Why coupon sites are full of dead codes

Open any big coupon page and you'll see the same thing: forty codes listed, a wall of green "verified" badges, and — if you actually try them — maybe three that work. It's tempting to call that incompetence. It isn't. It's incentives. The coupon-aggregator model is built to reward listing codes, not proving them, and once you see the money flow, the dead codes stop being a bug and start looking like the point.

The affiliate math rewards the click, not the discount

Here's the mechanic almost no coupon site will explain to you. When you click through one of their codes to a store and buy something, the site earns an affiliate commission — whether or not the code you clicked actually worked. The payout is attached to the click and the purchase, not to the saving. So the rational move, if truth isn't your constraint, is to list as many codes as possible. More codes, more clicks, more commission. A dead code that still sends you to the store still pays. That single fact explains most of what's wrong with the category.

Scrapers built a hall of mirrors

The codes themselves mostly aren't hand-found. They're scraped — one site copies another, which copied another, which copied a merchant email from 2023. A code that died months ago keeps getting re-listed because its reflection is still bouncing between sites that all cite each other as the source. Nobody in the chain ever put the code through a register. The "verified" badge just means the code was seen, not that it was tried. We call the codes that come out of this machine phantoms: strings that were never real, kept alive forever by pages that never check.

Vote counters lie, politely

The fix the industry landed on was community voting — thumbs up, thumbs down, a "92% success" number next to each code. It sounds honest and it mostly isn't. Votes lag reality by weeks; a code can rack up green votes in the morning and be dead by dinner when the merchant pulls it. Votes are easy to game, and worse, they average away the truth: "70% worked" is useless to you, the one person standing at the checkout, for whom the code either fires or it doesn't.

And then there was the extension that got caught

The most-installed shopping extension of its era faced a very public backlash in late 2024 when it came out that, at checkout, it could overwrite the affiliate credit for the sale onto itself — taking the commission from the creator who actually sent you there. The same class of tool has also been shown to surface worse codes than a quick search would find, because the incentive was never to save you the most money. When the business model rewards the click, the customer's actual saving becomes optional.

What we do instead

We started Clipper because we wanted the opposite constraint: be right, even when being right costs a click. So the rules are inverted. A code is only working if a live hunt found current evidence or a real person carried it through a real checkout. The final authority is a receipt — a human outcome at an actual register — and it outranks every badge and vote. Codes that fail don't vanish; they stay published in the graveyard, dead and labeled, because knowing a code is dead saves you the ninety seconds every other site is happy to waste. Every code shows when it was last checked, because a code proven in March is folklore by July. And when a code is real but fenced to the wrong cart, we mark it scoped instead of pretending it never existed.

The catch, said out loud

Being honest is worse for short-term clicks. A page that admits "these three are dead" gets fewer hopeful taps than one that shows forty green badges. We're betting the trade the other way — that a person who once watched us publish our own failures comes back, and a person burned by a wall of phantom badges doesn't. The whole moat is the receipts. Everything else is decoration.

See the method in the open — working, untried, scoped, and dead, timestamps and all — at clipper.deals/codes. ✂