Price tests for clients: legal, ad-feed and reputation checks

If a client sells into the EU and plans to announce Black Friday reductions, the safest course is often to finish price tests on EU-facing products before October 28, or to keep those products out of the test.

Price tests suit clients with steady sales on a few products, known product costs and a team that can answer a customer who asks why a friend paid less. They don't suit a client in the middle of a rebrand, a client with a history of price complaints, or a store whose best sellers are also its biggest ad-feed products and nobody owns the feed.

If the real question is "how deep should the Black Friday discount be", an offer test answers it with less exposure. Our article on whether to test prices or discounts first sets out the difference.

Most price-test complaints come from the price shown next to the tested price, not the tested price itself.

United States. The FTC's guide on former price comparisons says a former price is a legitimate basis for comparison when the article was "offered to the public on a regular basis for a reasonably substantial period of time" (16 CFR 233.1, read October 12, 2026). A strike-through invented for the test is the kind of comparison the guide warns about.

European Union. Article 6a of the Price Indication Directive, added by Directive (EU) 2019/2161, requires any announcement of a price reduction to show the prior price, meaning the lowest price the trader applied during at least the 30 days before the reduction (EUR-Lex consolidated text, read October 12, 2026). Member states enforce it, and some set their own details.

Two practical consequences for a client price test:

• Show a strike-through in an arm only when it is a price the store really charged. For most tests, the safest choice is the store's existing strike-through or none at all.
• Ask counsel whether a tested price counts as a "price applied". If a share of EU shoppers paid $48 in late October, it is an open question whether $48 becomes the reference for a November reduction. We don't know of guidance that settles this for A/B tests, so treat it as a question for the client's lawyer, and keep EU-facing products out of price tests close to a sale if the answer is unclear.

Google says the price in your product data has to match the price on the landing page, and products with a mismatch may be disapproved; it also notes that prices changed by JavaScript after the page loads can trigger the error (Google Merchant Center Help, read October 12, 2026).

A price test changes the shown price for shoppers in a test arm. Before launch, decide with the client and whoever runs their Shopping campaigns:

• Which tested products are in the feed, and how much of their traffic comes from Shopping ads.
• Whether to leave feed products out, or to keep Shopping traffic out of the test with traffic-source targeting.
• What happens when the store price is raised for the test. If you test a higher price by letting the tool raise the Shopify price, the feed, the Shop app and point of sale see the raised price for everyone outside the test while it runs.
• Who watches Merchant Center diagnostics during the test and has authority to pause it.

Shoppers forgive a price they don't like more easily than a price that changes under them. Check that:

• A shopper's price follows them from collection page to product page to cart and checkout, and stays the same when they come back on the same browser.
• Checkout charges what the page showed. A test that changes the number on the page while checkout charges the old price measures a display, not a price, and it is the fastest way to a complaint.
• Support has a script. "We sometimes test prices; you paid the price shown to you, and we're happy to help" is better than an agent guessing. Agree with the client whether support may refund the difference, and record that decision.
• Repeat buyers and subscribers are thought through. A subscriber who sees a lower first-order price in one arm and a normal renewal later needs to have been told on the product page.

A price test changes what the client's customers pay, so the client signs, not just the marketing contact. The sign-off should name:

1. 1.The products and variants, the arm prices and any strike-through shown.
2. 2.The markets and currencies in the test, and those left out.
3. 3.Whether the Shopify price is raised during the test, and what it returns to.
4. 4.The start date, the end date (no later than November 15) and who can stop it early.
5. 5.The primary goal: gross profit per visitor, or revenue per visitor if product costs aren't complete.
6. 6.The support script and refund policy for price questions.
7. 7.That the client's counsel has reviewed the reference-price and feed questions above.

Store it with the test, and link it from the client report.

When the test ends, every shopper should see and pay the store's normal price again, and any raised Shopify price should be back where it was. Check the product pages, the cart and the Merchant Center feed the next morning. Don't carry a "winning" price into the sale without the client's decision: a result from late October describes late-October shoppers, and BFCM shoppers behave differently. The guide to reading a result covers what an interval does and doesn't tell you.

• A strike-through invented for the test. Show only reference prices the store really charged.
• Forgetting the EU look-back. A tested price shortly before a sale can complicate the sale's reduction claims.
• Ignoring the feed. Shopping traffic landing on a different price is a policy problem and a trust problem.
• Sign-off from the wrong person. Price changes need the client's decision-maker, in writing.
• Running into the freeze. End by November 15 and confirm prices are back.

• [ ] The client is a good fit for a price test; an offer test was considered
• [ ] Strike-through prices are genuine or switched off
• [ ] EU exposure and the 30-day look-back reviewed by the client's counsel
• [ ] Feed products identified; Shopping traffic handled; someone watching Merchant Center
• [ ] Price consistency checked from collection page to checkout
• [ ] Support script and refund policy agreed
• [ ] Written sign-off stored with the test
• [ ] Test ends by November 15; prices and feed checked the next day

Running a price test for one brand rather than a client list? The brand and small-store tracks start from the BFCM hub. The general dates are in the guide to testing before Black Friday, and the term is defined in our glossary entry on price tests.