Build the scorecard before you talk to anyone
The most useful thing you can do before the first sales call is decide how you will judge the answers. Write down your real profile: target countries, expected monthly volume, average order value, product category, refund rate, the payment methods your research says customers use, and whether you need subscriptions. Then give each criterion a weight out of ten based on what would hurt you most. A merchant selling two-dollar digital top-ups should weight method coverage and settlement speed heavily. A merchant selling furniture on instalments should weight cross-border eligibility and dispute handling. Without weights, whichever provider gives the most confident presentation wins, which is not a method. Also set a pass mark on the things that are disqualifying rather than scored, such as whether your category is accepted at all, because no amount of pricing advantage compensates for an account that gets closed.
- Write down your real volume, average order value, category and refund rate before the first call.
- Weight each criterion out of ten based on what would actually hurt your business.
- Set pass or fail gates for category acceptance and eligibility rather than scoring them.
- Fill the scorecard in during the call, not afterwards, while the answers are fresh.
Coverage: the methods your customers use
Coverage is not the length of a provider's logo wall. It is whether the two or three methods your customers actually use are live, on your entity, in your category, at your volumes. Test this concretely. Give each provider your target market list and ask them to confirm, method by method, whether they can board you today, and to name the legal entity that will contract and the one that will settle. Cross-check the answer against the scheme operator's own information, because central banks publish which institutions participate. Then check depth rather than breadth: does the provider support the specific variant you need, such as dynamic QR generation with a reference, recurring payments on a given rail, or refunds issued through an API rather than by email. A provider with fewer methods but the right ones beats a provider with many that are shallow.
- Ask for method-by-method confirmation that you can be boarded today, not a coverage list.
- Cross-check participation against the national scheme operator's own published information.
- Test for depth: dynamic QR with a reference, recurring support, API refunds, detailed settlement data.
- Name the methods your research says matter, and score providers only on those.
Cross-border eligibility
This is the criterion most likely to invalidate everything else. A provider may be excellent and still unable to contract with a company registered where yours is, or to settle to a bank in your country, or to support your category in that market. Ask three specific questions. Will your registration country be accepted for this market and these methods? Which legal entity in the group contracts with you, and which one settles to you, because they are frequently different? And does using the provider remove or merely reduce any requirement for a local entity, local tax registration or local bank account? Get the answers in writing, because a verbal assurance from a salesperson is not something you can hold anyone to later. If a provider is vague here, treat that as a low score rather than as a detail to resolve after signing.
- Confirm in writing that your registration country is accepted for each market and method.
- Identify the contracting entity and the settling entity separately, because they are often different companies.
- Ask whether local entity, tax or bank requirements are removed or only reduced.
- Score vagueness on eligibility as a low mark rather than an open item.
Onboarding friction
Time to live is a real cost, and providers vary enormously. Measure it in three parts: how long the document review takes, how long technical integration takes, and how long it takes to get a production account fully enabled for each method. Ask what typically causes delays on accounts like yours, which is a question good providers answer honestly and weak ones deflect. Then reduce your own contribution to the delay by preparing one complete document pack: registry extract, ownership chart down to natural persons, identification for owners and directors, proof of business address, bank account confirmation, published refund policy and a volume forecast. Providers ask for the same things in different orders, so one well-maintained pack is the single biggest lever you have on onboarding time when you are comparing several at once.
- Ask for typical review timelines and the most common causes of delay on accounts like yours.
- Measure time to live, not time to contract signature, because methods get enabled in stages.
- Prepare one complete document pack before applying, then reuse it across providers.
- Track how quickly each provider responds to document questions, since that predicts later support.
Pricing structure, not headline price
Do not compare percentages. Build one model using your own numbers and put every provider through it: monthly volume, average order value, method mix, refund rate, dispute rate, settlement currency and payout frequency. Then ask for a complete fee list covering transaction fees, per-transaction fixed fees, cross-border and international card surcharges, currency conversion, refund and dispute handling, payout fees, monthly minimums, and any reserve. Public pricing is only a first filter. Actual pricing changes with country, industry, volume, settlement currency, integration method and contract terms. Confirm against the provider's official pricing page or a written quote before you go live. Then calculate two outputs: total monthly cost, and net receipts, which is the amount that actually reaches your bank. After your first full month, reconcile the model against the real bill and query any gap before you scale volume.
- Model total monthly cost and net receipts using your own volume and method mix.
- Ask for every fee category explicitly, including minimums, payouts and dispute handling.
- Get a written quote rather than relying on the public pricing page.
- Reconcile the model against your first full invoice and query differences before scaling.
Settlement, support and exit terms
The last three criteria are the ones merchants skip and later regret. For settlement, ask the settlement cycle in days, the daily cut-off and timezone, whether days are business or calendar, the payout currency, and any reserve or minimum payout threshold. For support, test it before you sign: send a technical question through the real support channel and time the response. Ask whether you get a named contact, what the escalation path is, and whether support covers your timezone. For exit terms, read them before you need them: notice period, whether you can export your transaction history and customer tokens, how long balances are held after the final transaction, and whether there is any cost to leave. A provider that is easy to leave is usually also one that has to keep earning your business, which is a better position to be in than one with a lock-in.
- Ask for the settlement cycle, cut-off time, timezone, payout currency, reserve and payout threshold.
- Test support response times through the real channel before you commit to a provider.
- Read the exit terms: notice period, data and token export, and post-termination balance hold.
- Prefer terms you can leave, because portability keeps a provider competing for your business.
Questions merchants ask
What should I compare when choosing a payment provider?
Coverage of the methods your customers actually use, eligibility to board your entity, onboarding time, the full pricing structure, settlement timing and currency, support quality, and exit terms. Score each against your own numbers rather than comparing headline percentages, and set pass or fail gates for category acceptance. Keep your own records of orders and settlements so the comparison never depends on a provider dashboard.
Is the cheapest payment provider the best choice?
Rarely. A headline percentage excludes fixed per-transaction fees, cross-border surcharges, currency conversion, refund and dispute costs, monthly minimums and reserve terms. Model the total monthly cost and the amount that actually reaches your bank account, then weigh that against method coverage, settlement speed, support and how easy the provider is to leave.
Should I use a global PSP or a regional payment aggregator?
It depends on where you sell. Global providers suit card-heavy developed markets, while regional specialists often hold the local licences and relationships needed for instant bank transfers, wallets and cash collection in emerging markets. Many merchants use both, one per region, and keep a fallback. If you sell into several regions, running both is common and gives you a fallback when one has an outage.
How long does it take to switch payment providers?
Long enough that you should read the exit terms before you sign. Plan for onboarding and integration time with the new provider, migration of stored credentials if you have subscriptions, and the period your old provider holds final balances. Keep your own records so a switch does not depend on their reporting.
Can I use two payment providers at once?
Yes, and many merchants should. A second provider gives you a fallback if one has an outage, leverage in pricing conversations, and a route to methods the other lacks. The cost is extra integration and reconciliation work, so keep your order and settlement data in your own systems rather than relying on a provider dashboard.
What is the biggest mistake merchants make when comparing providers?
Comparing a marketing coverage list instead of confirmed eligibility. A provider may show a method on its website and still be unable to board you for it, because eligibility depends on your registration country, category and who settles to you. Ask for method-by-method confirmation in writing before you build anything. Confirm eligibility first, then compare everything else, because nothing else matters if you cannot be boarded.