What you're buying when you buy acceptance
The price of accepting a card payment is made up of several separate charges, each set by a different party. They are not negotiated in the same place, and some are not negotiable at all. On a four-party card transaction, three layers are universal. Interchange goes to the issuing bank. Scheme fees are charged by the network on both sides at once, and the acquiring margin pays the provider. A fourth layer, the fee schedule, covers charges billed per event, and that is often where two offers really differ. This breakdown applies only to four-party systems. In three-party models, where the network is both issuer and acquirer, there is no interchange. The merchant gets a single price that it can only accept or reject.
| Layer | Who sets it | Negotiable by the merchant | What drives it |
|---|---|---|---|
| Interchange | The network, within any limits set by a regulator | No, never directly | Card category, issuing country, channel, data submitted, incentive program |
| Scheme fees | The network alone, through a unilateral fee schedule | No, but how they're passed through is | Volume, brand mix, services enabled, program penalties |
| Acquiring margin | The acquirer or provider | Yes, and it's the only line that really gets negotiated | Volume, industry risk, average ticket, contract length |
| Fee schedule | The provider, sometimes the network | Yes, item by item | Disputes, declines, currency conversion, payouts, minimums, optional services |
Six ways to bill the same payment
The billing model is how a provider builds the price it charges the merchant and how it presents that price on the statement. It doesn't change the transaction's underlying cost, but it determines who sees the detail and who captures the variations. Under blended pricing, the provider applies a single rate and keeps the difference from each transaction's actual cost. Under interchange++, it passes through actual interchange, then actual scheme fees, then adds a margin set in the contract: those are the two “pluses” in the name. Interchange+ stops at the first layer, and network fees are folded into the margin. A flat rate itemizes nothing, while a subscription shifts part of the price into a recurring fee. Tiered pricing, specific to the US market, sorts each transaction by processing conditions the merchant has little control over.
| Model | What's passed through at cost | Who keeps an interchange cut | Auditable from the statement | When it makes sense |
|---|---|---|---|---|
| Blended | Nothing: one rate per broad category | The provider | No | Low volume, stable mix, low admin overhead as the priority |
| Interchange+ | Interchange only | The merchant, on one layer | Partially | Mid-size volume; a trade-off between readability and reconciliation workload |
| Interchange++ | Interchange and scheme fees | The merchant, on two layers | Yes, line by line | Once volume justifies monthly monitoring |
| Flat rate | Nothing, and nothing is shown | The provider, entirely | No | Very low volume, or a secondary channel |
| Subscription + cost pass-through | Varies by contract | Depends on the pass-through clause | Depends on the contract | Steady, predictable volume; high average ticket |
| Tiered pricing | Nothing: a proprietary classification | The provider | No | Never; reject it in favor of an itemized quote |
The two extreme models both have published rate cards, available without talking to sales. Adyen (Adyen N.V., 2006) publishes an itemized structure. Its fixed processing fee is $0.13 per transaction, whatever the payment method. On top of that come “Interchange++ + 0.60%” on Visa and Mastercard, and €0.22 on iDEAL (Currence iDEAL B.V., 2005). The same rate card states that there are no monthly, setup, integration, or closure fees, while reserving the right to a minimum invoice depending on industry and business model. Stripe (Stripe, Inc., 2010) publishes the opposite structure in the UK: an all-in rate that varies by card origin. That means 1.5% + 20p on standard UK cards and 2.8% + 20p on premium UK cards. EEA cards cost 2.5% + 20p, and international cards 3.15% + 20p. IC+ pricing is available only on request. The two rate cards split information and monitoring work differently. The first shows the merchant the underlying cost and leaves the monitoring to them. The second delivers a single price whose makeup the payer never sees.
- Get the margin written out separately, in basis points and as a fixed amount, with volume tiers you can verify on the statement.
- Require the schedule of scheme fees passed through, not just the principle of passing them through: a “pass-through” without a schedule can't be checked.
- Require automatic pass-through of any regulatory cut in interchange, with a specific deadline. This clause is negotiated before signing, never after.
- Limit unilateral changes: a minimum notice period, a cap on annual increases, and the right to exit without penalty if the cap is exceeded.
- Reject proprietary tiers: a qualified / non-qualified classification moves a transaction to a more expensive rate even though the underlying interchange hasn't changed.
Scheme fees: the layer no one caps
Scheme fees are what a card network charges for routing, for its brand, and for the services it mandates or sells. They differ from interchange in three ways, and those differences shape how they should be handled in a contract. They are charged on both sides of the transaction, issuer and acquirer. They are set unilaterally, with no negotiation or contractual consideration. And so far no regulator caps them: the EU's Interchange Fee Regulation (IFR) explicitly leaves this layer out of scope.
- Authorization fees: billed per request, including on declined transactions. A high decline rate costs you twice.
- Clearing fees: billed at presentment, usually as a fixed amount.
- Brand license fees: an ad valorem charge based on acquired volume, regardless of transaction count.
- Cross-border and currency fees: triggered by the card's issuing country or the settlement currency, and stacked on top of interchange that is already higher.
- Connectivity and data fees: network access, message volumes, reporting.
- Value-added services: tokenization through Visa Token Service or Mastercard Digital Enablement Service (both launched in 2014), authentication, account updater for stored cards, network fraud tools.
- Program penalties: exceeding dispute or fraud thresholds, excessive authorization attempts, authorizations left without capture or reversal.
The growth of this layer has been measured publicly. In its MR22/2 market review, the UK Payment Systems Regulator found that Visa and Mastercard scheme and processing fees had risen by more than 30% in real terms, with no matching improvement in service (PSR, 2024). The same regulator quantified another increase. After Brexit, card-not-present payments between the UK and the EEA moved to inter-regional rates, rising from 0.2% and 0.3% to 1.15% and 1.50%. The annual cost to UK merchants and their customers is £150 million to £200 million (PSR, MR22/2, final report of December 13, 2024). That increase required no renegotiation. It followed from a change in geographic scope.
A second check concerns who the acquirer actually is. Ask in writing for the name of the acquiring bank and the network agreement that covers it. A provider that resells another company's acquiring can't quote interchange++, can't produce a breakdown by network, and can't resolve a billing dispute quickly, because it doesn't have the information itself. The question applies in every market, and the answer determines what the merchant will later be able to see, challenge, and negotiate.
The fee schedule: where the real difference lies
The fee schedule is the part of the acceptance contract that lists the fees billed outside the percentage applied to processed volume. It covers charges triggered by an event rather than by an amount: a dispute, a refund, a payout, a currency conversion, a month below the minimum. These lines are negotiated item by item, not as a rate. To cost them before signing, you have to apply them to the merchant's actual volumes; otherwise their weight is unknown.
| Card type | Billing basis | Trigger | Merchant lever |
|---|---|---|---|
| Disputes | Fixed fee per case, often due whatever the outcome | Every dispute received, including abusive ones | Upstream deflection, structured evidence, network prevention tools |
| Refunds | Fixed fee, sometimes with the original fee not returned | Every refund issued | Void before clearing presentment rather than refund afterward |
| Authorization | Fixed fee per attempt | Every request, approved or declined | Cut unnecessary attempts; stop retrying hard declines |
| Currency conversion | Percentage markup on the applied rate | Transaction currency differs from settlement currency | Multi-currency settlement accounts, separate FX contract |
| Payouts and batches | Fixed fee per batch or per payout, or a percentage of the amount paid out | Chosen settlement frequency, accelerated settlement, payout currency | Pay out less often; weigh cash flow against cost |
| Monthly minimum | Shortfall charged if volume falls below the threshold | Seasonality, secondary channel, low-activity entity | Negotiate the threshold per entity, not at group level |
| Account and compliance | Subscription, inactive account fee, non-compliance fee | Security program, threshold breaches, dormancy | Audit services billed but not used |
Published rate cards make it possible to cost this schedule without guesswork. On its UK rate card, Stripe charges £20 per dispute received, refunded if the merchant wins. An instant payout costs 1% of the amount, with a 40p minimum. Settlement in a currency other than the transaction currency costs 1% of the amount paid out, and currency conversion adds 2%. Adyen advertises the opposite structure: no monthly fees, no setup fees, and no closure fees, but a minimum invoice that depends on the industry. The two providers put the price on different lines: one on settlement and FX events, the other on a billing floor. Comparing them requires applying both to the same traffic.
- Have every fee schedule line costed against the merchant's actual volumes for the past year, before comparing rates.
- Check what happens to the original fee on a refund: returned, partly returned, or kept. The difference shows up in sectors with high return rates.
- Separate the cost of FX from the cost of payment: a provider's conversion markup should be compared with a bank FX contract, not with an acquiring rate.
- Count the payouts: daily settlement multiplies fixed fees twentyfold compared with weekly settlement, sometimes for no cash flow benefit at all.
- Hunt down services billed but unused: duplicate fraud tools, premium reporting, idle terminals, merchant accounts opened and never closed.
The cost you actually bear, and how to calculate it
The effective cost of acceptance divides all payment-related fees over a period by the gross volume processed over the same period. It is used to compare an offer with the cost actually borne, then to track contract performance month to month. It includes disputes and authorization fees, including those on declined transactions. Minimums, subscriptions, and payout fees also count. The denominator is gross volume submitted, before any deduction; otherwise the result understates the real cost. Some jurisdictions already make providers do this calculation. The Reserve Bank of Australia's standard requires acquirers to give each merchant an annual statement of its average cost of acceptance, by card system and expressed as a percentage. That figure sets the ceiling on permitted surcharges. It gives merchants, at no cost, a measure that merchants in other markets have to rebuild from their own data.
TRAFFIC ASSUMPTION (replace with your last twelve months)
transactions ............ 100,000 / month
average ticket .......... 42.00 GBP
gross volume ............ 4,200,000 GBP
mix by card origin 78% UK standard | 9% UK premium
8% EEA | 5% international
disputes ................ 0.08% of transactions -> 80 cases
converted volume ........ 5% of gross volume
RATE CARD APPLIED (Stripe, published UK pricing, August 2026)
UK standard ..... 1.50% + 0.20 GBP
UK premium ...... 2.80% + 0.20 GBP
EEA ............. 2.50% + 0.20 GBP
international ... 3.15% + 0.20 GBP
dispute ......... 20.00 GBP per case received
conversion ...... + 2.00% of converted amount
CALCULATION
weighted ad valorem component
0.78x1.50 + 0.09x2.80 + 0.08x2.50 + 0.05x3.15 ....... 1.78%
fixed component
100,000 x 0.20 = 20,000 GBP on 4,200,000 GBP ....... 0.48%
disputes
80 x 20 = 1,600 GBP on 4,200,000 GBP ............... 0.04%
conversion
5% of volume x 2% ................................. 0.10%
---------------------------------------------------------------
EFFECTIVE COST OF ACCEPTANCE ......................... 2.39%
advertised headline rate ............................. 1.50%
gap .................................. +0.89 pt, i.e. +59%The calculation reveals two effects that hold whatever rate card is used. The card origin mix moves the cost without any line of the contract changing: as a customer base becomes more international, a growing share of cards falls into the highest rate bands. The average ticket determines the weight of the fixed fee, which dominates below a certain amount. For a merchant with small tickets, the negotiation should therefore focus on the fixed per-transaction fee rather than on the percentage.
| Average ticket | Ad valorem component | Fixed fee as % of amount | Total cost |
|---|---|---|---|
| £8 | 1,50 % | 2,50 % | 4,00 % |
| £25 | 1,50 % | 0,80 % | 2,30 % |
| £42 | 1,50 % | 0,48 % | 1,98 % |
| £120 | 1,50 % | 0,17 % | 1,67 % |
| £600 | 1,50 % | 0,03 % | 1,53 % |
Putting two offers on the same basis
Comparing two proposals means calculating, for each one, what it would have charged on traffic you've already processed. A comparison that stops at published rate cards compares price structures, not amounts. Normalizing requires a transaction-level file covering a rolling 12 months, seasonality included. A merchant that doesn't know its own effective cost has no hard benchmark to hold up against a quoted rate. The comparison then comes down to how the offers are presented rather than what they cost.
- Extract 12 months of transactions at transaction level, with gross amount, currency, brand, card category, BIN issuing country, channel, authorization result, and raw response code.
- Add the events: disputes, refunds, voids, payouts, conversions.
- Rebuild the cost actually paid over the period by reconciling the batch report, the payout notices, and the provider's statement. That is the benchmark everything is measured against.
- Apply each candidate rate card to the same file, line by line, with no rounded mix and no theoretical average ticket.
- Fill in the itemized models: an interchange++ quote is incomplete until the applicable interchange and the scheme fee schedule have been provided. Without them, it can't be compared with an all-in rate.
- Add the fee schedule: disputes, minimums, subscriptions, payouts, conversion, and the optional services actually used.
- Neutralize currencies: convert every rate card into a single currency at a dated exchange rate, and document that rate.
- Rerun the calculation under two volume scenarios, one lower and one higher, to surface thresholds, minimums, and volume discount tiers.
| Item | All-in rate card (e.g., Stripe UK) | Itemized rate card (e.g., Adyen) | What you need in order to compare |
|---|---|---|---|
| Ad valorem component | 1.5% to 3.15% depending on card origin | 0.60% margin advertised on Visa and Mastercard | The interchange rates that apply to the merchant's actual mix |
| Fixed fee | £0.20 per transaction | $0.13 per transaction, all payment methods | A dated exchange rate and a single reference currency |
| Scheme fees | Included in the advertised rate | Passed through at cost, not published | The pass-through schedule, line by line, with its calculation basis |
| Recurring fees | Per-module subscriptions, from £450 per month for recurring billing | No monthly fees advertised, but a minimum invoice depending on industry | The minimum amount, and the legal entity it applies to |
| Disputes | £20 per case, refunded if the dispute is won | Not published | The per-case fee, and what happens to the original fee |
| Currency exchange | +2% for conversion, 1% for multi-currency settlement | Multi-currency settlement accounts, presented as a way to avoid conversion | The list of settlement currencies available without conversion |
PER TRANSACTION
unique reference ......... network identifier, unique, not reconstructed
processing date .......... distinct from the sale date
gross amount + currency .. before any deduction
brand .................... network actually used, not the logo on the card
card category ............ consumer / commercial, debit / credit / prepaid
BIN issuing country ...... this is what pushes a transaction outside a cap
channel .................. card-present / card-not-present, authenticated or not
result ................... approved / declined + raw response code
interchange applied ...... rate and amount
scheme fees .............. by component, with its original name
provider margin .......... rate and amount, shown separately
PER PERIOD
disputes ................. count, amount, outcome, fees charged
refunds .................. count, amount, original fee returned or not
payouts .................. count, currency, fees, actual time to credit
conversions .............. volume converted, rate applied, markup
recurring lines .......... subscriptions, minimums, account fees
FINAL CHECK
sum of fees = gross submitted minus net paid outWhen a central bank sets the price
Outside cards, on a growing share of retail payment rails, the merchant service charge (MSC) is published by a public authority or imposed by the system's owner. It is not negotiated with a provider. That is a different regime from cards, where the price comes out of a commercial contract. Price negotiations then focus on items other than the MSC, since it is fixed outside the contract. A third family follows neither logic: private wallets and mobile money. There, the operator sets the price itself, with no public cap and no industry counterweight. M-Pesa (Safaricom plc / M-Pesa Africa, 2007), Alipay (Ant Group, 2004), GCash (G-Xchange, Inc., 2004), and Mercado Pago (MercadoLibre Inc., 2004) fall into this third category. Merchants are handed the rate card with no discussion, and negotiation becomes possible only at very large scale.
| System | Operator, launch year | Who sets the price | Level | Source |
|---|---|---|---|---|
| Pix | Banco Central do Brasil, 2020 | The central bank for the rail, the market for the service | R$0.01 per 10 credits received, charged by the BCB to the participant; charging individuals and sole proprietors is prohibited; 2026 published market rates of roughly 0.89% to 1.45% for businesses, with a per-transaction cap in absolute terms | Resolução BCB nº 19 of October 1, 2020; institutions' published rate schedules, 2026 |
| Unified Payments Interface (UPI) and RuPay debit | National Payments Corporation of India, 2016 and 2012 | The legislature | Zero merchant service charge by law since January 1, 2020 | Payment and Settlement Systems Act 2007, Section 10A, inserted by the Finance Act 2019 |
| QRIS (Quick Response Code Indonesian Standard) | Bank Indonesia with the Asosiasi Sistem Pembayaran Indonesia, 2019 | The central bank, through a published rate schedule | 0% for micro-businesses below Rp 500,000, 0.3% above that, 0.7% for organized retail, 0.6% for education, 0.4% for gas stations, 0% for government payments and donations | Bank Indonesia, “MDR QRIS bagi Merchant” schedule |
| PromptPay | National ITMX, under a Bank of Thailand mandate, 2017 | The central bank | Free by mandate below a cap on person-to-person transfers, which in practice sets the market's price ceiling | Bank of Thailand |
| DuitNow QR | PayNet, 2019 | The system owner, under supervision | Mandatory national QR standard: banks and wallets accept the same code | PayNet |
| PayNow | Association of Banks in Singapore, operated by BCS, 2017 | The bank consortium | Priced to participating banks; more than 45% of Singapore's account-to-account transfer market in 2025 | Association of Banks in Singapore |
| BI-FAST | Bank Indonesia, 2021 | The central bank | Cap of Rp 2,500 chargeable to the customer, Rp 2,100 for bulk transfers; Bank Indonesia charges participants Rp 19, or Rp 16 for bulk | Bank Indonesia |
The Indonesian case shows where the economic decision moves to. QRIS pricing depends on the merchant category recorded at onboarding, based on the documents provided. The gap between 0.3% and 0.7% quickly exceeds the provider's own fee. A misclassified business pays the higher rate on all its past transactions, and retroactive correction is not the norm. Bank Indonesia also sets a second rule, on who bears the fee. The QRIS fee is paid by the merchant and cannot be passed on to the consumer. Surcharging the customer breaks that rule.
Where the rail is free or capped, the provider rebuilds its margin on other lines of the contract. It bills for the API and call volumes, for reconciliation and automatic matching, for splitting funds on marketplaces, for accelerated payouts, and for data delivery. Brazil offers the accounting proof. In 2025, Rede reported R$2.058 billion in service revenue and R$3.770 billion in financial intermediation revenue, with net income of R$1.888 billion. Credit and receivables prepayment bring in nearly twice as much as acceptance itself. A negotiation focused only on the MSC leaves the settlement delay and the prepayment rate off the table. The MSC is then only a fraction of the total price the merchant pays.
Managing pricing over time
Pricing governance means monitoring the cost of acceptance systematically between negotiations. It exists because an acquiring contract's real cost changes even when its terms don't. The card mix shifts, the customer base becomes more international, networks revise their fees, and regulators move the caps. Without monitoring, a gap between the negotiated price and the price actually paid opens up within 18 months, without anyone signing anything. The setup rests on three elements: clauses that limit the provider's unilateral decisions, a monthly set of metrics, and a documented annual review.
- Notice and cap on price changes: a minimum notice period, a limit on annual increases, and the right to exit without penalty above it. Canada's Code of Conduct provides model wording you can use anywhere.
- Pass-through of regulatory cuts: an automatic clause, a specific deadline, and an obligation to notify decreases just as for increases.
- Traceable network fees: every line passed through keeps its original name and calculation basis, with no bundling.
- Raw data delivery: response codes per transaction, card category, BIN issuing country, brand actually used. Without this clause, no audit is possible.
- Exit terms: commitment period, termination notice, portability of stored card tokens. If tokens lock you in, any renegotiation is purely theoretical.
| Indicator | Calculation | What drift signals |
|---|---|---|
| Effective cost of acceptance | Total fees for the month / gross volume for the month | The only figure to compare with the contractual commitment |
| Effective cost by segment | Same calculation by channel, country, and ticket size band | A stable average can hide a segment that's getting worse |
| Mix by card category | Consumer / commercial and domestic / foreign split | A more international customer base drives up cost without any rate change |
| Brand mix on co-badged cards | Share of each network used | Misconfigured routing, or drift in the default selection |
| Scheme fee trend | Amount by component, relative to volume | This layer drifts slowly, silently, and without notice |
| Cost of attempts per completed sale | Total authorization fees / number of successful sales | Pointless retries on hard declines, and a risk of program penalties |
| Dispute rate and cost | Number and cost of disputes / transactions | Approaching network program thresholds, well before any penalty |