Add Variants to a Web Experience

You can add variants to a Web experience and define the percentage of site traffic to participate in each variant within the WHAT settings.

Automated Personalization experiences have a limit of 10 variants per experience. All other experience types have a limit of 20 variants per experience.

To add a variant to an experience, first place your mouse pointer on the plus sign (+) below the WHAT settings. Click ADD VARIANT when it appears, and then select the type of variant that you want to add.

  • For Automated Personalization experiences, Dynamic Testing experiences, and Standard Test experiences, you can add a blank variant or a variant with actions from another variant in the experience.
  • For 100% Experiences, you can add a blank variant, a variant with actions from another variant in the experience, or a control.

Animated demonstration of a user placing their mouse pointer on the plus sign, clicking the ADD VARIANT button, and then viewing the variant options available

By default each variant is assigned a letter: Variant A, Variant B, Variant C, etc. To append the name of a variant to better identify it, first click it in the WHAT settings, then click the pencil icon next to its name in the action configuration panel, type the name into the field, and then click the green checkmark to save the change.

Animated demonstration of a user clicking a variant and then renaming it

To configure an action for a blank variant or to modify the existing action from another variant within the experience, click that variant in the WHAT settings and then either click ADD ACTION or click the existing action name to edit it.

Animated demonstration of a user adding an action to a blank variant

Allocating Traffic

If you add variants to a 100% Experience or a Standard Test experience, you can adjust the traffic allocation to those variants by selecting an option from Traffic allocation. The default traffic allocation is even.

Each variant in an experience has its own control group. Therefore, the control group consists of separate control groups for each variant.

Callout of the 'Traffic allocation' selector

Traffic Allocation Considerations

You can adjust traffic allocation for three experience configurations.

A vs. Control

If you add only a control, then you can select Custom from Traffic allocation and then configure the allocation in the Edit Custom Traffic Allocation modal.

The Edit Custom Traffic Allocation modal with configuration fields for variant A and the control

A vs. B

If you add only a second variant, variant B, then Traffic allocation contains a number of uneven variant allocation options as well as the Custom option.

Animated demonstration of the options available in the Traffic allocation selector for an experience configured with only two variants

A/B vs. Control or A/B/n vs. Control

Once you add two or more variants with a control to an experience, Traffic allocation contains a different set of options. Instead of selecting the percentage of traffic for each variant, you select how much of the experience's traffic goes to the control.

Callout of the options in the 'Traffic allocation' selector when the experience has a variant A, a variant B, and a control

The control group consists of separate control groups for each variant. As a result, the control group traffic selection is evenly divided across variants. The traffic not allocated to the control is divided equally across all variants.

In the case of an A/B/n test without a control, the traffic distribution remains even and Traffic allocation is disabled. This guarantees that each variant receives enough traffic to achieve a statistically relevant result in the least amount of time.

When you create an experience with multiple variants, keep in mind that the size of those variants influences the sample size for the experiment and control groups. A small control group affects how long an experience takes to reach significance. For this reason, if an experience is not a test, you may not need to use a control group.

Variants with low traffic percentages offer little insight into the behavior of your site's visitors.

You should consider the following factors before adjusting the traffic percentages for variants in an experience:

  • How much traffic will this experience see? If the page(s) you want to test only receives a small amount of traffic, a 50% control group is much more likely to yield actionable data.
  • Is this experiment risky? If the experiment offers the potential for negative consequences, you may want to opt for a 60% or 70% control group.
  • Do you want insight into site visitor behavior, or do you want to optimize your site? If you want insight, a control group closer to 50% may be helpful. However, if you want to optimize site content for the majority of your site visitors, a smaller control group may be best.

Exporting and Importing Variant Actions

You can export an action used in a variant and import it to a new experience.

To export a variant action, first click the additional options menu (...), select Export Actions, and then select the variant you want to export.

Callout of the menu of additional options, with 'Export Actions' highlighted and its variants to export shown

The variant JSON downloads as a GZ file.

You can add the exported variant action and settings to a variant in a different experience. With the recipient experience open on the Experience Editor page, click the additional options menu (...), select Import Actions, and then select the variant into which you want to import the action.

Callout of the menu of additional options, with 'Import Actions' highlighted and its import recipient variants shown

Next, click Choose File in the Import Actions modal, and then navigate to and select the GZ file.

The Import Actions modal