Create a new Holdout

Creates a new holdout experiment to establish a control group for measuring the impact of your feature flags and experiments.

A holdout reserves a percentage of your traffic (specified in basis points) that will be excluded from seeing any experiments, allowing you to measure the overall impact of your experimentation program against a baseline control group.

The holdout will be created in a draft status and can be configured with:

• Traffic allocation: Percentage of users to exclude from experiments (0-10000 basis points)
• Scope: Global (affects all experiments) or local (specific experiments only)
• Audiences: Target specific user segments using audience conditions
• Metrics: Track and analyze the impact on key business metrics
• Environment: Specify which environment the holdout applies to

Key Validation Rules:

• Holdout key must be unique within the project
• Required fields: key, name
• traffic_allocation must be between 0-10000 (basis points)
• key must match pattern: ^[a-zA-Z0-9_\-]+$ (max 64 characters)
• name and description have character limits (255 and 1000 respectively)

Metrics Integration: If metrics are provided, they will be automatically saved to the MetricsHub for tracking and analysis.