> For the complete documentation index, see [llms.txt](https://docs.elevateab.com/elevate-helpcenter/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.elevateab.com/elevate-helpcenter/split-url-testing/preview-and-qa-a-split-url-experiment.md).

# Preview and QA a Split URL Experiment

Pre-launch checklist for split URL experiments in Elevate A/B Testing, covering redirect behaviour, URL matching, and tracking across pages.

***

Work through this in **Preview**, before launching. A split URL experiment moves visitors between pages, so a misconfigured redirect sends real traffic to the wrong place.

### Redirect check

Use a fresh incognito window for each check. A normal window may already hold an assignment.

| Check                         | What passes                                                                                                                                               |
| ----------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Control redirects             | Landing on the control URL sends you to a variation                                                                                                       |
| The redirect is clean         | You are not left looking at the control page before it moves                                                                                              |
| Behavior matches your setting | If you chose **Control URL only**, going directly to a variation URL does not redirect. If you chose **Redirect once**, a second visit does not redirect. |
| Assignment is sticky          | Returning lands you on the same page as before                                                                                                            |
| Every variation is reachable  | Check each variation, not only the first one you happened to get                                                                                          |

### URL matching check

| Check                           | What passes                                                                                                                            |
| ------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------- |
| The control matches as intended | The URL you expect to trigger the experiment does trigger it                                                                           |
| Parameters behave               | Visit the control URL with a campaign parameter appended and confirm the experiment still runs                                         |
| Nothing over matches            | If you used **Contains**, **Starts with**, or **Regex**, visit pages that should be outside the experiment and confirm nothing happens |
| Multiple control URLs work      | If the control holds several URLs, test each one                                                                                       |

Over matching is the expensive failure here. A pattern broader than you intended will redirect visitors from pages you never meant to include, and it will look like the experiment is working.

### Page check

| Check                   | What passes                                                              |
| ----------------------- | ------------------------------------------------------------------------ |
| Both pages are finished | Control and every variation load correctly and completely                |
| Mobile and desktop      | Every page in the experiment renders correctly at both widths            |
| Apps and widgets        | Reviews, chat, upsells, and subscription selectors work on every page    |
| The full path works     | Add to cart and start checkout from the control and from every variation |

### Paid traffic check

If ads or emails point at any URL in this experiment, confirm the destination still behaves as your campaigns expect after the redirect, and that any conversion tracking on those pages still fires.

### Tracking check

| Check                         | What passes                                                        |
| ----------------------------- | ------------------------------------------------------------------ |
| Events arrive from every page | Visiting each URL in the experiment produces an event              |
| Correct variation recorded    | The event records the page you actually landed on                  |
| Add to cart is recorded       | Adding a product from any page in the experiment produces an event |

Experiment data is cached and refreshes every two hours. Select the refresh control next to **Experiment Settings** to pull live data.

### Next steps

* [Troubleshooting Split URL Experiments](/elevate-helpcenter/split-url-testing/troubleshooting-split-url-experiments-and-faqs.md)