Skip to content
All topics

TOPIC BRIEFING

PSP & Infrastructure

Your PSP stack is not just a vendor choice. It decides who controls authorisation, settlement, tokens, redundancy, data portability, and market expansion.

Payment infrastructure is the operating layer behind acceptance. PSP selection, acquiring setup, MoR vs direct processing, token strategy, orchestration, settlement flows, and contract control decide how much flexibility an operator has when volume, markets, and risk increase.

32 briefings PSP vs MoRGateway vs AcquirerOrchestration & RoutingTokens, Settlement & Lock-In

Stack map

The payment stack decisions that compound

Each layer below has its own provider, its own contract, and its own operator implication. Compounding effects: the choice you make at one layer shapes the leverage you have at every layer downstream.

  1. 01

    Checkout

    Where payment intent starts — UX, conversion, and the first data captured.

  2. 02

    Gateway

    API, encryption, routing interface — the integration surface you maintain.

  3. 03

    PSP / Acquirer

    Authorisation, acquiring, settlement, and chargeback exposure sit here.

  4. 04

    Orchestration

    Redundancy, routing logic, and vendor abstraction across multiple PSPs.

  5. 05

    Token Vault

    Portability, recurring-payment resilience, and the deepest source of lock-in.

  6. 06

    Settlement

    Cash timing, reserves, and reconciliation complexity across providers and markets.

  7. 07

    Reporting & Reconciliation

    The source of truth across PSPs, markets, and currencies — and the audit trail.

The operator thesis

Three operator takes

01

Outsource processing, not control

A PSP can handle acquiring, routing, risk tooling, settlement files, and reporting — but operators should stay clear on which parts of the stack remain portable, negotiable, and independently measurable.

02

One PSP is simple. It is also concentration risk.

Single-PSP setups reduce operational complexity, but they also concentrate outage exposure, pricing leverage, routing options, token dependency, and renewal risk in one vendor relationship.

03

Portability is designed before you need it

Token strategy, data access, reconciliation structure, failover design, and contract exit clauses determine whether switching PSPs is a controlled migration or a painful rebuild.

Start here

Reading paths for PSP & Infrastructure

Choosing your PSP stack

Decide your operating model before evaluating providers. These three articles in order.

Switching or scaling your stack

Understanding the real cost of your current PSP before adding complexity or moving.

Merchant of record decisions

For software companies evaluating Paddle, Polar, or Lemon Squeezy — and when to migrate off.

Briefings, grouped by decision

32 briefings in PSP & Infrastructure

Other briefings in this topic

Reference

Frequently asked

What is the difference between a payment gateway, a PSP, and an acquirer?

A payment gateway is the technical layer that transmits payment data between your checkout and the acquiring bank — it handles encryption, routing, and the API interface. A PSP (Payment Service Provider) bundles the gateway with acquiring, often adding fraud tools, reporting, and local payment method support in a single contract. An acquirer (or acquiring bank) is the licensed financial institution that holds the merchant account, settles funds, and takes on chargeback liability. Most merchants deal with a PSP that packages all three functions — Stripe, Adyen, and Checkout.com are PSPs with their own acquiring licences in most markets they operate in.

When should a business use a Merchant of Record instead of a direct PSP?

A Merchant of Record (MoR) makes sense when tax and compliance complexity exceeds your team's capacity to manage it. If you sell digital goods internationally, an MoR (Paddle, Polar, FastSpring) handles VAT registration in 100+ countries, local compliance, and tax remittance on your behalf — you invoice the MoR, not the customer. The tradeoff is cost (MoR take rates are typically 4–6% vs 1–3% for direct PSP) and control (the MoR owns the customer relationship for payment purposes). At scale above roughly $3–5M ARR with international digital sales, most operators evaluate whether direct PSP with dedicated tax infrastructure is more economical.

What are the most important clauses to check in a PSP contract?

The five highest-risk clauses: (1) Rolling reserve — how much of your funds the PSP holds and for how long; (2) Termination rights — whether the PSP can terminate with minimal notice and freeze funds; (3) Data portability — whether you can export your tokenised card vault if you switch PSPs; (4) Fee change notice period — how much notice you get before rate increases; (5) Chargeback liability — what thresholds trigger account review or termination. Most PSP contracts are heavily weighted toward the PSP's risk management needs, not merchant flexibility. The contracts for Stripe, Adyen, and Checkout.com all have specific red flags documented in the PSP contract red flags briefing.

What is payment orchestration and when does it make sense?

Payment orchestration is a middleware layer (Spreedly, Primer, Gr4vy) that sits between your checkout and multiple PSPs, routing transactions intelligently — by geography, payment method, transaction value, or real-time success rates. It makes sense when you have: (1) multi-PSP redundancy needs (if one PSP goes down, failover automatically); (2) market-specific acquiring requirements (needing local acquirers in different countries for better auth rates); or (3) complex routing logic that your primary PSP can't express. For single-PSP merchants below roughly $10M in processing volume, orchestration adds integration overhead without proportional benefit. At scale, it is a meaningful authorisation rate and cost lever.

How do network tokens differ from PSP vault tokens — and does it matter?

A PSP vault token (Stripe's tok_xxx, Adyen's recurring reference) is specific to that PSP — it represents your customer's card as stored in that PSP's system. If you switch PSPs, the token is useless and you have to re-vault all cards (expecting 10–25% customer drop-off during migration). A network token (Visa Token Service, Mastercard MDES) is issued by the card scheme and works across any PSP that supports network tokenisation. Network tokens also self-update when cards are reissued, eliminating subscription churn from card expiry. The practical implication: operators with large stored-card subscriber bases should prioritise PSPs that automatically provision network tokens, as it reduces both migration lock-in and recurring billing failure rates.

PSP selection, acquiring structure, gateway architecture, and routing logic are not commodity decisions — they directly determine authorisation rates, processing costs, vendor leverage, and how quickly an operator can adapt when a market shifts. The stack architecture you build in year one is often the one you’re working around in year three.

The decisions in this space have compounding consequences. A PSP contract with a punitive rolling reserve provision can tie up six figures of working capital. A vault tokenisation model that stores PSP-specific tokens rather than network tokens creates a de facto lock-in that costs 15–25% customer churn to escape. A single-acquirer architecture without routing logic leaves 2–5% of authorisation rate improvement unrealised. These are not theoretical risks — they are operational costs that most operators only discover after the contract is signed.

The briefings in this topic cover the full stack: how to evaluate and choose a PSP, how to structure multi-acquirer routing for resilience and rate improvement, how to read a PSP contract before you sign it, and when the complexity of direct acquiring outweighs the benefits of staying with a payment facilitator or merchant of record. The framing question throughout is the same: at every layer of the stack, who controls the next decision — you, or your PSP?