The audience tab of your experiment is where you go to define the targeting for the experiment. Here you will control which groups see which behaviors, under which conditions.
- Goto Rollout's dashboard, on the left sidebar, click on Experiments, at the very top of that list. This will bring you to the Experiments page.
- Here, you will see a list of your experiments. Click on the one whose audience you want to configure. (Or, Create a new one.)
- When you create a new experiment for editing, you'll see a single condition. This is the default condition, and it serves as the application's default behavior. You cannot delete it.
- If you click the Add New Condition link, you'll be able to add a new condition.
- Notice that when you do this, the new condition is framed as the if while the default condition becomes the else.
- In this new condition, you can set a version and a group to target, and then specify the value of the flag. Note that the flag value can be False, True, or Split.
- When selected Split, a percentage selector will appear and allow you to select the percentage of the group that should evaluate to true. Also, note that you can only enter the true percentage, with false updating automatically.
- In the target group selector, you can add multiple groups which will have an OR relationship between them.
- You can continue to add conditions in this fashion, with each subsequent addition resulting in an else if condition.
- You can also delete any condition with the garbage can icon to the right of the condition.
- In order to change the order at which the conditions are being evaluated, please note that each condition is draggable and can be moved before/after other conditions. To change a condition order, click and drag the three vertical dots next to an if/else statement.
- If your flag is a Multi-Variant Flag the audience tab would present all the flag variants (as they appear in the code)
- When you are satisfied with your changes, click the Update Targeting button to persist your changes.
- Alternatively, you can click the Reset link to go back to the state from before you started making changes. Note that Reset only appears when you've made unsaved modifications.
Scheduled Release allows a team to open a feature at a specific time.
In order to create a scheduled feature release:
- Go to the Audience Tab on an experiment
- At the Value drop-down menu (True, False, etc), you will also find a Scheduled value
- After clicking on the Scheduled value, you would be able to select the percentage of the True value Now and the percentage of the True value that should be applied at a specific date
- You can also configure the system to change the deployment percentage at specific times by adding more than one deployment date (see the + Add button)
- Keep in mind you can add to a scheduled release a target group or a flag dependency
Scheduled Release Time
The Scheduled Release time is always in UTC.