Skip to main content
You can define URL rules and visitor rules inline on every experiment. For a first test, inline is fine. As soon as you have more than one test that targets the same set of pages or the same audience, saving them starts paying off.

​
Where they live

  • Site → Locations, reusable URL rule sets.
  • Site → Audiences, reusable visitor rule sets.
Locations list

Saved Locations. Attach the same list to many experiments.

Each is a table with a name, a rule count, and the experiments currently attached.

​
Creating one

Under the tab, New. Give it a name that reads clearly on an experiment’s summary, “Product detail pages”, “US mobile visitors”. and add rules with the same builder used inside the experiment editor. Rules combine with AND. To combine with OR, create two saved lists and attach both to the experiment; the experiment matches if either matches.

​
Attaching one

Inside an experiment, under Where it runs or Who sees it, use the picker instead of writing rules inline. The picker lists saved Locations or Audiences on this site. An experiment can attach more than one saved list. It can also mix saved and inline rules, attach the saved list, then add an extra inline rule for something specific to this test.

​
Editing a saved list

Under Site → Locations (or Audiences), click a row to open the editor. Changing the rules changes them for every attached experiment on the next config publish. Because the change affects many experiments at once, editing a saved list requires the Edit locations (or Edit audiences) permission. See permissions catalog.

​
Deleting

Deletion is refused (409 in_use) while any experiment references the list. Detach from the experiments first, then delete.

​
Naming

A saved list’s name shows up on the experiment’s summary line, so make it explicit. “Product detail” is fine; “location-1” is not.

​
When not to save

  • One-off test. If you know this is a single-shot test on a URL you will never target again, inline rules are less overhead.
  • You need per-experiment tweaks. If every test on the “pricing” page needs slightly different rules, one includes the trial page, one excludes it, an inline rule per experiment is clearer than a saved list plus an override.

​
The two-way ownership

A saved list is a reference, not a snapshot. If you edit the list, every attached experiment picks up the change on the next publish. Sometimes you specifically want the opposite, you want the experiment to freeze the rules at the moment it was set up. Two options:
  • Detach the saved list and copy the rules inline on the specific experiment. That freezes them.
  • Clone the saved list, edit the copy, and attach the copy to the new experiment. That gives you a divergent branch without touching the original.
Either is fine. Pick the one your team finds easier to read six months later.