Pages

Modification content defines the code that is added to a page to change the contents, appearance, or behavior of the page. The code can consist of HTML, CSS, and JavaScript, in any combination. For example, the content can be a DOM element that is displayed to visitors, such a hero banner or a product carousel, or it can be a script that adds invisible functionality to the page, such as data tracking or event handling.

A modification can have one or more variations. Variations are different content versions of the same modification. If a modification has only one variation, all visitors get that variation. If a modification has multiple variations, you can define how the variations are distributed among visitors, that is, what percentage of visitors gets which variation.

Apart from multi-armed bandit modifications, a visitor is assigned to a specific variation until they clean their browser cache, open a new window in incognito mode, or visit the site from another device.

You can create three types of variations:

  • Custom content. The content is defined directly in the variation. The content is only available to that variation.
  • Content template. The content is defined outside the modification in a template. A variation links to the template for its content. The template, and therefore the content, is also available to other modifications and variations on the site.
  • Original content. No content is defined for the modification. The variation is used for tracking the performance of the original page content. This allows you to compare how the other variations perform against the original content.

On this page:

Creating a variation

To define content for your modification, you must first create a content variation. The number of variations you can create for a modification depends on the selected modification case:

  • A/B test: Create two or more variations to test which variation performs best, and manually define the variation distribution.
  • Multi-armed bandit: Create two or more variations. The selected multi-armed bandit algorithm decides the best-performing variation and adjusts the variation distribution accordingly.
  • Personalization: The modification can only have a single variation. The variation is pre-created with the modification, and you can modify it, but not remove it.

By default, 10% of visitors are automatically included in a comparison group that will get the original content of a web page, not the modification. To disable the comparison group for the modification, deactivate the comparison group variation.

To create a variation:

  1. In the Basic settings view, click Create variation.

    Creating a variation

    If the modification case is Personalization, the modification can only have a single variation. The variation is pre-created with the modification. You cannot add more variations.

  2. Define the content type and content settings for the variation.
  3. When you're done, click Save.

Editing a variation

To edit a variation:

  1. In the Basic settings view, find the variation you want to edit, and click its name.
  2. Define the content type and content settings for the variation.
  3. When you're done, click Save.

Removing a variation

If the modification case is Personalization, the modification can only have a single variation. The variation is pre-created with the modification. You cannot remove the variation.

To remove a variation:

  1. In the Basic settings view, find the variation you want to remove.
  2. Click the quick menu button for the variation, and select Remove.

    Removing a variation

Defining the content type and content settings for a variation

To define the content type and content settings for a variation:

  1. Optionally, enter a title for the variation.

    Entering the variation title

    If your site has a Google Analytics integration with the Frosmo Platform, the variation title is shown as a label in Google Analytics. If you change the title in the Control Panel, the label will also change.

  2. Select or change the content type:
    1. Click Select type.
    2. Select the content type.

      Selecting the content type for a variation

    3. Click Select.
  3. Define the content based on the selected content type. The available content settings depend on the selected content type and template.
    1. Custom content: Define your own CSS, HTML, and JavaScript content in a code editor.

      When using CSS and JavaScript in variation content, the code must be wrapped in <style> and <script> elements, respectively.

    2. Original content: Create a variation for tracking the original page content. Since the variation is only used for collecting data, you don't define content for it.

      To track original content, the display method of the selected placement must be Replace content, Replace element, or Hide & replace original. You cannot append a variation to the original content.

    3. Content templates: Define content based on a predefined template.

  4. When you're done, click Save.

Activating and deactivating a variation

Before activating the modification and thus making it visible on the site, you must activate each variation you want to use.

By default, 10% of visitors are automatically included in a comparison group that will see the original content of a web page, not a modification. Once a visitor is included in the comparison group, that visitor will not see any modifications. The comparison group is treated as a variation and you can deactivate it as any other variation.

To activate a variation:

  1. In the Basic settings view, find the variation you want to activate.
  2. Click the quick menu button for the variation, and select Activate.

    Activating a variation

To deactivate a variation:

  1. In the Basic settings view, find the variation you want to deactivate.
  2. Click the quick menu button for the variation, and select Deactivate.

    Deactivating a variation

Defining the variation distribution for an A/B test modification

You can manually define the variation distribution only when the A/B test modification has two or more active variations (in addition to the comparison group).

To define the variation distribution for an A/B test modification:

  1. In the Basic settings view, click Define distribution.

  2. Set the distribution between the variations. If you increase the distribution for one variation, the distribution of other variations is automatically adjusted so that the total distribution is 100%.

    If the comparison group is enabled for the modification, 10% of the visitors are automatically included in the comparison group, so the total distribution between the other variations is 90%. You can deactivate the comparison group in the Basic settings view.

  3. When you're done, click Save.

After defining the variation distribution, you can view the distribution for each variation.

Defining the variation distribution for a multi-armed bandit modification

For a multi-armed bandit modification, you define how to measure variation performance ("reward") and which algorithm to use to automatically optimize the variation distribution. The Frosmo Platform continuously adjusts variation distribution for that modification based on how well the variations perform over time. Performance is measured based on the number of clicks or conversions attributed to each variation.

For more information about the rewards and algorithms, see Multi-armed bandit optimization.

To define the variation distribution for a multi-armed bandit modification:

  1. In the Basic settings view, click Define distribution.

    Defining the variation distribution for a multi-armed bandit modification

  2. Select the reward and the algorithm.

    Defining the variation distribution for a multi-armed bandit modification

  3. When you're done, click Save.

You can view the distribution for each variation when the variations, and the modification, are active. After data starts to accumulate, the distribution changes over time as the platform measures which variation performs the best. The calculation is performed every two hours.

Defining the variation distribution for a multi-armed bandit modification

  • No labels