> ## Knowledge Base Index
> Fetch the complete knowledge base index at: https://help.personizely.net/sitemap.xml
> Use this file to discover available pages before exploring further.
> Pure-Markdown content can be obtained by appending a '.md' suffix to the content URLs listed in the sitemap (without the trailing slash).

# How to QA an A/B Test before launching it?

Follow the steps below to QA your A/B test before launching it to your real visitors.

1. After you have finished drafting the A/B test, in the **campaign configuration**, go to the **Target** subsection and select **Specific target**.


![](https://storage.crisp.chat/users/helpdesk/website/-/3/0/a/0/30a06a2316be0a00/screenshot251_1yz0ibf.png =1000x565)

2. Click on **Create new target**.

In the new target, select the **Query string** filter, and set the query string to `testing`. Set the condition to **is not empty**.

Save the target, and apply it to the campaign.


![](https://storage.crisp.chat/users/helpdesk/website/-/3/0/a/0/30a06a2316be0a00/screenshot252_11t7ilb.png =804x800)

3. Next, change the **Variation split**. Uncheck **Even Split**, and change the **Variation** to **99%**, keeping **Original** at **1%**. It's important for Original to have 1% so the test runs properly on the website.


![](https://storage.crisp.chat/users/helpdesk/website/-/3/0/a/0/30a06a2316be0a00/screenshot253_1uc3dcc.png)

4. **Save** your changes and **launch** the test.

5. In incognito/private mode of your browser, access the A/B test URL with the following query string: `?testing=true`. 
    Example of a full URL: `https://example.com/?testing=true`

This will ensure you are part of the test target and trigger the A/B test. It will only run when using that query string, so it won't affect your real visitors.


**Additional notes:**

* Make sure to remove or adjust the **Specific target** and **Variation split** settings once QA is complete, so the test runs normally for all visitors. As well as reset the campaign stats.
* This method allows you to safely live test variations without affecting your real visitors.

Congratulations, you're now ready to confidently QA your A/B tests before going live!