What Are Metrics?
Metrics show how feature rollouts affect users—like clicks, sign-ups, or purchases—by analyzing events and impressions.They’re essential for pinpointing optimal flag settings and driving data-backed experiments.
Defining a Metric
Metrics start with Events you’ve tracked (e.g.,signup-cta-click). Here’s what defines them:
Follow along in the Dashboard, click ”+” next to Metrics or go to Create a Metric.
Measurement Type
Choose how to analyze data—each type powers experiments differently:Conversion Rate
Measures the percentage of users acting after seeing a flag (events ÷ impressions).
Numeric
Aggregates event values per user—e.g., counts, sums, or averages.
Aggregation Method (Numeric Only)
For numeric metrics, pick how to process values:- Count: Number of events per user (e.g., purchases).
- Sum: Total value per user (e.g., spend).
- Average: Average value per event (e.g., price).
- See Numeric Metrics for details.
Metric Impact
Set the goal—does success mean more or less? This guides how results are interpreted in Experiments:Increase
Higher values are better—e.g., more
signup-complete events for registrations.Decrease
Lower values are better—e.g., fewer
error-occurred events for stability.Next Steps
- Explore types in Conversion Rate or Numeric.
- Track events in Quickstart: Send Events.
- Define a metric in Quickstart: Define Metrics.
- Test with Experiments.