Split URL Testing
Split URL Testing sends a defined portion of your traffic to a completely different URL, letting you compare two page experiences that live at different addresses. Unlike Theme A/B Testing — which keeps visitors on the same URL and swaps the theme — Split URL redirects the visitor's browser to the variant URL before the page renders.
Common use cases
Comparing a dedicated campaign landing page against your standard product page
Testing two different category page structures at separate URLs
Validating a redesigned collection page before making it the permanent URL
A/B testing personalised landing pages for different paid traffic sources
How it works
Visitor identification and assignment — Same as Theme A/B Testing: Cooee reads the visitor's ID from localStorage / cookie and deterministically assigns them to an arm. The assignment is sticky — returning visitors always go to the same URL.
Redirect — Visitors assigned to a variant arm are redirected to the variant URL using
window.location.replace(). This does not add an entry to the browser's history, so pressing the back button does not loop visitors back through the redirect.Control arm — Control visitors stay on the original URL with no change.
Anti-flicker — The same anti-flicker protection used in Theme A/B Testing applies here: the page is briefly hidden until the redirect decision is made, then revealed.
Assignment recorded — The same server-side recording runs in the background, enabling per-arm analytics in your Cooee dashboard.
Before you start
The variant URL must be live, fully functional, and Cooee-enabled before you publish the experiment — visitors will land on it immediately
Both the control and variant URLs must be on the same Shopify store domain
Ensure Cooee's theme extension is also active on any custom pages at the variant URL
Setting up a Split URL Test
Step 1 — Create the experiment
In Cooee, go to Experiments → New Experiment
Select Split URL Testing
Enter a Name and Hypothesis
Choose a Primary Metric
Click Create
Step 2 — Configure the variant URL
Click on Variant A
Enter the full Variant URL — must start with
https://and be on the same domain as your storeRename the arm if helpful (e.g. "Campaign LP v2")
Adjust the traffic split as needed
Step 3 — Publish
Click Publish. Cooee starts redirecting variant-assigned visitors immediately.
Reading your results
The same set of per-arm metrics are available as in Theme A/B Testing:
Metric | What it measures |
Impressions | Unique visitors who landed on the arm's URL in a real session |
Conversion Rate | % who completed a purchase |
Revenue per Visitor | Total revenue ÷ total visitors |
Add to Cart Rate | % who added to cart |
Bounce Rate | % who left without engaging |
Session Duration | Average time on site |
Important considerations
SEO impact
Search engines may discover and index the variant URL. If the experiment runs for more than a few weeks, consider adding a rel="canonical" tag on the variant page pointing back to the control URL. This signals to Google that the control URL is the preferred version and prevents keyword cannibalisation.
Session continuity
Once a visitor is redirected to the variant URL, all their subsequent sessions will originate at the variant URL (because their assignment is sticky). Analytics tools like Google Analytics will attribute these sessions to the variant URL's path. Account for this when reviewing external analytics data alongside Cooee data.
Variant URL must stay live
If the variant URL becomes unavailable during the experiment (e.g. the page is deleted or the URL changes), variant-arm visitors will land on a broken page. Monitor the variant URL's availability throughout the experiment.
Minimum runtime
As with Theme A/B Testing, run for at least 2 full business cycles (typically 14 days) before evaluating results.