Making changes to a running Experiment

You can update a running experiment at any time, including changes to variant code, targeting rules or tracking settings. However, these updates do not automatically apply to visitors until they are explicitly enforced. This ensures experiment consistency and prevents accidental mid-test changes from affecting data validity.

Editing variants

If you modify a variant through the Pertento Editor (for example updating HTML, CSS or JavaScript), the changes are saved but not immediately deployed to active users. These updates must be enforced for the new version to replace the currently active variant code.

Updating targeting or tracking

Changes to targeting rules, URL patterns, audience definitions or event reporting also require enforcement.

Without enforcement, the experiment will continue using the previous configuration to ensure result stability.

Enforcing changes

To activate any updates during an ongoing experiment:

1. 1Go to the experiment’s Details page
2. 2Locate the Enforce action
3. 3Click Enforce to push all pending changes live

After enforcement, all visitors matching the experiment’s targeting will receive the updated configuration going forward.

Why enforcement is required

Enforcement prevents unintentional mid-experiment shifts that could:

• invalidate results
• split users across inconsistent variants
• alter tracking logic mid-test
• compromise significance calculations

By explicitly confirming the update, you ensure that all users see the same, stable experience from that point onward.

Notes

• Enforcement affects only future visitors; it does not retroactively change historical data.
• Always enforce after making changes that should be live immediately.
• If you do not want a change to affect the current run, leave it unenforced until the next test.