Price Testing on Shopify: How to Test Prices Safely

Shopify price testing means showing randomly split groups of shoppers different prices for the same product at the same time, then comparing revenue or profit per visitor. You can do it safely with an app that keeps each shopper on one price and charges the price they saw at checkout. Stay clear of personalized pricing and fake "was" prices. And if a price test feels too risky, test shipping fees and offers first. They often move revenue just as much.

Below we cover what a price test is, the legal and fairness limits, safer alternatives, the tools that can run one, and real results from our database.

What a price test is

A price test is an A/B test where the variable is the price itself. Group A sees $40, group B sees $45. Everything else stays the same: the page, the photos, the copy.

Shopify has no built-in way to do this. Apps handle it in one of two ways. Some use Shopify's Cart Transform Functions or checkout scripts to change the price in the cart. Others swap in a hidden duplicate product at add to cart. Intelligems' docs describe both methods: on Shopify Plus it can adjust the price with a checkout script, and on other plans it swaps in duplicate products tagged "price_test".

The metric matters more than in most tests. A lower price almost always lifts conversion rate, so conversion rate alone will mislead you. Judge price tests on revenue per visitor, or better, profit per visitor.

Price tests also need volume. The test has to be big enough to catch a revenue difference, not just a conversion difference, and revenue data is noisier. Plan on running at least two full business cycles. Our guide to how long to run an A/B test covers the math.

A quick example

Say a $40 product converts at 3% and a $45 version converts at 2.8%. Per 1,000 visitors, $40 earns $1,200 and $45 earns $1,260. The higher price "loses" on conversion rate but wins on revenue, and wins by more on margin, since the extra $5 is close to pure profit.

Legal and fairness limits

We're not lawyers, and this isn't legal advice. But a few rules come up again and again when stores test prices, and they are worth knowing before you launch.

Random splits vs. personalized prices

A standard A/B test assigns prices at random. That is different from personalized pricing, where a price is set from data about the shopper. Regulators treat the second much more strictly.

• EU: since the Omnibus Directive, sellers must tell consumers when a price was personalized on the basis of automated decision-making.
• New York: the Algorithmic Pricing Disclosure Act took effect on November 10, 2025. Businesses that set prices with an algorithm using a consumer's personal data must show the notice "THIS PRICE WAS SET BY AN ALGORITHM USING YOUR PERSONAL DATA," with civil penalties of up to $1,000 per violation, according to Jones Day.

The safe default: split traffic at random, never by device, location, browsing history or customer profile, unless you have checked the disclosure rules for your markets.

Sale prices and "compare at" prices

If one variant shows a discount, the reference price has to be real. In the US, the FTC's Guides Against Deceptive Pricing say a former price used in a comparison must be one the product was genuinely offered at, not a made-up number. In the EU, Article 6a of the Price Indication Directive requires a price reduction to show the lowest price applied over the previous 30 days.

This matters for price tests because a "compare at" price that only exists in a test variant may not count as a genuine prior price.

The trust problem

Even a legal test can go wrong if shoppers notice it. In September 2000, Amazon ran random price tests on 68 DVDs. Customers compared notes, and Amazon apologized and refunded about 6,900 customers the difference between what they paid and the lowest test price.

To lower that risk:

• Keep each shopper on the same price across sessions and devices where your tool allows it.
• Make sure the price on the product page, cart and checkout always matches.
• Exclude logged-in repeat customers and email or SMS traffic, who are the most likely to compare prices.
• Decide in advance how support will answer "why did my friend pay less?" One option is to refund the difference, as Amazon did.
• Keep the gaps modest. A small gap is easier to explain than a large one.

Shipping-price and offer tests as alternatives

If changing the product price feels too risky, test what surrounds it. Shipping fees, free-shipping thresholds, bundles and discount framing change what the customer pays without two people seeing different sticker prices for the same item. Our database has more winners here than in pure price tests.

Shipping fees

Dave Diederen has run many shipping tests, and charging for shipping keeps winning. Dave Diederen tested shipping price levels and found a $9.49 fee beat the control by +43.81% revenue per visitor. In another store, flat $2, $5 and $7 fees were tested against free shipping, and the $5 fee added an estimated +$59,894/month in revenue.

Thresholds work too. Raising the free-shipping threshold to $165 produced +6.27% profit per visitor. More results are on our free-shipping thresholds page.