---
description: "Run one test across several countries, languages or domains: one targeting group, one variant with copy per market, results by country, and deployment only where it won."
---

# One Test Across Markets

::: info The short answer
One test can run across all your markets. Put every domain or locale path in one targeting group, and let one variant fill in the right copy per market. You reach a result faster than with a separate test per country, then see results by country and deploy only where the change won. Run separate tests when the markets have different pages or offers.
:::

## Why one test

A store selling in five countries splits its traffic five ways. Five separate tests each get a fifth of the visitors and take far longer to finish. One test across all five gets the full traffic, and the country breakdown still shows where it worked.

## 1. One targeting group for every market

| Your site uses | Set up the group with |
| --- | --- |
| **Locale paths**, such as `/fi-FI/`, `/sv-SE/`, `/de/` | One **Regex** rule that matches every locale (example below) |
| **Separate domains**, such as `example.com` and `example.ca` | A ruleset per domain. Click **Add a ruleset on another domain** for each one. A page matches if any ruleset does. |

For five locales, a rule covering every product page:

```
^/(fi-FI|sv-SE|de|en|pl-PL)/products/
```

Check the reach panel to confirm every market is counted. See [Targeting Groups](/experiments/targeting).

## 2. One variant, copy per market

Describe the change once and give the agent the copy for every market in the same prompt. For example, for a Nordic outdoor-clothing brand with five locales:

> Add a highlights block under the product title with three benefits. Pick the language from the URL path: `/fi-FI/` Finnish, `/sv-SE/` Swedish, `/de/` German, `/pl-PL/` Polish, `/en/` English. Here is the copy for each: ...

The agent writes the code that reads the path (or the domain) and fills in the matching copy. Preview the variant on a page of each market before launching.

The agent writes in the language of the page and answers in the language of your prompt.

## 3. Results by country

Open the results and choose **By country**. It re-runs the comparison per visitor country, for your top countries, with the rest under **Other**.

This is available on newer experiments. On older ones, the tab says results by country are available for new experiments.

To find a market's tests on the Experiments page, use the **Market** filter. It only filters the list and doesn't change who sees a test.

## 4. Deploy only where it won

When the test finishes, deploy to the countries where it won and leave the others on the original. See [Deploy to a segment](/personalise/by-audience#deploy-to-a-segment).

## When separate tests are better

| Situation | Why |
| --- | --- |
| The pages differ per market (layout, products, offers) | One change can't mean the same thing everywhere |
| You need a separate verdict per market | One test gives one verdict. The country breakdown is a slice, not its own test. |
| One market has most of the traffic | Start there, learn what works, then roll out to the others |

## Limits

- **Results split by visitor country, not by domain.** A visitor from Germany on your English site counts as Germany.
- **Brand guidelines can't be set per market yet.** [Brand Intelligence](/editor/brand-intelligence) applies to the whole account.
- **Country isn't a test audience yet.** To limit a test to one market, use the domain or locale path in the targeting group.