configuring-experiment-analytics

To configure experiment analytics, you define exposure criteria using feature flags or custom events, select metric types like funnel or ratio, and establish a workflow to interpret statistical significance and determine impact.

Setting up funnel and ratio metrics requires specifying sequential steps for funnels and automatic exposure steps, then classifying each metric as primary or secondary to correctly measure the experiment's impact.

Handling test accounts and custom exposure events involves configuring explicit guardrails to exclude test data and defining custom exposure events to accurately count users entering the experiment.

You need data schema discovery during the interpretation workflow to identify available events, define metric types, and apply a decision framework for shipping or ending experiments based on statistical significance.

Yes, feature flags can define exposure criteria for mean and retention metrics by counting users who trigger the flag, ensuring reliable measurement across multiple variants within your experiment analytics.

Distinguishing primary versus secondary metrics is required to apply a structured decision framework, ensuring that shipping decisions rely on the main impact metric while secondary metrics provide contextual guardrails.