Payment Processing Fees Explained

One price, three different companies

A merchant sees a single deduction from a sale and reasonably calls it "the processing fee." That single number is actually three separate charges, collected together by one company and paid onward to three different destinations. Almost every useful decision a business can make about payment costs follows from knowing which component is which.

Roughly speaking, interchange is by far the largest of the three, assessments are the smallest, and markup sits in between and varies enormously. That ordering is stable even though the individual numbers change: card networks revise their interchange schedules on a regular published cycle, typically twice a year, and processor pricing changes whenever a contract permits it. Any guide quoting exact percentages as current fact is out of date, which is why the useful skill is reading structure rather than memorizing tables.

The mistake most businesses make here is negotiating the wrong number. They push a sales representative on "the rate," accept a small reduction in the headline figure, and never ask which of the three components moved. If the headline came down while the markup stayed the same, nothing was won.

Interchange: the largest component, and the one nobody negotiates

Interchange exists to compensate the issuing bank for funding the transaction, carrying the fraud and credit risk on the cardholder side, and — increasingly — for funding the rewards program that persuaded the cardholder to use that particular card. It flows from the acquirer to the issuer on every transaction, and the card network merely publishes the schedule and enforces it.

Two consequences follow, and they are the reason interchange dominates every conversation about card costs. First, no processor can sell interchange more cheaply than another, because they all pay the same published schedule for the same transaction. A provider that claims to have negotiated better interchange is either describing a genuine but narrow program — certain government, utility, education, charity, and small-ticket categories have their own published rates — or is misdescribing its own markup. Second, because interchange is the same for everyone, the entire competitive difference between providers sits in the markup and the fee list, which are exactly the parts that tend not to be published.

Interchange has been the subject of litigation and regulation for two decades. The US interchange antitrust multidistrict litigation has run since 2005 and remains unresolved in its injunctive-relief phase, the Department of Justice sued Visa in 2024 over its position in debit, and the European Union has capped consumer card interchange on cross-border transactions since 2015. None of that changes the structure a merchant deals with today: the schedule is published by the network, it applies to everyone, and the merchant's leverage is elsewhere.

What determines which interchange rate applies

Interchange is not one rate. Each network publishes hundreds of categories, and which one a given transaction falls into is decided mechanically from the data attached to it. Five variables do most of the work.

Merchant category code

The merchant category code, or MCC, is a four-digit code assigned when the merchant is boarded that classifies what the business sells. It is the first thing a business should verify on its own account, because interchange programs are written per category: supermarkets, fuel, charities, education, utilities, restaurants, and travel all have their own treatment, and some categories carry rate ceilings or special small-ticket programs that others do not. A business coded incorrectly at boarding — a specialty food retailer coded as general merchandise, a nonprofit coded as a retail store — can pay a materially different cost on every transaction for years. Changing an MCC after the fact is possible but requires the acquirer to reboard the account, and no salesperson raises it unprompted.

Card-present versus card-not-present

A transaction where the chip was read or the card was tapped at a terminal is card-present. A transaction keyed by hand, taken over the phone, or paid online is card-not-present. Card-not-present interchange is higher across essentially every category, for the straightforward reason that the fraud rate is higher and the issuer is carrying more of that exposure. This is why the same card, on the same day, costs a business more when the customer reads the number aloud than when they tap it — and why a business that keys a meaningful share of its in-person transactions has a fixable cost problem rather than a pricing problem.

Card type and issuer

Consumer debit, consumer credit, rewards credit, and commercial cards occupy different interchange tiers. Rewards and premium consumer credit cards carry the highest consumer interchange because the interchange is funding the points, the miles, and the lounge access. Commercial, corporate, and purchasing cards are higher again, but they are also the one category where a merchant can actively reduce cost by supplying enhanced data — the invoice, tax, and line-item detail known as Level 2 and Level 3 — which qualifies the transaction for a cheaper commercial program. Businesses that sell to other businesses and do not pass Level 3 data are paying for a discount they are entitled to and not claiming.

Transaction and authorization quality

Whether address verification ran, whether a security code was collected, whether the authorization matches the settled amount, whether a card-on-file transaction was correctly flagged as such, and how quickly the transaction was captured all affect the final category. These are integration settings, not commercial terms.

Regulated versus unregulated debit

US debit is a category of its own because of the Durbin Amendment to the 2010 Dodd-Frank Act and the Federal Reserve rule implementing it, Regulation II. Durbin did two things that a merchant needs to understand.

First, it capped debit interchange for covered issuers — banks and credit unions holding assets above a statutory threshold. Debit cards issued by those institutions are described in the industry as regulated debit, and their interchange is a capped amount rather than a schedule the network sets freely. Debit cards from smaller institutions below the threshold are unregulated or exempt debit, and their interchange is set the ordinary way and is meaningfully higher. A merchant has no way of knowing at the point of sale which kind of debit card is in front of it, and no way of influencing it.

Second, Durbin required every debit card to be enabled on at least two unaffiliated networks, so that the transaction can be routed over more than one path. That routing choice belongs to the merchant and its processor, not the issuer, and the cost of the two paths is not identical. Debit routing optimization is one of the few genuine cost levers in card acceptance, and it exists only because of that provision.

Network assessments: small, fixed, and unavoidable

The card networks earn their revenue not from interchange but from the fees they charge issuers and acquirers for using the network. On a merchant statement these appear as assessments and as a set of per-transaction network charges, which the industry variously calls network access fees, data-processing fees, acquirer processing fees, and settlement or switching fees, depending on the network.

Assessments are a small percentage of volume plus a small flat amount per authorization. They are identical for every merchant of a given profile, they are not negotiable, and no processor discounts them. Two structural points are worth knowing. Cross-border and currency-conversion charges apply when the issuing bank is in a different country from the merchant, and those additions are not trivial for a business selling internationally from a domestic account. And Visa and Mastercard do not publish the fee schedules they charge acquirers, which means that although assessments are pass-through in principle, a merchant cannot audit them from public documents. A processor padding a network fee by a fraction of a cent per transaction is very difficult to catch, which is one more reason to insist that a statement itemize network fees separately rather than fold them into a blended line.

Processor markup: the only component you can actually negotiate

Everything left after interchange and assessments belongs to the acquirer, the processor, the gateway, and the sales organization that signed the account. It takes three shapes, and they are not mutually exclusive.

• A percentage of volume — the "plus" in interchange-plus, or the invisible spread inside a flat rate.
• A per-transaction amount — a few cents per authorization, which matters far more to a business with a small average ticket than to one selling in thousands.
• Fixed periodic fees — monthly account, statement, gateway, PCI compliance, PCI non-compliance, minimum-volume, batch, terminal rental, and annual fees.

How visible the markup is depends entirely on the pricing model. Under interchange-plus it is stated explicitly and can be compared between providers directly; Helcim publishes its margin table by volume tier, Adyen states its processing fee separately from pass-through interchange and scheme fees, and Stax Payments replaces the percentage markup with a fixed monthly subscription plus a per-transaction amount. Under a blended flat rate — Square, Stripe on standard terms, Shopify Payments, and the published small-business rates of several bank acquirers — the markup is not disclosed at all, because the whole point of the model is that the merchant is quoted one number regardless of what the underlying transaction cost. Under tiered pricing the markup is not merely undisclosed, it is actively obscured.

The mistake most businesses make here is to treat the fixed fees as noise. A short list of monthly charges, each individually small, will comfortably exceed the entire percentage markup for a business processing modest volume. Add the fixed fees up, divide by monthly volume, and they become a rate — which is the only fair way to compare them against a competitor's quote.

Downgrades: why the effective rate exceeds the quote

A downgrade is a transaction clearing into a more expensive interchange category than the best one it was eligible for, because a data or timing requirement was not met. Downgrades are the single largest source of the gap between the rate a merchant was quoted and the rate it actually pays, and they are largely self-inflicted and largely fixable.

The recurring causes:

1. Late settlement. Transactions batched a day or more after authorization fall out of the qualifying window. Automatic daily batch closing solves this permanently.
2. Missing address verification or security code on keyed and online sales. The cheaper card-not-present categories require them.
3. Missing Level 2 and Level 3 data on commercial cards. Business, corporate, and purchasing cards qualify for cheaper programs only when tax amount, customer code, and line-item detail are passed. Most gateways support this and most merchants never enable it.
4. Authorization and settlement amount mismatch. Tip adjustments, partial shipments, and incremental authorizations must be handled with the correct transaction flags rather than by settling a different figure.
5. Card-on-file and recurring transactions flagged incorrectly. Stored-credential frameworks exist precisely so that these clear correctly; unflagged, they can be treated as ordinary keyed transactions.
6. International cards. Not a downgrade in the strict sense, but they carry cross-border and currency costs a domestic-only quote never mentioned.

A merchant on interchange-plus can see downgrades on the statement, because the interchange line is itemized by category. A merchant on tiered pricing cannot see them at all — they simply appear as more volume in the expensive bucket, with no way to distinguish a genuine premium-card transaction from an avoidable data failure. That opacity is not a side effect of tiered pricing; it is what tiered pricing sells.

The fees that are not per-transaction

The percentage argument absorbs all the attention, and meanwhile the fixed and incidental charges quietly do their work. The list worth auditing line by line:

Equipment leasing deserves a specific warning. A terminal lease is often a separate agreement with a separate finance company, does not terminate when the processing agreement does, and can outlast the hardware itself. Providers that sell hardware outright rather than leasing it remove an entire category of dispute.

What to do with all of this

Four actions, in order of return on effort:

1. Verify the merchant category code on the account. It is one field, it determines which interchange programs apply, and it is wrong more often than anyone admits.
2. Fix the transaction-quality problems. Automatic daily batching, address verification and security codes on card-not-present sales, Level 2 and Level 3 data if selling to businesses, correct stored-credential flags. None of this requires a renegotiation and all of it reduces interchange.
3. Get onto pricing that itemizes. A statement that shows interchange by category, network fees separately, and markup as its own line is the precondition for every other improvement. Without it, there is nothing to audit.
4. Negotiate the markup and the fixed fees, and re-check annually. Interchange and assessments are fixed for everyone. Markup, monthly fees, and minimums are the whole negotiation, and pricing that was competitive when the account was opened rarely stays that way.

The businesses that pay the least are not the ones with the best-negotiated rate. They are the ones whose transactions are clean, whose category code is correct, whose statement is legible, and who know that the only number worth arguing about is the one their processor keeps.

Frequently asked questions