • Keeping the control variant:

        1. Click the ellipsis icon icon_ellipsis.png next to each variant.

        2. Click Remove. After you remove all variants, the control variant becomes the default option, and there is no Content Test option for that block.

        3. Publish the page to reflect the changes.

      • Keeping a variant:

        1. Click the ellipsis icon icon_ellipsis.png next to the variant you want to keep.

        2. Click Copy Variant as Block. A new block appears in the Block card, with the same configuration as the variant.

        3. In the control variant, click the ellipsis icon icon_ellipsis.png > Remove. All the variants associated with the control variant are also removed.

        4. Publish the page to reflect the changes.

Important

Remove all variables and close the test to prevent it from running in the background. Unclosed tests continue to operate and leave variables visible on the Editor UI.

Close the test to prevent the overhead of maintaining it after determining a winner and to keep non-winning variables from cluttering your Editor UI.

As a reminder, when you refresh your page or return to the page, you are prompted to close any tests that have not been closed even after a winner has been found.