managing-experiment-lifecycle

Managing the experiment lifecycle involves guiding teams through transitions like launch, pause, resume, end, ship, and archive. This framework specifies preconditions, variant handling, and decision criteria to reduce missteps during analytics workflows.

You should pause, resume, or end an experiment based on decision criteria for running experiments. The lifecycle guidance clarifies outcomes affecting variant assignment and analytics to help you select the correct action in common analytics workflows.

Pausing an experiment does affect variant assignment and analytics data. The lifecycle framework defines explicit preconditions and outcomes for each action, ensuring you understand the analysis implications before modifying a running experiment.

The best way to ship or archive an experiment after analysis is to apply the post-analysis cleanup decision criteria. This ensures lifecycle actions are executed with clear outcomes, streamlining experiment management across teams.

You can reset or duplicate an experiment within your analytics workflow. The lifecycle framework provides a reusable decision structure that specifies the preconditions and outcomes required to perform these actions safely.

When you end or archive an experiment, the lifecycle guidance specifies how variant handling and feature flag configurations are affected. This ensures decision criteria are met and analytics implications are clear before completing the action.