Split URL Testing on Shopify: Two Pages, One Winner

Most new product pages on Shopify get judged by a calendar.

You launch the new page on the 1st. You compare the month to last month. Sales are up 11%, so the new page "won." Except the 1st was payday week, your Meta budget went up $2,000, and a creator posted your product on the 9th. You didn't test the page. You tested October against September. That's the exact mistake split URL testing Shopify pages is built to catch.

Split URL testing on Shopify fixes that. You keep the old page at its URL, put the new page at a second URL, and split visitors between the two at the same time, usually 50/50 with a redirect. Same week. Same ads. Same weather. The only thing that's different is the page. Then you pick the winner on revenue per visitor, conversion rate times average order value.

Here's the whole method in 5 steps:

1. Build the new page at its own URL.
2. Split visitors 50/50 with a split URL tool.
3. Use a 302 or JavaScript redirect and a canonical tag.
4. Set a stop date from the sample size math.
5. Promote the winner, then delete the test.

A before-and-after comparison is a rigged fight. A split URL test is a fair one. Same crowd, same week, two pages.

When should you use a split URL test instead of an A/B test?

When the new page is a rebuild, not a tweak.

A classic A/B test changes something on one page at one URL: a headline, an image, a button. A split URL test compares two whole pages that live at two different URLs. That's the right tool when you've changed the layout, the story, the offer and the photos all at once, because squeezing a full rebuild into an on-page editor gets slow and breaks easily.

I'll be blunt about why this matters more now. Two years ago, nobody split URL tested a product page because building a second page took a designer, a copywriter and three weeks. So founders tested button colors. AI changed that. With RevenueFlows AI, you drop in a product URL and the AI writes the copy, designs the images and builds the page in less than 15 minutes. The challenger page is cheap now. Proving it wins is the part you still have to do right.

Use a split URL test for:

• A full product page rebuild against the original
• A new offer page, like a bundle-first page, against the single-unit page
• A landing page for paid traffic against the standard product page, the landing page vs product page question settled with data
• A new theme or template against the current one

Use a regular A/B test for one element at a time. The step-by-step for that lives in how to A/B test your Shopify product page.

How do you set up a redirect test on Shopify?

There are two clean ways to give the new page its own URL.

Option 1: an alternate template. Shopify lets one product render through a second template, called with a ?view= parameter. Convert's docs on testing Shopify templates with split URL experiments walk through it: duplicate the product template as product.alternate, and the variation URL becomes the same product URL with ?view=alternate on the end. Convert says this template method works with all Shopify plans. Same product, same inventory, same checkout. That's why it's my default.

Option 2: a separate landing page. If the new page is a standalone landing page (built in a page builder or as a Shopify page), it already has its own URL. Point the test at both.

Then the split. A testing app sends each new visitor to one version and keeps them there. Shoplift's FAQ says it randomly assigns each new visitor on their first page load, 50/50 by default, and the same visitor sees the same version every time they return. That stickiness matters. A shopper who sees page A on Monday and page B on Thursday pollutes both numbers.

Two setup traps Convert flags that are worth checking on day one: exclude URLs that already carry the view= parameter so the redirect doesn't loop, and pass the original URL's parameters through to the variant so your UTM tags survive the redirect.

Will a split URL test hurt your Google rankings?

Not if you follow Google's own rules. Google publishes guidance on website testing, and it boils down to four things:

• No cloaking. Don't show Googlebot one set of URLs and humans another.
• Canonical on the variant. Add rel="canonical" on the test URL pointing to the original. Google prefers this over noindex, which it warns can have unexpected bad effects here.
• 302, not 301. A 302 tells search engines the redirect is temporary, so the original stays indexed. JavaScript redirects are fine too.
• End it. Run only as long as you need, then remove the test. Google warns that a test left running far longer than needed can look like an attempt to deceive search engines.

Quick check before launch: open the variant URL, view the page source, search for "canonical", and confirm it points to the original product URL.

Google doesn't punish testing. It punishes tests that never end and tests that lie to the crawler.

Which tools run split URL tests on Shopify in 2026?

First, the one you might remember. Google's free testing tool is gone. Google's own help page says it and its paid 360 version stopped being available on September 30, 2023. Any guide still telling you to set it up is out of date.

Here's what I checked this week:

Tool	What it compares	Plan or requirement	Worth knowing
Convert	Original URL vs ?view= template URL	Template tests on all Shopify plans	Regex URL matching, passes UTM parameters to the variant
Shoplift	Templates, themes and URL tests	Core plan and up, 14-day free trial	Sticky 50/50 assignment, anti-flicker, 4.9 rating from 134 reviews
Shopify Rollouts	Two versions of a theme or checkout change set	Grow plan or higher	Free and native, 50/50 by default, ends after 90 days unless you change it
Google's old free testing tool	Nothing	Discontinued	Shut down September 30, 2023

Shopify Rollouts deserves a note. It's free and built in, but it compares a change set (your theme, checkout, catalogs, discounts) against the current state. It doesn't point two arbitrary URLs at each other. For a standalone landing page against your product page, you need an app. For the full ranking on statistical rigor, see the best Shopify A/B testing apps.

How long should you run a split URL test?

Pick the stop date before you launch, from the math. Never from a good-looking Tuesday.

Evan Miller's rule of thumb in How Not to Run an A/B Test: visitors per version equals 16 times the variance, divided by the square of the difference you want to detect. For a conversion rate, the variance is the rate times one minus the rate.

Run the math on a store like this, a hypothetical: a ceramic cookware set with 15,000 monthly visitors to its product page. The old page converts at 1.8%. You want to know if the new page hits 2.3%.

16 times 0.018 times 0.982, divided by 0.005 squared, is about 11,313 visitors per version. That's 22,626 total. At 15,000 a month, about 3,462 a week, the test needs roughly 6.5 weeks.

Under 10,000 visitors a month? Read the low traffic testing playbook before you start. And don't peek and stop early. Miller shows that checking a no-difference test after every visitor and stopping the moment it looks significant can push the false positive rate to 26.1%.

How do you pick the winner?

Revenue per visitor. Never conversion rate alone.

Same hypothetical cookware store, 10,000 visitors to each page.

Old page: conversion rate 1.8%, average order value $92. That means revenue per visitor is $1.66. On 10,000 visitors, that's $16,560.

New page with a bundle-first offer: conversion rate 2.3%, average order value $118. That means revenue per visitor is $2.71. On the same 10,000 visitors, that's $27,140.

Old page $16,560

New page $27,140

Conversion rate x average order value x visitors. Old 1.8% x $92. New 2.3% x $118.

Watch what happens. Conversion rate rose 28%. Average order value rose 28%. Revenue per visitor rose 64%, because the two levers multiply. Now flip it: a new page that converts at 2.3% but drags order size down to $70 earns $1.61 per visitor. It wins the conversion report and loses you money. Full argument in revenue per visitor vs conversion rate.

What does a rebuilt page look like against the original?

Real client numbers, not typical results, and not a promise of what your store will do. On one bedding product, Cooling Bamboo Sheets, conversion rate went from 1.0% to 4.3% and average order value went from $125 to $254. Revenue per visitor went from $1.25 to $10.92. On 10,000 visitors, that's $109,200 instead of $12,500. You can see the full case study numbers.

That wasn't a headline tweak. It was a new page. And a gap that size is exactly what a split URL test exists to prove.

For the whole testing system around this, start with the ecommerce A/B testing guide.

What to do next

Open your analytics and write down one number: last month's visitors to the page you rebuilt. Divide the total sample size from the math by it. If the answer is under two months, set up the split URL test this week. If it's longer, make the new page bolder before you test it.

---

Book Your Profit Audit

A split URL test only pays off when the challenger page is a real rebuild, and that's where most stores stall. Book a free profit audit and we'll show you where your revenue per visitor is leaking, then show you how to rebuild a high-converting product sales page in less than 15 minutes.

Book Your Profit Audit →

Or go here to check it out → revenueflows.ai

P.S. Last month versus this month is a rigged fight. Two pages, one week, one split is a fair one. Only the fair fight tells you which page earns more.