Experiment Overrides

Learn how to override group allocation for experiments during development and testing.

Override group allocation for an experiment

During development, explicitly stating which experiment group a user falls into can be useful. For example, if a developer is testing that each experiment condition works as expected, they might want to force themselves into each condition.

You can add overrides based on Feature Gates and Segments to create rules that force all users who pass a given Feature Gate or Segment into a given group. When fetching parameters for an experiment, if the user matches any override, Statsig returns the overridden result immediately.

Note:

For experiments that are in a layer, set the overrides at the layer level. Statsig can only override a user into one experiment within that layer (since layers always make experiments within them mutually exclusive).
Add overrides before the experiment starts, or at least before a user's first exposure. Statsig excludes users whose first exposures are controlled by overrides from Pulse results, because these users aren't randomized. Adding a large number of users to the override can affect the reliability of your experimental results.
- Warning: if you add overrides after a user has already been exposed in an experiment, Statsig doesn't exclude that user from Pulse results. Statsig continues to attribute that user's events and metrics to the group they were first assigned to. However, Statsig applies and honors the override after you define it. This can dilute or pollute your experiment results, because Statsig attributes the user's actions to one group while they were actually exposed to a different group. The overall impact depends on how many users fall into this category.
Statsig evaluates ID overrides first, then conditional overrides (from top to bottom).
ID type used for ID Overrides may or may not match the experiment’s ID type.

Adding an override

Log into the Statsig console at https://console.statsig.com
On the left-hand navigation panel, select Experiments
Select the experiment where you want to add an override, and navigate to the Setup tab
Click Manage Override, configure the override, and select Save

Deleting an override

If you add an override but later decide it isn't needed, you can remove it so Statsig evaluates the rules as normal.

Log into the Statsig console at https://console.statsig.com
On the left-hand navigation panel, select Experiments
Select the experiment or layer from which you want to delete the override
Click Manage Overrides
Click on the trash can icon next to the override you want to delete, and select Save

Testing an override

After adding your override, test it in the "Check Group for a User" window by navigating to the Diagnostics tab and adding the properties required for a user to pass the gate or segment set as an override.

Was this helpful?