The offer that sounds too good to refuse

Somewhere in your Stripe dashboard, or in a sales email, you've seen the promise: pay a small percentage on top of your normal processing, and Stripe will cover eligible fraudulent chargebacks. No evidence to gather. No deadline to chase. If a disputed charge qualifies, Stripe reimburses the amount and waives the dispute fee.

For a merchant who has just lost a week to a single chargeback, that sounds like buying your Sunday back. And sometimes it is. But "protection" is doing a lot of quiet work in that sentence, and the word covers less than most people assume when they switch it on. Before you trade a percentage of every sale for peace of mind, it's worth understanding exactly what you're buying — and what you're still on the hook for.

What the fee actually covers

Stripe's Chargeback Protection (part of Radar) is, mechanically, an insurance product. You pay an additional fee on each transaction — a flat percentage layered on top of your regular per-charge cost. In exchange, when a covered dispute lands, Stripe reimburses the disputed amount and eats the dispute fee, so a loss that would have cost you the sale plus the penalty costs you nothing beyond the premium you'd already paid.

The key word is covered. Protection applies to disputes filed under fraud reason codes — the "I didn't authorize this" family, where a cardholder claims they never made the purchase. That's a meaningful category. True unauthorized-use fraud is often the hardest kind to win on your own, because the cardholder's bank starts from the assumption that the customer is telling the truth, and your evidence has to overcome that presumption.

So if your losses cluster in genuine card fraud, protection is aimed squarely at your problem. The trouble is that most merchants' disputes don't cluster there.

The disputes it quietly leaves on your desk

Walk through the reasons customers actually dispute charges, and you'll notice how many fall outside the fraud bucket:

Product not received. The customer says the thing never arrived. That's a delivery dispute, not a fraud claim — and it isn't covered.

Not as described or defective. The customer got the product but says it wasn't what they were promised. Also a service dispute. Also not covered.

Subscription and recurring-billing complaints. "I thought I cancelled," "I forgot I signed up," "I didn't expect to be charged again." These are among the most common disputes for any business with a renewal, and they live outside the fraud category.

Friendly fraud dressed as fraud. Here's the subtle one. A customer who genuinely made a purchase but files under an unauthorized-use code may fall under protection — but the category a dispute is filed under is chosen by the cardholder's bank, not by the truth of the situation. You don't get to decide how your losses are labeled.

The result is a coverage gap that's easy to miss until you're standing in it. You can pay the premium on every single transaction and still find your actual disputes — the not-received, the not-as-described, the subscription regret — sitting in your dashboard uncovered, waiting for a response, with the clock running.

The trade you're actually making

Protection also asks something in return that's easy to overlook: on covered disputes, you generally stop fighting. Stripe absorbs the loss, and you waive the back-and-forth. That's the whole point — you're paying to not participate.

For a fraud dispute you'd likely lose anyway, giving up a fight you were going to lose is a good trade. But it changes your relationship to your own data. When you fight disputes yourself, you learn things: which products attract complaints, which customers charge back, where your billing descriptor confuses people, whether your delivery proof holds up. Handing that off wholesale can smooth your losses and dull your instincts at the same time. The premium buys calm; it doesn't buy the feedback loop.

The math that actually decides it

Strip away the marketing and the question becomes a straightforward expected-value calculation. You're paying a fixed percentage of all revenue to offset a variable cost that only hits some of it.

Start with three numbers you can pull from your own account:

Your fraud-dispute rate. Not your total chargeback rate — the share that comes specifically from unauthorized-use fraud codes, the kind protection covers. This is the only slice the premium is buying down.

Your average covered loss. The disputed amount plus the dispute fee, for those fraud cases.

The premium on your full volume. The protection percentage applied to everything you process, because you pay it on every transaction whether it's ever disputed or not.

If the premium across all your sales costs more than the fraud losses it would have absorbed, you're subsidizing an insurance pool — paying to protect transactions that were never at risk. If your genuine fraud losses are heavy and concentrated, the premium can easily pay for itself and then some.

This is the same logic behind any insurance decision. Insurance is rational when the loss is rare but catastrophic, or frequent enough that the premium beats the expected damage. It's a bad deal when you're insuring against something that rarely happens to you, or when the policy excludes the exact events you keep suffering. A business with a high true-fraud rate — digital goods, high-ticket items, card-not-present with weak verification — may find the trade clearly worth it. A business whose disputes are mostly subscription regret and "where's my order" is paying a fraud premium to solve a non-fraud problem.

The honest middle path

Here's the conclusion most merchants arrive at once they run the numbers: protection is a tool for one specific failure mode, not a blanket solution. It can genuinely take the worst fraud losses off your plate. What it cannot do is make the rest of your disputes disappear — and for a lot of businesses, the rest is where the real money leaks out.

That leaves you with a two-part reality. The fraud disputes you might offload with a premium. But the not-received, the not-as-described, the "I forgot I subscribed" — those still land in your inbox with a seven-day deadline, and the only thing that wins them is a clear, evidence-backed response filed on time. Protection doesn't write that response. Nothing about paying the premium changes the fact that an unanswered dispute is a lost dispute by default.

So the real question isn't "protection or nothing." It's whether you have a reliable way to handle the disputes protection was never going to cover — the majority, for most merchants. If those are the ones piling up, the premium is treating a symptom you don't have while the real one goes unaddressed.

Where this leaves you

That gap is exactly the one Argeback is built for. It watches for the disputes that land in your Stripe account, drafts an evidence-backed response tailored to the reason code — the not-received, the not-as-described, the subscription complaint — and files it before the deadline passes, from your phone. It doesn't ask you to pay a percentage of every sale to skip the fight; it makes the fight quick enough to actually win. If you're weighing chargeback protection and realizing it leaves most of your disputes uncovered, you can see how Argeback handles the rest at argeback.lumenlabs.works.