Action Group Distributions¶
Distributions control how the Tag Manager selects between multiple action groups within a rule. They enable A/B testing and traffic splitting.
Distribution Types¶
Dxtra Tag Manager supports three distribution types:
| Type | Behavior | Use Case |
|---|---|---|
| None | All action groups execute | Default -- no traffic splitting needed |
| Session | One group selected per session, consistent for the visitor's entire session | A/B testing where consistent experience matters |
| Page Load | One group selected randomly on each page load | Rotating content, impression-based testing |
None (Default)¶
When no distribution is set, all action groups in the rule execute. This is the standard behavior for most rules.
Session Distribution¶
Session distribution assigns each visitor session to one action group and keeps that assignment consistent throughout the session. This ensures a visitor sees the same variant across all pages during their visit.
When to use: A/B testing where you need consistent behavior per visitor (e.g., testing different tracking configurations or different banner content).
Page Load Distribution¶
Page load distribution randomly selects an action group on each page view. The selection can change from page to page.
When to use: Rotating content or scenarios where per-page randomization is acceptable (e.g., rotating promotional banners).
Setting Up a Distribution¶
- Open a rule with multiple action groups
- In the rule settings, find the Distribution option
- Select the distribution type (None, Session, or Page Load)
- Assign percentage weights to each action group
Weight Configuration¶
Weights determine how traffic is split between groups. For example, with two action groups:
| Group | Weight | Traffic Share |
|---|---|---|
| Group A (Original) | 50% | Half of sessions/page loads |
| Group B (Variant) | 50% | Half of sessions/page loads |
You can use uneven splits for staged rollouts:
| Group | Weight | Traffic Share |
|---|---|---|
| Group A (Current) | 90% | Most traffic |
| Group B (New) | 10% | Small test audience |
Verifying Distributions¶
Use the debug overlay (#dxtra-debug) to see which action group was selected for the current page load or session. The debug panel shows:
- Which distribution type is active
- Which group was selected
- The assigned weights
Best Practices¶
- Start with small test percentages (e.g., 10/90) before scaling up
- Use session distribution for A/B tests to avoid inconsistent visitor experiences
- Use page load distribution only when per-page randomization is acceptable
- Monitor results in the analytics dashboard before committing to a variant
- Remove distributions after testing concludes -- switch the winning variant to 100%
Continue to Analytics to learn about Dxtra's built-in privacy-first analytics.