URL redirect tests

Split traffic between two page URLs and measure which one converts and earns more.

What a URL redirect test is

A URL redirect test compares two full page URLs. Visitors who land on the control URL are split into two groups: the control group stays on that page, and the variant group is redirected to a second URL you specify. Both pages are real, published pages on your store — you build them however you like, then let AB Test decide which one each visitor sees.

Because the comparison is between two independent URLs, the two pages can differ as much as you want: a different layout, a different offer, a rebuilt landing page, or a different product page entirely.

When to use it

Reach for a URL redirect test when you have (or can build) two distinct pages and want to know which one performs better. Common cases:

Comparing two landing pages for the same campaign or ad.
Testing a redesigned product, collection, or home page against the current one.
Trying a different long-form vs. short-form sales page.
Sending traffic to a bundle page vs. a single-product page.

If you want to change one theme template rather than send visitors to a separate URL, use a template test or a theme test instead.

How it works

The control URL is the trigger path. When a visitor lands on that path, the storefront app embed assigns them to a variant and keeps that assignment stable for the rest of their visit:

Control group — stays on the control URL. Nothing changes.
Variant group — is redirected at runtime to the variant URL you configured.

The redirect happens in the browser as the page loads, before the visitor interacts with it. Assignment is sticky, so a returning visitor keeps seeing the same side of the test and your results stay clean.

Requests that already carry a ?view= suffix are left alone — the redirect does not fire for them, so preview and template-scoped URLs are unaffected.

Set up a URL test

In the AB Test admin, create a new experiment and choose URL redirect as the test type.
Enter the control URL — the page whose path triggers the test. This is where your existing traffic already lands.
Enter the variant URL — the page redirected visitors should see instead.
Optionally set audience targeting rules so the test runs only for the visitors you choose.
Optionally set an end date. Experiments auto-end at their scheduled end date.
Launch the experiment.

A URL test needs no code and no theme changes. The app never creates or edits theme files — it only redirects assigned visitors from one published page to another.

Requirements and notes

Both arms need a URL. The control URL is the path that triggers the test; the variant URL is where the variant group is sent.
Both pages must be real, reachable pages on your store. Make sure the variant page is published before you launch.
The control URL should be a page that already receives traffic — otherwise the test has no visitors to split.
Requests with a ?view= suffix are excluded from the redirect, so they never interfere with the test.
Keep the storefront app embed enabled so assignment and redirects can run. See the Quickstart for enabling the embed.

Point the variant URL at a different page than the control. If both arms resolve to the same path, the redirect has nothing to compare.

Targeting and results

Combine a URL test with audience targeting to run it only for specific visitors — by device, new vs. returning, country, UTM parameter, referrer, or cookie. This is useful when a variant page is built for one segment, such as a campaign landing page for paid traffic.

Once the test is running, open Results to see Bayesian statistics: each variant's probability to be best, its conversion rate and revenue, and segment breakdowns. Revenue is credited to the assigned variant even through Shop Pay and third-party checkout — see revenue attribution. When a winner is clear, apply it in one click.