A/B link testing distributes eligible visitors among alternative destinations. The redirect can report allocation and clicks, but a winner requires a destination-side outcome such as signup or purchase.
Plan before publishing
Write the hypothesis, audience and decision metric first. Keep the control as the baseline and change one meaningful factor so the result can explain what caused the difference.
- Define one hypothesis.
- Make weights total 100 percent.
- Instrument the destination outcome before launch.
Build and test the workflow
Configure at least two working destinations and weights totaling 100 percent. Use stable assignment when repeat visitors should see the same variant and keep campaign labels consistent.
Worked example
With a 50/50 split, 1,000 visitors may not produce exactly 500 per page. Test every destination, form and mobile layout, and check whether geo or device overrides bypass the experiment.
Measure the right outcome
DrutoLink reports clicks and estimated uniques by variant, not native conversions or revenue. Join the variant context to destination analytics and compare conversion rates only after enough eligible traffic.
Troubleshooting checklist
- No traffic: validate weights and link status.
- Visitors switch: inspect assignment mode and identifier limits.
- Early winner reverses: extend the test.
Product limits and responsible use
Do not run a redirect experiment when traffic is too low, regulated visitors require identical information, or a platform-native test can measure the complete funnel more reliably.