Pack's A/B testing calculates a "Chance to Win" metric for each variant — the Bayesian probability that a variant outperforms the control. For revenue-based goals, this tells you the likelihood that a variant generates more revenue per user than the control.
The calculation happens in two stages:
- BigQuery aggregates raw event data (total revenue, visitor counts, etc.) per variant
- Bayesian Engine computes the probability that each variant beats the control