Preview & Launch

Once you have completed the configuration of your experiment and have set the desired allocation and filters for your experiment, you should verify the behavior of your variants before launching the experiment.

Preview

For native mobile experiments, you can use Preview Variants to apply any variant to your device and check that your variants, filters and events are setup correctly. If everything looks good, you’re ready to launch!

../_images/Preview-Variants.png

Launch

Next, you can schedule the launch and/or end dates for your experiment or choose to launch it immediately. By default, the experiment is set to launch immediately and will need to be stopped manually. Simply click Launch Experiment and confirm the targeting criteria. Users should start seeing your experiment within 15 minutes.

../_images/Launch-Now.png

Scheduling the experiment

In the launch section, you can schedule the launch and/or end date & time for your experiment. To schedule a launch date for the feature, choose the ‘Scheduled to launch on’ option and set the desired launch date/time. Similarly, you can choose the ‘Scheduled to end on’ option and set an end date/time. If both the launch and end dates are scheduled, and the schedule is < 7 days, you can choose to repeat the schedule weekly.

../_images/Launch-Schedule.png

Note that when you set a schedule for an experiment, the schedule is set and executed based on your local time. It does not take end-user time-zone into consideration.

To clear a schedule, you can click “Clear Schedule” at the bottom of the screen. You will be asked to review and confirm your changes before the schedule is removed. The state of the experiment (draft, running, etc.) will remain unchanged when you clear the schedule.

../_images/Experiment-Clear.png

To stop the experiment, click “Stop Experiment” and users will stop seeing your feature within about 15 minutes, or the next time they are online.

../_images/Experiment-Stop.png

See our results page on viewing and interpreting results.