#A/B tests

Go to A/B Tests → New A/B test, name it, and pick two to four scripts from your library. They become variants A, B, C and D.

#How variants are assigned

Each rep is always given the variant they have used least. Every rep spreads their calls evenly, so one star rep can't make a script look better than it is.

#How the winner is decided

Variants are compared on meetings booked per dial. Clarity shows:

• Needs more calls until every variant has at least 30 dials
• Leads, not significant yet when one variant is ahead but the gap could still be luck
• Wins (95% confidence) when a two-proportion z-test says the difference is real

When you end a test with a clear winner, Clarity records it. Roll the winner out to the team and test it against a new challenger.