Ecommerce A/B testing ideas: 30 Shopify tests by page

The best A/B test ideas for a Shopify store answer a question shoppers are already asking: is this the right product, what will it really cost, when will it arrive, and can I trust this store? Start on the product page and the cart, where those questions get answered or abandoned. Judge each test on revenue per visitor, and only run the tests your traffic can actually detect. The 30 ideas below are sorted by page, and each one lists the metric that decides it, the traffic it needs, and whether it needs Shopify Plus.

Most "test idea" lists stop at the idea. That is the easy part. The hard part is knowing which ideas your store can test at all, which number should decide the winner, and which ideas will quietly waste a month of traffic. This guide covers all three.

Two stores selling in the same market still do not have the same visitors. Positioning, colours, and photos change how people perceive the brand, and how they act. Treat the 30 ideas below as inspiration, not a to-do list. Do not apply them all at once and hope. Test each one on its own. Some will lift conversions, and some will tank them. Keep the winners. Every idea here is a hypothesis, and an A/B test is how you learn whether that hypothesis is likely true.

Before you pick an idea: three filters

Run every idea through these three filters before it goes on your roadmap. They remove more bad tests than any prioritization score.

1. Can your traffic detect it?

A test can only find a lift that is large compared with the natural noise in your data. Smaller stores need bigger, bolder changes. As a worked example, using a standard sample size formula at 95% confidence and 80% power:

Two things follow. First, a store with 30,000 monthly sessions can realistically detect a big change in conversion rate, not a small one. Second, metrics with a higher baseline, like add to cart rate, need far fewer visitors, which is why they are useful early signals. Revenue per visitor usually needs more traffic than conversion rate because order values vary so much. Run your own numbers in the sample size calculator before you commit.

A practical rule of thumb: a product-page test needs roughly 40,000 monthly visitors to that page to reach significance in a reasonable time. Set the sample size before you launch, run the test for at least two weeks so you see different kinds of visitors, and call the result at 95% significance, the same bar as the table above. Do not peek and stop early.

2. Which number decides it?

Pick one primary metric before launch. For anything that touches price, offers, or product choice, use revenue per visitor, because conversion rate can rise while order value falls. Add one or two guardrails that must not get worse, such as refund rate, discount use, or page speed.

3. Can Shopify run it on your plan?

Everything in your theme (product pages, collections, cart, homepage, navigation) can be tested on any plan. Checkout is different. Apps that customize the information, shipping, and payment steps of checkout are available only on Shopify Plus [Shopify]. If you are not on Plus, spend your testing budget before the checkout button.

In the tables below, "Traffic" is a rough guide:

• Any: a bold change that can show up at modest traffic, especially on add to cart rate.
• Medium: needs tens of thousands of visitors per variant on the page.
• High: a subtle change. Only worth testing at high traffic, or bundle it into a larger test.

Product page test ideas

The product page is where most purchase decisions are made, and almost all of it lives in your theme. Start here.

What to skip: button color and font tweaks. They are cheap to build but rarely move revenue enough for a store to detect. If you want to change them, change them outside a test. Do not spend a test on a tweak your traffic cannot detect.

Where the ideas come from: your own evidence beats any list, including this one. Search your support tickets and reviews for the questions shoppers ask before buying. The CRO audit framework walks through how to turn those into test hypotheses.

Collection page test ideas

Collections decide which products a shopper even considers. Changes here are usually measured on click-through to product pages first, then on revenue.

Cart and cart drawer test ideas

The cart is where cost surprises surface. Shoppers who get this far want to buy, so small reassurance changes can matter.

Homepage and navigation test ideas

The homepage matters most for branded and returning traffic. Measure it on how quickly visitors reach a product, then on revenue.

Offer and pricing test ideas

Offer tests can move revenue the most, and can also lose the most. Always judge them on revenue per visitor, and check margin before you ship a winner, because revenue is not profit.

A warning on price tests: showing two different prices for the same product at the same time can confuse returning shoppers and create support tickets. Testing how an offer is framed, or testing a shipping threshold, is usually safer than testing the base price. If you do test price, use a tool built for it; the Shopify A/B testing tools guide covers which ones support price and shipping tests.

Checkout test ideas (Shopify Plus)

These need Shopify Plus, because they customize the information, shipping, or payment steps. On other plans, test the cart instead.

Run one test, then the next

You cannot run all 30, and you should not write them into a long list and grind through it in order. A list written upfront is a pile of guesses. The result of one test should decide the next hypothesis.

The loop is short:

1. Write one hypothesis. Look at your Shopify funnel: sessions, add to cart, reached checkout, and orders. The biggest drop is where to start. Match one idea to that leak. Low add to cart points at product pages (ideas 1 to 8). A big drop from cart to checkout points at the cart (14 to 18). Keep only an idea your traffic can detect, and prefer a bold one marked "Any" if the page is quiet. "Because [evidence], we believe [change] will increase [metric] for [audience]." If you cannot fill in the evidence, it is a guess, not a test. The A/B testing best practices guide covers how to write one.
2. Launch that change alone. One isolated variable. Set the sample size before you start, run for at least two weeks, and call the result once, at 95% significance. Do not peek.
3. Learn, then form the next hypothesis. If it wins, keep it. If it loses or is flat, remove the variant and write down what you learned. The next idea should come from that, not from the next row on the list.

A worked example

Say your store gets 60,000 monthly sessions, converts at 2.5%, and your support inbox is full of "when will this arrive?" emails. Your funnel shows a healthy add to cart rate but a big drop between cart and checkout.

• The leak: cart to checkout.
• The evidence: delivery questions in support tickets.
• The test: idea 15, showing the estimated delivery date and shipping cost in the cart.
• The metric: checkout start rate as an early read, revenue per visitor to decide.
• Feasibility: with roughly 30,000 visitors per variant each month, this store can detect a large lift within about a month, but not a small one. The change is visible and specific, so it is a reasonable bet.

If it wins, keep it, and let that result suggest the next hypothesis, for example moving the delivery promise onto the product page (idea 1). If it loses or is flat, take the variant down, write down what shoppers ignored, and form a different hypothesis. Do not start the next numbered idea just because it was next on the list.

Expect most tests to lose

This is normal, not a failure. Researchers who have run experimentation programs at companies like Microsoft report that only a minority of tested ideas improve the metric they were designed to improve [Kohavi et al., 2020]. A losing test still tells you something: what shoppers do not care about. That saves you from shipping changes that would have cost money.

Be wary of lists that promise lifts of 50% or 200% from a single change. Those numbers usually come from one store, one test, and a low baseline. Your result depends on your store, your shoppers, and your traffic.

Test idea checklist

Before any test goes live, check:

• The idea is one hypothesis, drawn from evidence (analytics, support tickets, reviews, or session recordings), and it is the only change in the test
• One primary metric is chosen, usually revenue per visitor
• Guardrail metrics are listed
• The sample size is set before launch, and the test will run for at least two weeks
• The result will be called once, at 95% significance, without peeking
• The variant works on mobile, with discount codes, and in every market you sell in
• No major promotion or launch is scheduled during the test
• A losing or flat variant will be removed, and the learning will be written down

For the mechanics of running the test itself, from splitting traffic to reading the result, see how to run Shopify A/B tests that prove revenue.

If you would rather not run it yourself

Tools such as Intelligems, Shoplift, and Visually deserve credit for getting stores to test at all. Using one yourself is still a time sink: someone on your team has to build the variant, check it on every device, and read the result. Retainer agencies are paid every month whether or not revenue moves. They are often slow to build, and they tend to work through a fixed list of ideas.

Cascayd runs the same loop, and it is hands-off for the founder. Test builds are AI-assisted, with a human in the loop, and each test is proposed in Slack for your approval before it goes live. Cascayd is paid only from a share of measured uplift, with no monthly retainer. There is no promise that a given idea will win. Some hypotheses will lift conversions, and some will tank them. The test is how you find out which, and you keep the winners.

Book a call and tell us about your store.