Check whether a conversion-rate difference between two variants is statistically real, then bound and rank the results.
Run a two-proportion z-test on conversions and visitors for variant A vs. B.
ab_test({"conversions_a": 100, "visitors_a": 100, "conversions_b": 100, "visitors_b": 100})Get a confidence interval around the observed conversion rate.
confidence_interval({"mode": "mean", "mean": 100, "std_dev": 100})Rank the winning variant's conversion rate against historical runs.
percentile({"numbers": []})