A/B testing
A/B testing lets you split traffic between 2-4 destination URLs with configurable weight percentages. Each variant has a name, target URL, and weight - and every click records which variant was served.
Requires a Team or Agency plan.
Setting it up
- Create or edit a link
- Scroll to the A/B Testing section
- Click Add variant to add destination URLs
- Give each variant a name, target URL, and weight (percentage)
- Weights must sum to 100
Example:
| Variant | URL | Weight |
|---|---|---|
| Control | https://yoursite.com/page-v1 | 60% |
| Variant A | https://yoursite.com/page-v2 | 40% |
In this example, 60% of visitors go to page-v1 and 40% go to page-v2.
Viewing results
Open the link’s edit modal to see variant performance. The stats table shows:
- Variant name
- Click count per variant
- Percentage of total clicks
Use this to determine which destination performs best, then update the link accordingly.
Limits
- 2-4 variants per link
- Weights must be integers from 1-99
- Weights must sum to exactly 100
- Variant names must be unique within a link
Interaction with routing rules
If a link has routing rules, routing rules take priority over A/B testing. A/B variants only apply when no routing rule matches the visitor.
Priority order: Rules engine > legacy geo/device/OS routing > A/B testing > default target.
Example: If you have a rules engine rule for US visitors AND A/B variants, US visitors always go to the rule’s target. Everyone else gets split across the A/B variants. The same applies to legacy geo/device routing rules.
Removing A/B testing
Remove all variant rows in the edit modal and save. The link reverts to its standard target URL.