A/B tests in Wedge exist to answer a positioning question, not to squeeze out a colour-change lift.
Create variants#
Generate a second variant with a different angle — a different pain, a different customer, or a different promise. Traffic is split between variants automatically.
What's tracked#
- Views per variant
- Signups per variant
- Conversion rate per variant
Reading results#
Signup rate is the signal. A variant that wins on views but loses on signups lost. Treat small samples with suspicion: a handful of visitors won't separate two decent variants.
Feeding the score#
Per-variant conversion flows into your Validation Score. Strangers converting on a specific promise is strong evidence that the promise is right.