Time series endpointAnalyze how your chosen metrics evolve over time while the test runs.New in v3: GET /3/abtests/{id}/timeseries with date range and metric selection.
Create A/B testChoose which metrics the experiment should measure, and add statistical controls for evaluation.
v2: create a test by naming it, defining its variants, and setting an end date.
v3: in addition, you explicitly select the metrics to analyze (for example, conversion rate) and can configure statistical options such as minimum detectable effect (MDE), and multiple-testing correction.
Estimation endpointEstimate required sample sizes and duration to detect your target effect.
v2: basic estimator.
v3: uses your selected metrics and desired effect size/MDE for more tailored estimates.
Listing and retrievalSame operations with more flexible sorting/filtering and responses that include your selected metrics and evaluation settings.
Multi-variant testing
v2: mixed setups such as A/A/B are rejected. You must either use the same index for all variants (for example, A/A/A) or use distinct indices for all variants (for example, A/B/C). Mixed shapes fail with an invalid configuration error.