Standard
A coupon is only as good as the check behind it.
This page is the working standard, not marketing. It explains exactly what we record, how long we trust it, and the situations where the honest answer is “we don't know”.
What a check actually is
Copying a code proves nothing. Visiting a store's promo page proves nothing either. A check means taking a real basket through a real checkout far enough to see the discount appear, then recording the numbers below.
- Start from a documented offer: an official page, an email, a permitted feed, or a shopper report.
- Build a basket that the offer's terms allow, and note the subtotal before tax.
- Enter the code at the cart or checkout stage where the store accepts it.
- Record whether the discount appeared, how much came off, and what changed in the total.
- Store the result — including failures — with the timestamp and the tester's identity.
What every ledger entry contains
| Field | Why it exists |
|---|---|
| Offer and store | Which recorded offer was tested, and at which merchant. |
| Exact time | Stored in UTC, displayed with the timezone so a check cannot be mistaken for a claim about now. |
| Who tested | A named human tester or an approved automated process. Never an anonymous tick. |
| Source | The URL or document the offer came from, plus its type: official page, email, feed, or shopper report. |
| Market and channel | US market, USD, and whether the basket was run on desktop, mobile or the store's app. |
| Account eligibility | Whether the tester was a new or returning customer, because that changes results. |
| Basket | The products or category, quantities and subtotal the code was applied to. |
| Money | Discount applied, shipping if known, and the pre-tax total after the discount. |
| Outcome | Passed, failed, inconclusive or not tested — with a reason code when it did not pass. |
| Stacking | Whether a sale price was already active, and whether combining them was actually tested. |
Labels, and what they never mean
- Checkout tested — Applied successfully in a documented basket.
- Merchant confirmed — Published by the merchant. Not yet applied at checkout by us.
- Shopper reported — Reported by a shopper. Useful, but limited assurance.
- Not yet tested — Listed, but no check is recorded. Treat as a lead, not a result.
- Needs recheck — Evidence is stale or conflicting. We re-test before recommending it.
- Ended — No longer active. Kept for the record, not shown as a live offer.
A timestamp, a feed import, a page visit, an affiliate record or a code copy is never a successful checkout test. We do not display a single “success rate” without the sample, the window and the failures that produced it, so we display no such score at all yet.
How long evidence stays useful
These are operating defaults, not guarantees. Prominent offers on high-traffic store pages are rechecked every 24 hours; lower-volume offers every 72 hours. When a store publishes an expiration date, that date wins — an offer does not become valid forever because our recheck is late. Anything past its due date is listed as needs recheck rather than quietly left in place.
Failures, conflicts and eligibility
One failed checkout does not prove a code is dead for everyone. A code that fails for a returning customer may be intended for first orders; a code that fails on sale items may work at full price. So a failure records the basket and the account type that produced it, and the offer moves to needs recheck instead of being deleted. Two conflicting results trigger a review rather than a coin toss.
Automation limits
Automated checks are only used where a merchant permits them and a scenario can be reproduced without bypassing access controls. We do not defeat CAPTCHAs, place real orders, or use private customer accounts. When a store cannot be checked automatically, it goes to the manual queue — which is slower, and we say so.
Reporting a result
Shopper reports are recorded as shopper reports, kept separate from staff tests, and moderated. We strip names, emails, order numbers, addresses and payment details from anything we publish. If a report changes a label, the reason is written down.
Send a report through the corrections and reports form. The full editorial standard is in our editorial policy.