Stripe vs Adyen vs Checkout.com: Compare PSP Pricing on the Same Basis

A pricing deck can make three payment service providers look comparable when the cells are measuring different things. A US standard card rate is an offer under published conditions. An indicative Interchange++ row shows components but not the final bill for your card mix. “Contact sales” is a request for a contract quote, not a zero or an unknowable price.

That is the useful distinction in a Stripe, Adyen and Checkout.com comparison. The question is how much each would charge this merchant, for this traffic and these events. No public page can answer it for every geography, product and contract.

What each public pricing page establishes

The table should not be read horizontally as “2.9% versus 0.60% versus unknown.” Adyen's 0.60% is one component added to interchange, scheme charges and a processing fee. Stripe's standard number bundles some costs and may have conditional additions. Checkout.com's quote may be structured in either of the models it describes. Even where two offers use Interchange++, the definition of passed-through fees and the fixed-event charges need to match.

Build one transaction profile before asking for prices

Use the PSP quote-cost calculator to compare two complete fee schedules on the same transaction profile. It shows the arithmetic and exclusions; its demo rates are fictional and it does not supply current provider quotes.

Give all three providers the same data pack. State the merchant's legal and processing countries, online versus in-person split, card-present status, expected monthly count and value, average and distribution of ticket sizes, domestic versus international issuer mix, consumer versus commercial cards, and payment and settlement currencies. Identify the payment methods and optional fraud, authentication or payout products actually needed.

Then request a written schedule for:

1. The percentage and fixed charges on authorization, capture or successful settlement, and the exact event that triggers each. Ask whether a refused or risk-declined attempt, an expiry, a void, a retry or a refund request can still create a fee. Use the written fee schedule to identify those events, then reconcile the charges to the itemized invoice.
2. Interchange and scheme pass-throughs where relevant. Ask how card type, issuer country, merchant category, card-present status and authentication affect the line items. Do not borrow a consumer-card interchange assumption for commercial cards.
3. Cross-border and FX terms: the card or account country that triggers an international surcharge, the conversion condition, the exchange-rate reference and the settlement currency. For Stripe's US standard schedule, an international card surcharge and conversion charge are separate conditions, not automatic add-ons on every foreign card.
4. Ancillaries and commitments: disputes received and contested, refunds, fraud tools, 3DS, payouts, monthly minimums, volume tiers and any gateway or platform charge. Request a sample settlement statement that demonstrates how each shows up.

A merchant with a thousand USD 20 orders will care more about a per-transaction fixed charge than one with a thousand USD 2,000 orders. A merchant whose customers mostly use regulated domestic consumer debit has a different pass-through profile from one selling to international commercial-card holders. Use counts and value by segment; one blended annual volume number hides both effects. The MDR stack reference shows how to reconcile the resulting invoice.

Worked example: the lower percentage does not always win

The following Quote A and Quote B are invented for arithmetic only. They are not offers from Stripe, Adyen or Checkout.com. Assume USD online payments, one successful capture per approved order, a 95% approval rate on unique attempts, and no separately billed attempts, refunds, FX, disputes, pass-through fees or minimums. Those exclusions make the calculation transparent, not a complete merchant cost estimate.

• Quote A: 2.4% of successful settled value, plus USD 0.10 for each successful payment.
• Quote B: 1.9% of successful settled value, plus USD 0.25 for each successful payment.

Quote A: 2.4% + USD 0.10 Quote B: 1.9% + USD 0.25

USD 30 is the crossing point. Below it, A costs less. Above it, B costs less. At USD 30, each costs USD 0.82.

• USD 10 paymentA costs less

  Quote A

  USD 0.34

  Quote B

  USD 0.44

• USD 30 paymentSame fee

  Quote A

  USD 0.82

  Quote B

  USD 0.82

• USD 100 paymentB costs less

  Quote A

  USD 2.50

  Quote B

  USD 2.15

Fee per successful payment, USD

Ticket	Quote A	Quote B
$10	$0.34	$0.44
$20	$0.58	$0.63
$30	$0.82	$0.82
$50	$1.30	$1.20
$100	$2.50	$2.15
$200	$4.90	$4.05

Calculation: rate × ticket amount + fixed fee. Excludes failed attempts, refunds, disputes, FX, pass-through fees and minimums. Compare the same ticket distribution and contracted chargeable events.

Profile 1 — USD 20 tickets: 10,000 attempts × 95% approval = 9,500 successes, settling USD 190,000. Quote A costs USD 4,560 in percentage fees plus USD 950 in fixed fees, or USD 5,510. Quote B costs USD 3,610 plus USD 2,375, or USD 5,985. A is lower by USD 475 on these known fees.

Profile 2 — USD 200 tickets: 1,000 attempts × 95% approval = 950 successes, also settling USD 190,000. Quote A costs USD 4,560 plus USD 95, or USD 4,655. Quote B costs USD 3,610 plus USD 237.50, or USD 3,847.50. B is lower by USD 807.50 on these known fees.

Both profiles settle USD 190,000, yet the fixed fee is applied 9,500 times in one and 950 times in the other. In this deliberately narrow model, the fee lines cross at a USD 30 ticket. These subtotals are not a forecast of savings from switching provider.

If Quote B also bills an unknown fee per declined authorization or has an undisclosed monthly minimum, the total is unknown pending quote. The low-ticket profile has 500 unsuccessful attempts and the high-ticket profile has 50; an attempt-level charge would not affect them equally. Approval was held at 95% solely to isolate pricing, not to assert that different providers approve the same transactions. An itemized invoice helps reconcile actual charges, but the written agreement must establish whether attempts or refund requests are billable. Send the same assumptions and event counts through the PSP RFP template before making a real selection.

The calculator's multi-event and incomplete-contract example prices attempts, refunds, disputes, fixed fees and FX together: 94.90 hypothetical currency units. It shows why an unknown refund fee leaves the comparison unavailable, while an explicitly confirmed zero produces 92.40 CU under the same exclusions. These are fictional terms, not provider prices.

Model approval changes as a separate hypothesis

Provider optimization tools may improve authorization, but a vendor-reported uplift is not a transfer-ready estimate for another merchant. Clarify whether a stated figure is a relative percentage or a percentage-point change; what counted in the denominator; whether the comparator was a matched control; and whether the result is for the same card, region, channel and period. Approval by attempt, by unique order and by payment value are different metrics.

An illustrative calculation shows the boundary. Suppose 100,000 unique USD 100 orders generate USD 10 million in attempted value, and approval rises from 90% to 92% on comparable traffic. That is up to 2,000 additional initially approved orders and USD 200,000 of additionally approved value. It is not automatically USD 200,000 of incremental revenue. Some orders would succeed on a later retry or through another provider; some approved payments will be refunded or disputed; and contribution after goods, fulfillment and payment costs is less than gross sales. Measure the genuinely incremental completed orders and their contribution before comparing that gain with the fee delta. The authorization-rate economics guide covers denominator design in detail.

Pricing and approval therefore need two parallel models: a fee ledger on matched transaction events, and a controlled performance test on matched traffic. A cheaper quote need not win if it performs worse on your audience; a vendor's marketing uplift does not prove its higher fee is worthwhile.

How to compare three quotes without inventing a winner

Normalize each quote on the same month of historical traffic. Reprice every transaction segment and add separately billed attempts, refunds, disputes, FX, payout and fixed charges. Calculate total payment cost divided by successful settled value, but keep gross attempted value, unique orders and successful orders visible alongside it. A provider can appear cheaper per settled dollar by rejecting more of the harder payments; another can appear to have a higher bill because it successfully processes more orders.

Ask each provider to explain a hypothetical declined payment, void, capture and partial refund on its proposed schedule. The new failed-payment and refund fee reference maps those events and separates returned customer principal from retained merchant fees. If any line is inaccessible until contract, mark it unknown pending quote rather than filling the cell with zero.

The decision belongs to the merchant's traffic and agreement. The public pages define starting points and questions for the RFP. They do not establish a universal cheapest processor, a guaranteed authorization gain or a one-size-fits-all enterprise threshold.