E-commerce processing costs: where online sellers actually overpay

Selling online starts with a structural handicap: card-not-present transactions carry higher interchange than in-person ones, because the fraud risk is higher. That part is real and nobody can waive it. But most online sellers pay well above even that honest premium — in the stacking, not the rate. Here is where the money hides.

The five stacking points

• The fixed piece on small orders. At 2.9% + 30¢, a $18 order pays 4.57% effective before anything else is added. Low average order value is the single biggest driver of online processing cost — and the reason per-transaction fees deserve more scrutiny than the percentage.
• International and currency fees. Cross-border cards add percentage fees (Stripe publishes +1.5% for international cards, plus conversion), so a store with a global customer base blends far above its domestic headline without anything being "wrong."
• Platform + gateway + processor, three layers deep. A store platform's own transaction fee (where it applies), a gateway fee, and the processor's rate can all touch the same sale. Sellers routinely pay one layer they could switch off simply by using the platform's built-in payments — or pay a gateway twice, once standalone and once bundled.
• Fraud tooling priced per screen. Radar-style tools charge per screened transaction or per order. Necessary for many stores — but it belongs in your all-in cost math, and its settings (block thresholds) quietly decide how many good orders you decline.
• Refunds that keep the fee. Under major processors' published policies, the original processing fee is generally not returned on a refund. A store with a 10% return rate pays processing on a tenth of its volume twice over — once on the sale, once in the kept fee.

The number that settles it

Add every processing-related deduction for a complete month — processor fees, gateway, platform transaction fees, fraud screening — and divide by gross processed volume. That effective rate is the only figure that can be honestly compared across setups. Online sellers who compute it for the first time commonly land between 3.4% and 4.5%; where you sit inside that range is mostly ticket size and customer geography, and the rest is stack.

Then compare models, not brands: for card-not-present volume, interchange-plus passes through the (higher) online interchange and adds a stated markup — the comparison against your flat rate is arithmetic your own statement can settle. Run your two numbers in the calculator, pre-set for online sellers.