# Managing lifecycle - Docs

After you've analyzed your experiment metrics and determined a winning variant, you can manage the experiment's lifecycle by ending, pausing, or resetting it.

## Ending an experiment

After you've determined a winning variant, click the **End experiment** button on the experiment page. This lets you choose a variant to keep, roll it out to all users, and end the experiment.

Beyond this, we recommend:

1.  Sharing the results with your team.

2.  Documenting conclusions and findings in the description field of your experiment. This helps preserve historical context for future team members.

3.  Removing the experiment and losing variant's code.

4.  Archiving the experiment.

5.  Disabling the feature flag once the winning variant is hardcoded in your application and the flag check has been removed from your code. Every active flag counts toward your [feature flag billing](/docs/feature-flags/cutting-costs.md), even if it's rolled out to 100% and no longer doing anything useful.

Remember, experimentation is an iterative process. Each experiment teaches you something about your users, even when results aren't what you expected.

### Ending an already rolled out experiment

If you've already shipped a variant to 100% of users, you can still end the experiment using the **End experiment** button. In this case, the experiment is marked as complete without changing the feature flag.

## Pausing an experiment

You can temporarily pause an experiment to stop collecting data without ending it completely. This is useful when:

-   You need to fix a bug in one of the variants
-   External factors might temporarily affect results (e.g., a major marketing campaign)
-   You want to review interim results before continuing

To pause an experiment, click the **Pause experiment** button on the experiment page. This will disable the feature flag behind the experiment, so no new data will be collected.

When you're ready to continue, click **Resume experiment** to start collecting data again. The experiment will continue from where it left off.

## Resetting analysis

Resetting analysis clears the experiment's results and returning it to a draft state. This doesn't affect the feature flag or variant rollout, so variants stay visible to users. It does also not delete any collected data.

Technically speaking it resets the experiment's start and end dates. All events collected thus far will still exist, but won't be applied to the experiment unless you manually change the start date after launching the experiment again.

This is useful when you need to restart data collection, for example after fixing a bug in your experiment setup or changing metric definitions.

## Comparing pausing, ending, and resetting

Here's how the actions differ:

| Status | Feature flag | Users see | Exposure tracking | Results | Can resume |
| --- | --- | --- | --- | --- | --- |
| Running | Enabled | Multiple variants | Continues | Updating | N/A |
| Pausing | Disabled | Control variant | Stops | Fixed | Yes |
| Ending | Modified to roll out chosen variant to all users | Chosen variant | Continues (chosen variant only) | Fixed | No |
| Resetting | Unchanged | Multiple variants | Continues | Cleared | Yes (relaunch) |

### Community questions

Ask a question

### Was this page useful?

HelpfulCould be better