​
Experiment Creation

Set up an experiment in FlagSync to test feature flag variants against metrics. This process integrates Feature Flags, Events, and Metrics to evaluate performance and optimize your product with data.
1

Experiment Details

Define a name (e.g., “Button Text Test”) and hypothesis (e.g., “Changing ‘Buy Now’ to ‘Complete Purchase’ will boost conversions due to clarity”).
Experiment details form
A specific hypothesis aids result interpretation.
2

Select Metrics

Pick existing Metrics (e.g., conversion rates, totals) or create new ones, to measure success.
Experiment metrics form
3

Choose Feature Flag

Select or create a flag with variants to test (e.g., “Buy Now” vs. “Add to Cart”).
Experiment flag form
4

Choose Baseline

Pick the control variant (e.g., current “0% discount”) to compare against others.
Experiment baseline form
Typically the current production value, or best-performing variant.
5

Configure Rollout

Allocate percentages to each variant (e.g., 50% baseline, 50% new)—must total 100%.
Experiment rollout form
Changing traffic or variants starts a new experiment iteration for data integrity.
6

Activation

  • Save and start the experiment to collect data
  • Once active, FlagSync tracks impressions and events, aggregating metrics to compare variants.
Run for a fixed time (e.g., two weeks) or until significant statistical results are achieved.
FlagSync locks linked flags and metrics while running.

​
Next Steps

Get started with a guided flow in Quickstart: Overview, then: