E-commerce: CNP interchange,
fraud tools, and your
true blended rate

A founder running the same storefront in person and online will usually notice their blended effective rate creeping up as web sales grow, and assume their processor did something. It didn't. The card networks price a keyed-in or clicked-through sale differently from a dipped or tapped one — a mechanic we broke down in full in card-not-present interchange — and on top of that floor, e-commerce carries its own layer of chargeback exposure and fraud-tooling spend that in-store retail simply doesn't. Put both together and you get the real number: not the rate on your agreement, but the true blended cost of accepting cards online.

Layer one: the CNP interchange floor

Every card-not-present sale — a web checkout, a phone order, a card typed into a virtual terminal — is priced into a higher interchange category than the same card would earn if it were dipped or tapped in person, because the issuer has no cryptographic proof the physical card was present. Published 2026 Visa interchange schedules put standard card-present consumer credit around 1.5–1.8%, while card-not-present consumer credit categories on the same schedules commonly run from roughly 1.89% up to 2.60%, plus a per-transaction fee on both. That gap is not a MidPay decision, a competitor's markup, or anything negotiable — it sits underneath every acquirer's pricing identically, the same way MidPay's own locked benchmark figures (1.49% debit / 2.69% credit) sit on top of whatever floor the network sets for a given transaction type.

This floor alone explains why a merchant who adds a webstore to an existing retail location sees their blended effective rate rise even though nothing in their agreement changed — a mechanic covered from the pricing side in why your effective rate rises. The CNP floor is layer one. It is real, it is published, and it applies before anything else in this article does.

Layer two: chargebacks are structurally worse online

Card-not-present transactions carry a materially higher chargeback rate than card-present ones. Industry chargeback trackers compiled from network and processor data put typical CNP dispute rates in the 0.6–1.0% range of transactions, versus roughly 0.5% for card-present retail — and fraud is the dominant driver, with the large majority of disputes tracing back to some form of fraud rather than simple buyer's remorse. The mechanism is the same one that sets CNP interchange higher in the first place: no physical card, no chip cryptogram, no proof of possession, which means a stolen card number is exactly as usable on a checkout form as a legitimate one until something else — AVS, CVV, 3-D Secure, device fingerprinting — catches it.

Every chargeback an e-commerce merchant absorbs costs more than the disputed sale itself: the merchandise (often already shipped and unrecoverable), the dispute fee charged by the acquirer, and the staff time to research and respond, on top of the risk that a rising ratio trips a card brand monitoring program — the escalation path we walked through in chargebacks as a hidden tax. None of that appears as a percentage on a rate sheet. It appears months later as an unexplained drag on net revenue.

The rate sheet prices the card. It does not price the fraud, and fraud is where most of an e-commerce merchant's real cost is hiding.

Layer three: fraud-prevention tooling is a real, ongoing line item

Because CNP fraud is structurally easier to attempt than card-present fraud, most e-commerce merchants of any real volume run some combination of AVS/CVV checks, 3-D Secure (Verified by Visa / Mastercard Identity Check), device fingerprinting, velocity rules, and manual review queues — each carrying its own cost, whether a per-transaction fee from a fraud-scoring vendor, a gateway surcharge for 3-D Secure calls, or the payroll cost of a human reviewing flagged orders. None of these tools change the underlying interchange percentage. What they change is how much of the chargeback and fraud-loss cost in layer two actually lands on the merchant instead of getting caught before it ships.

3-D Secure is the clearest example of the tradeoff: when authentication completes successfully, liability for fraud on that transaction shifts from the merchant to the card-issuing bank — a real financial protection — but it can also add friction at checkout that costs a percentage of conversions, which is its own cost even when no fraud ever occurs. Some CNP interchange programs also require a qualifying authentication signal just to reach the standard CNP rate rather than falling to a costlier non-qualified tier, a mechanic covered in full in the downgrade problem. For recurring-billing sellers, keeping card-on-file data current through a card updater service closes a different leak: an expired or reissued card doesn't just fail quietly, it can trigger decline and retry fees stacked on top of everything else in this article.

Building your own true blended-rate number

1. Start with your CNP interchange floor. Pull an itemized interchange-plus statement and separate card-present from card-not-present volume — a blended flat-rate statement hides this entirely.
2. Add your actual chargeback cost, not just the disputed dollar amount: dispute fees, unrecoverable merchandise, and staff time, tracked over a rolling 90 days.
3. Add your fraud-tooling spend — 3-D Secure fees, fraud-scoring vendor costs, and any manual review labor — as a real monthly line, not a rounding error.
4. Compare the total against your quoted rate. The gap between what you were quoted and what you actually paid, once all three layers are added, is your true blended rate — and it is almost always higher for e-commerce than for the identical business selling in person.
5. Re-run this quarterly as your online share of volume changes — the CNP floor doesn't move, but your fraud-tooling and chargeback costs will, especially around seasonal volume spikes.

Frequently asked questions

Why is my e-commerce effective rate so much higher than my in-store rate?

Because a card typed into a checkout form is priced as card-not-present interchange, which runs meaningfully above card-present interchange on the same network schedules, and because e-commerce carries a higher chargeback rate than card-present retail, so more of your volume is quietly funding dispute costs and fraud-prevention tooling on top of the interchange floor. Your quoted rate never changed — the channel mix behind your blended number did.

What actually makes up an e-commerce merchant's true blended rate?

Three layers stack on top of each other: the card-not-present interchange floor set by Visa and Mastercard (roughly 1.9-2.6% on consumer credit, plus a per-transaction fee), the network assessment and processor markup layered on top of that floor, and then the indirect costs of chargebacks and fraud-prevention tooling that a card-present retailer mostly doesn't carry. The published rate on your agreement is only the second layer — the first and third are what most merchants never see broken out.

How much does card-not-present fraud actually cost e-commerce merchants?

LexisNexis Risk Solutions' 2026 True Cost of Fraud study found retail and e-commerce merchants in the US now spend an average of $5.13 in total cost for every $1 lost directly to fraud — the first time that multiplier has crossed $5, up from $4.61 in the prior year's study. That multiplier includes the fraud loss itself plus investigation labor, remediation, fees, and lost merchandise, not just the chargeback amount.

Do fraud-prevention tools like 3-D Secure and AVS actually lower the blended rate?

They lower the total cost, not the base interchange rate. 3-D Secure shifts fraud liability to the card issuer when it completes successfully, and passing complete AVS and CVV data helps a transaction qualify for the standard card-not-present tier instead of downgrading to a costlier non-qualified one. Neither changes the published interchange percentage, but both reduce the chargeback and fraud-tooling costs that sit on top of it — which is where most of an e-commerce merchant's controllable cost actually lives.