Skip to content
Working Coupons

How we test promo codes | Working Coupons

What a test looks like

We open the store on the storefront for the country being tested, in that country's currency, and add one ordinary, in-stock item to the basket as a guest. If the basket or the code box is only available after signing in, we record that and stop. We do not create accounts.

Before any real code, we enter a made-up control code. It must be rejected. If a store accepts the control code, it is accepting anything, so that session tells us nothing: we record no code as working and log the session as invalid.

Then we enter each listed code one at a time, removing the previous one. If a code fails only because of a minimum order, we raise the quantity once and try again, and we note the minimum. A minimum-order or sign-in block is not counted as a failure; we only mark a code as failed when the store clearly rejects it.

We read which code the discount line actually names, because some stores apply a code of their own when yours does not work. For every code we take a cropped, dated screenshot of the coupon line and the order summary: subtotal, discount and total. Then we empty the basket.

What we never do

We never pay. We never sign in. We never enter personal data. A test stops at the basket, before checkout asks for anything about you. That keeps the test the same for every store and every country, and it means a result describes what a guest shopper would see when they type the code in.

Stores that block automated browsers

Some stores block automated browsers with security checks or blank carts. We never try to get past a human check. Those stores are tested by hand on a phone, following the same steps: guest basket, control code first, then each code, then a dated screenshot.

How results reach you

Each test becomes a row in our public ledger at /tested/, and the same data is published as CSV and JSON under CC BY 4.0, so anyone can check or reuse it. On a store page, the result shows up as a label: "Works · tested {date}" or "Failed our test · {date}". Codes we have not tried yet say "Awaiting our test".

Labels expire. A tested result stays current for 21 days; our own network codes are re-confirmed nightly; a code we have only seen posted on other sites is shown for 30 days. Once an offer ends, it reads "Ran until {date}".

Store facts, such as delivery or return terms, are read on that country's storefront, in its currency, and shown with the source link and the date we read them.

Working rate and reader votes

A store's working rate is the number of codes that worked divided by the number tested in the last 30 days, per store and per country. We show it only after at least three tests, so one result cannot swing it.

Readers can tell us whether a code worked. Votes are limited to one per visitor per day per code, stored as a one-way hash rather than an IP address, and shown only after three votes. A code a reader sends us stays unverified until we test it ourselves.