Our Editorial Policy and Content Standards

This page sets the rules our content must meet before publication, names the people accountable, and lists what we refuse to publish. It applies to coupon listings, store pages, category pages, and guides, without exception. It is written so you can hold us to it: every commitment below is either something you can check on a page or something you can report when we get it wrong.

What this policy covers, and who it binds

The policy covers four things: the offers we list, the descriptions and terms attached to them, the guides we write, and the people whose names appear on all of it.

It binds everyone who publishes here, including contributors and anyone using automation tools. Where a commercial relationship conflicts with this policy, this policy prevails. Where this policy and a reader's experience disagree, the reader is usually right, and we correct the page.

Who is responsible for this site

CouponFlint is published by CouponFlint, based in the United States. Editorial decisions are the responsibility of Vengatesh Kamatchi, who can be reached at vengatesh@couponflint.com. Our about page covers the site's background and how it operates.

We publish this because a reader deserves to know who is behind a page before deciding whether to trust it. A site that will not say who runs it has told you something.

How we source and verify coupon offers

Every offer is tested before it is published, and the test date is shown with the offer. Offers come from three places: retailer affiliate programs, the retailer's own public channels, and reader submissions. None of the three is published on trust alone, including offers supplied directly by a retailer's affiliate program, because network entries stay live after promotions end.

The full process, including what a passing test does not prove, is documented in how we test and verify codes. Everything listed /stores follows it.

How we source and fact-check written content

Guides and store descriptions are built on named sources, and the source is linked at the claim rather than collected in a footer. We work in a tier system.

Three rules follow from that table. Any claim about a retailer's policy links to that retailer's own page. Any statistical links should point to the organization that produced them, not to whoever repeated them. Any claim we cannot trace to a source gets cut rather than hedged into something vague.

Time-sensitive facts carry the date they were checked. Retail policies, sale dates, and cash-back rates change without notice, and an undated claim about them is a claim with a hidden expiry.

Standards for numbers, claims, and comparisons

Numbers are the easiest thing to invent in this industry and the hardest for a reader to check, so we hold them to a stricter standard than prose.

We do not publish a discount percentage, price, savings total, user count, or accuracy rate without a source and a date. We do not publish success rate percentages next to codes, because a percentage implies a measured redemption rate across real checkouts, which is not data a coupon site holds. Where a number would be guesswork, we describe the mechanism instead and omit the magnitude.

This standard is not just editorial preference. Under the FTC's advertising substantiation policy, an advertiser must have a reasonable basis for an objective claim before publishing it. Where a claim implies a particular level of proof, the advertiser actually has to hold that level of proof. Site statistics, verified badges, and "most trusted" style superlatives are all objective claims by that standard.

Superlatives need a stated basis. "The best cash-back site" is not something we will publish unless we can say what measure, how it was tested, and on what date.

Independence from commercial relationships

We earn commissions from retailers when readers click through and make a purchase. That funds the site, and the affiliate disclosure sets out the mechanics.

What the commission cannot do:

• It cannot change a test result. A code was either applied to the cart, or it wasn't.
• It cannot buy a placement, a ranking position, or a favorable description. If a retailer ever pays for placement, that placement carries a visible label identifying it as paid, and it does not appear inside editorial content.
• It cannot remove criticism. If a retailer's return policy is restrictive or its codes routinely exclude the products people actually want, we say so on the page that earns us money.
• It cannot stop a correction. A commercial relationship is never a reason to leave an error standing.

What commission does affect, honestly: which retailers we cover at all. Coverage follows the programs available to us, so an absence from the site is not a judgment about a retailer.

How we use AI and automation

We use automation in production, and we say so rather than leaving readers to guess. Google's guidance on helpful, reliable, people-first content asks publishers to be clear about the "Who, How, and Why" behind their content, including where automation played a role. This section is our answer to the "How."

Our rules for it:

• A person reviews and is accountable for everything published, and nothing goes live without review.
• Automation is not allowed to generate a fact, a statistic, a policy detail, or a source. Facts come from the sources described above and are checked against them.
• Automation never produces a claim of experience. A sentence saying a code was tested only exists when a person has tested it.
• Automation never writes a byline, a credential, or a biography.

Authors, bylines, and credentials

Every guide carries a byline that names a real person, links to a real author profile, and describes a real background. Store pages name the editor who maintains them.

We do not invent authors, borrow the name or photograph of a real person who has not agreed to it, or attach years-of-experience figures and job titles that nobody holds. This has become a common practice in this industry, and it is straightforward deception: a reader deciding whether to trust a page is entitled to know the byline belongs to someone real.

Dates, updates, and review cadence

Every page shows when it was last updated, and that date only moves when something substantive changes. Restamping a page to make it look fresh is a manipulation, and readers notice it.

Content that becomes outdated and cannot be brought up to date is removed rather than left up.

Corrections, and how to report an error

Send the page, the claim, and what you believe is correct through /contact, and include the store name and code if it is an offer that failed.

What happens then:

1. We check the claim against a primary source, typically the retailer's website.
2. If we were wrong, we fix the page and update the last updated date.
3. For a factual error in a guide, we add a dated correction note at the foot of the page saying what was wrong and when it changed. Silently editing an error away is not a correction; it is a cover-up.
4. If the report concerns a broken code, the offer is returned to the test queue and relabeled or removed based on the retest results.

We do not require you to explain who you are or why you care.

What we will not publish

• Single-use codes are issued to one shopper, then fail for everyone else afterward, costing the recipient their discount.
• Employee, student, military, or first responder discounts presented as open to everyone, when legitimate versions require verification through the retailer's own provider.
• Codes obtained by generating or guessing strings, or by any other method, would be treated as abuse.
• Offers whose terms we cannot find or read, since publishing them results in the failure being displayed on your checkout screen.
• Fabricated testing, invented reader stories, and manufactured quotes.
• Unverifiable site statistics, including totals for shoppers, helped, money saved, or codes listed.
• Content designed to rank rather than to answer, including pages that exist only to hold a keyword.
• Instructions for abusing a retailer's returns, pricing, or promotional systems.

Reader submissions via /submit-coupon are added to the same list.

How does this policy map to outside standards

Two external frameworks shape the sections above, and naming them lets you judge our policy against something other than our own opinion of it.

The Trust Project's Trust Indicators are a transparency standard used by news organizations, covering best practices, author expertise, type of work, citations and references, methods, and actionable feedback. The sections on responsibility, sourcing, bylines, and corrections map onto those indicators.

We are not a Trust Project member and do not claim certification from any organization. We use the framework because it is a public standard, not because we have been awarded anything.

The FTC's guidance on endorsements and advertising substantiation governs disclosure and numerical standards. Those are legal requirements for a commercial publisher in the US, not aspirations.

Frequently asked questions

Does a retailer pay you to be listed or ranked higher?

No. Retailers pay commissions on referred sales, which is a payment for outcome rather than placement. Any paid placement, if it ever runs, carries a visible label indicating it is paid and sits outside editorial content. Commission rates do not change how offers are tested, described, or ordered.

Do you use AI to write your content?

We use automation as a production tool, with a person reviewing and taking responsibility for anything published. Automation is not permitted to generate facts, statistics, sources, claims of testing, or author identities. Anything stating that a code was tested reflects a test that a person actually ran.

How do I report something that is wrong?

Send the page URL and the claim through the contact form, with the store name and code if an offer failed. Factual errors in guides get a dated correction note on the page. Broken codes are returned to the test queue and relabeled or removed based on the retest.

Why does the site not show how many shoppers it has helped?

Because we cannot substantiate those numbers, and an unverifiable statistic used to build trust is exactly the kind of claim this policy exists to prevent. Where a credibility claim is useful, we describe a process you can check, rather than a total you cannot.

Who writes the guides on this site?

Named people with real profiles, linked from every byline. Store pages also name the editor responsible for maintaining them. If a page carries a byline you cannot trace to a profile, that is a defect worth reporting.

Reaching the editorial team

Corrections, policy questions, and complaints about anything on this page can be sent /contact. If a commitment on this page is not being met anywhere on the site, please let us know, as it is the fastest way to get it fixed.