All docs
Landing pages

A/B testing & conversion

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.