Every small business that takes money online or in person pays a payment processor, and over a few years those fees become one of the largest line items the business is barely tracking. A 1% difference in processing fees on a business doing £200,000 a year is £2,000 — a real, recurring cost that compounds annually and is invisible because it leaks out a few pounds at a time on every transaction. Most small business owners choose their payment processor early in the business’s life, often by accepting whatever was offered with their booking system or e-commerce platform, and then never revisit the choice for years. The cumulative cost of inertia in this decision is meaningful, and the comparison work — once you understand the categories — takes a few hours rather than the project it appears to be.

This guide covers the main payment processors available to small businesses, the structures of their fees, the categories in which each is genuinely strong, and the considerations that should shape the choice for businesses operating in different markets and with different transaction profiles. The geographic dimension matters because processor availability and pricing vary significantly between markets — the right choice for a UK service business is often wrong for a Kenyan retailer, and vice versa. The transactional dimension matters because the optimal processor for high-volume small transactions is rarely the same as for low-volume large ones. Both are addressed below; the goal is to give you enough understanding to choose well rather than to pick a single answer that fits no business specifically.

§ 01Understanding the Fee Structure Before the Brand

The first step in evaluating any payment processor is understanding what the fees actually are, because the headline number on a processor’s website is rarely the number you actually pay. The headline fee — say, “2.9% + 30 cents per transaction” — is the base online card-present fee, and a meaningful percentage of your transactions will not be charged at this rate. International cards typically incur an additional 1-1.5%; American Express often costs more than Visa or Mastercard; manually entered cards cost more than dipped or tapped ones; recurring subscriptions may be priced differently from one-off transactions. The blended rate you actually pay is almost always higher than the headline rate, sometimes substantially.

The second part of the fee structure that surprises people is the supplementary costs. Chargebacks — disputed transactions that the processor refunds to the customer — typically cost £15-25 per dispute regardless of outcome, and even legitimate chargebacks against you accumulate. Currency conversion, when you accept payments from international customers, adds 1-3% on top of the base fee. Payouts to your bank account may incur fees on instant or expedited options. International transfers, if you sell across markets, add their own layer. None of these is large individually; in aggregate they can shift your effective rate by half a percentage point or more, which compounds meaningfully on annual volume.

The third dimension is contract structure. Most modern processors operate on month-to-month terms with no long-term contracts and clear, published pricing — Stripe, Square, Adyen, and most newer entrants fit this model. Older merchant-services providers, often sold by banks or independent agents, may push three- or five-year contracts with cancellation fees, opaque pricing tiers, and rates that mysteriously increase after the introductory period. The structural answer for most small businesses is to avoid the long-contract providers entirely; the modern direct-to-business processors offer better pricing, better software, and the operational freedom to switch if their service deteriorates. The ten-minute test is whether you can sign up, take a payment, and leave again with no penalties; if you can, the processor is structurally aligned with your interests.

§ 02Stripe — The Default for Online-First Businesses

Stripe has become the default payment processor for online-first businesses globally, and for businesses whose primary sales channel is a website, the case for choosing Stripe is strong enough that the burden of proof falls on the alternatives. The base pricing — currently around 2.9% + 30 cents on standard online card transactions in most markets, with regional variation — is competitive without being the absolute lowest. The strength of Stripe is in everything that surrounds the base transaction: developer-friendly APIs that integrate cleanly with almost every modern e-commerce and booking platform, sophisticated tools for subscriptions and recurring billing, strong fraud-prevention features, and a reputation for reliability that matters when your business depends on payments going through.

The areas where Stripe is genuinely strong are subscription businesses (Stripe Billing handles complex recurring scenarios that legacy processors struggle with), platforms that take payments on behalf of others (Stripe Connect handles the marketplace structure cleanly), and any business with international customers (Stripe handles 135+ currencies and most local payment methods natively). The integration ecosystem is the deepest in the industry — virtually every modern e-commerce, booking, and CRM platform supports Stripe natively, often with deeper feature integration than any other processor. For online businesses, this combination of pricing, reliability, and integration depth is the strongest case in the category.

The areas where Stripe is weaker include in-person payments (Stripe Terminal exists and is improving, but Square remains structurally better for businesses where most transactions happen at a physical till), some emerging markets where local payment methods dominate (M-Pesa in Kenya, mobile-money systems in West Africa, certain QR-code systems in parts of Asia), and the pricing for very high-volume businesses (where direct merchant-services contracts can produce lower rates if you’re willing to negotiate and accept longer terms). For most small businesses doing online sales, none of these weaknesses applies enough to dethrone Stripe as the default; for the businesses where they do apply, alternatives become genuinely important.

§ 03Square — When Most Transactions Happen In Person

Square’s market position is the inverse of Stripe’s: it began as a card reader for in-person transactions and expanded into online and other channels later, with the result that for businesses where most transactions happen at a physical till, Square is structurally well-suited in ways that online-first processors aren’t. The hardware ecosystem — card readers, point-of-sale terminals, kitchen displays, kiosk hardware — is mature and integrates with the same back-end as the online sales, which means a single dashboard covers in-person, online, and invoiced revenue. For small retailers, restaurants, food trucks, salons, and other in-person businesses, this consolidation is genuinely useful.

Square’s pricing on in-person transactions (currently around 2.6% + 10 cents in the US, similar in the UK and Australia) is competitive and often better than online rates because card-present transactions carry lower fraud risk. The point-of-sale software, which Square gives away free with the processing service, is genuinely usable for small retail and food-service operations and includes inventory, basic CRM, and reporting that would otherwise require separate tools. Square Appointments integrates booking and payment for service businesses; Square for Restaurants extends to table management. The all-in-one structure works well for businesses that don’t want to assemble a stack from separate vendors.

Square’s weaknesses appear when business operations grow beyond the small-merchant template the platform was designed for. Subscription billing is functional but not its strength; complex e-commerce with international customers and varied product configurations is better served by Stripe-based stacks; integration with sophisticated CRM and marketing tools is shallower than competitors. The pricing on online transactions specifically is similar to Stripe but the surrounding feature depth is weaker, so for businesses where online dominates, Stripe usually wins on capability. The cleanest test is the proportion of in-person to online: above 70% in-person, Square; below 70% online, Stripe; in between, the choice depends on which feature set matters more to your specific operations.

§ 04PayPal — Familiar But Expensive

PayPal occupies an awkward position in the modern payments landscape. Its strengths are name recognition (a meaningful percentage of online buyers will pay through PayPal who wouldn’t enter their card details on an unfamiliar site), broad availability across markets, and easy buyer-protection mechanisms that reassure cautious customers. Its weaknesses are substantial: pricing is consistently higher than Stripe for most transaction types, account holds and disputes are notoriously friction-heavy, customer-service quality is widely reported as poor, and the integration depth in modern platforms has not kept pace with newer processors.

The right framing for PayPal in 2026 is as a secondary payment option offered alongside a primary processor, rather than as the primary processor itself. The marginal customers who specifically prefer PayPal are real and worth capturing, but routing all transactions through it sacrifices margin to capture that small minority. The standard configuration for most small online businesses is Stripe or Square as the primary processor handling card transactions, with PayPal as an additional checkbox at checkout for customers who insist on it. This approach captures the brand-recognition value of PayPal without paying its higher fees on the bulk of transactions.

The exception is businesses where PayPal genuinely dominates the customer base — certain creator economies, some categories of cross-border B2C commerce, and specific niches where PayPal’s buyer protection is part of the customer expectation. For these businesses, PayPal as primary may be unavoidable. Even then, the operational discipline of monitoring chargebacks aggressively, maintaining detailed transaction records, and being prepared for occasional account holds is worth setting up early. PayPal works; it costs more than alternatives and creates more operational friction than alternatives; choose accordingly.

§ 05M-Pesa, Flutterwave, Paystack — The African Reality

For businesses operating in Kenya, Nigeria, Ghana, South Africa, and other African markets, the dominant payment infrastructure is structurally different from the card-network world that Stripe and Square operate in. M-Pesa in Kenya processes more transactions than the entire Kenyan banking system combined, and any business in Kenya not accepting M-Pesa is effectively excluding most of its potential customer base. The integration of M-Pesa into a website is handled either through Safaricom’s direct APIs (powerful but technical) or through aggregators like Flutterwave, Paystack, or DPO that wrap M-Pesa alongside card processing in a single integration.

Flutterwave is the dominant pan-African payment aggregator, with operations across most major African markets and integrations that include M-Pesa, MTN Mobile Money, Airtel Money, card networks, bank transfers, and emerging crypto rails. Its pricing varies by market and transaction type but is generally competitive, and the single integration covering multiple payment methods is operationally valuable for businesses serving multiple African markets. Paystack, now owned by Stripe, is similarly positioned with stronger Nigerian roots, excellent developer experience, and pricing that has stayed competitive through the acquisition. Both handle the operational reality of African e-commerce in ways that Western-first processors don’t.

The choice between them depends on the specific markets you’re serving. For Kenya-primary operations, the local depth of integration and customer trust around DPO Pay or direct Safaricom APIs may be worth it. For Nigeria-primary operations, Paystack’s depth is hard to beat. For multi-country operations across the continent, Flutterwave’s coverage and single-integration model is structurally attractive. The Western processors are increasingly available in African markets — Stripe operates in Kenya, Nigeria, and South Africa — but the local players still typically offer better customer-facing experiences for African buyers and lower friction on local payment methods. For African small businesses, the question is rarely whether to use a local-first processor; it’s which one.

§ 06Adyen and the Enterprise Tier

For small businesses growing into mid-sized ones, the question of when to graduate from Stripe or Square to a more enterprise-grade processor like Adyen, Worldpay, or Checkout.com eventually becomes relevant. The pricing logic shifts at higher volumes — somewhere around £1-2 million a year in transaction volume, the negotiated rates available from enterprise processors become meaningfully lower than the published rates of the small-business processors, and the saved fees start to justify the operational complexity of more sophisticated processing infrastructure. Below this threshold, the operational simplicity of Stripe or Square almost always wins; above it, the math may favour the move.

Adyen has built its reputation on serving large international businesses with complex needs — multi-currency, multi-region, multi-channel — and offers genuinely sophisticated tools for businesses operating at that scale. Its small-business tier exists but rarely competes effectively with Stripe at the lower volumes; its strength is in the £5+ million volume range. Checkout.com sits in similar territory, with strong European roots and enterprise-grade reliability. Worldpay is the legacy player with deep merchant-services experience, broad market coverage, and pricing that requires negotiation but can be aggressive at volume.

The mistake at this scale is over-anchoring on fee percentage. The differences between processors at higher volumes are often in operational quality — fraud-prevention sophistication, dispute-handling capability, reporting depth, and integration with the back-office tools larger businesses use — and these qualities can move the bottom line more than headline rates. Negotiate the rate, but choose primarily on operational fit; the difference between a processor that handles disputes cleanly and one that doesn’t is worth more than 0.1% on rate. Most small businesses will never reach this tier, and that’s fine; the small-business processors handle most operational reality competently. For the businesses that do grow into it, the move is genuinely worth the work.

§ 07Recurring Billing and Subscription Logic

For businesses with subscription, membership, or recurring-billing models, the payment processor decision is shaped by a different set of considerations. The base transaction rate matters less; what matters more is the sophistication of the recurring-billing logic — handling failed payments gracefully, managing card updates when expiry dates change, supporting upgrades and downgrades cleanly, prorating mid-cycle changes correctly, and handling cancellations without revenue leakage. Stripe Billing has built the most mature feature set in this category and is the dominant choice for SaaS, subscription boxes, membership organisations, and anything else with recurring revenue.

The “dunning” workflow — what happens when a recurring payment fails — is where subscription processors differentiate sharply. A failed payment that simply cancels the subscription loses the customer; a failed payment that triggers a sensible retry sequence (typically 3-4 retries over 2-3 weeks, with email notifications to the customer at appropriate points), updates the card if a new one is on file, and only cancels after multiple genuine failures recovers a meaningful percentage of revenue that would otherwise be lost. Mature subscription processors handle this automatically; less sophisticated tools require you to build it yourself or accept the revenue loss. Over a year, the difference between good and bad dunning is typically 2-5% of subscription revenue.

Other tools worth knowing in this category include Chargebee and Recurly, which sit on top of Stripe (or other processors) and add subscription-management sophistication beyond what the base processor handles — usage-based billing, complex pricing tiers, sales-tax handling across jurisdictions, revenue recognition for accounting. For small subscription businesses, Stripe Billing alone is usually enough; for more complex pricing structures or higher volumes, the addition of Chargebee or Recurly produces operational leverage that justifies the additional cost. The decision is similar to the booking-system question: start simple, upgrade when limitations become genuinely binding.

§ 08Chargebacks, Fraud, and the Boring Operational Reality

The unglamorous reality of payment processing is that chargebacks and fraud are continuous rather than occasional, and the processor’s tools for handling both meaningfully affect the operational cost of the payments side of your business. A chargeback rate above 1% — a common threshold across processors — triggers monitoring programmes that can result in higher fees, account holds, or eventual termination. Most small businesses operate well below this threshold without thinking about it, but specific industries (digital goods, high-value services, anything with delivery delays) drift higher and need active management.

The defences against chargebacks are operational and require no special technology. Clear product descriptions and policies displayed prominently on the website reduce disputes about what was delivered. Tracked shipping with delivery confirmation reduces “I never received it” disputes. Clear refund policies that make legitimate refunds easy reduce the friction that pushes some customers to chargebacks instead. Detailed transaction records make winning the disputes that do occur much more likely. The processor provides the workflow for fighting disputes; the operational discipline comes from your business. Spend the time on the operations rather than hoping the processor’s algorithms will handle everything.

Fraud detection has become substantially better across all major processors over the last decade, and the false-positive rates that used to plague small businesses (legitimate transactions blocked by overzealous fraud rules) have dropped meaningfully. Stripe Radar, Square’s Fraud Suite, and Adyen’s RevenueProtect are all genuinely sophisticated and require minimal configuration to handle most small-business fraud scenarios competently. The remaining vigilance — flagging unusually large transactions for manual review, watching for patterns of failed-then-successful card attempts, treating international cards on local-only businesses with appropriate scepticism — is straightforward operational discipline that any owner can maintain. Trust the processor’s tools; supplement with your own attention to anomalies.

§ 09How the AI Builder Wires Payments In

The reason payment processing becomes a problem for many small business websites is that the integration between the website, the payment processor, and the rest of the business systems (CRM, email, accounting, fulfillment) is rarely seamless when assembled from generic parts. Each integration point is its own configuration project, and the cumulative time cost of getting them right often exceeds the actual launch budget for the website. A purpose-built AI website builder for small business owners generates the website with payment processing integrated as a first-class element — Stripe and Square integrations pre-configured, the relevant African processors available where the business is regionally appropriate, and the connections to surrounding tools already in place.

The integration covers the operational realities most small businesses encounter: standard one-off purchases, subscription billing through Stripe Billing, deposits and partial payments for service bookings, M-Pesa integration where the business is Kenya-based, and the handling of international currency where customers cross borders. The platform-specific edge cases — Stripe’s required webhook configuration, Square’s OAuth handshake, Flutterwave’s redirect flow — are handled in ways that don’t require developer involvement to set up properly. You configure the processor account once; the website handles everything downstream.

The economic case is consistent with the rest of the series. Setting up clean payment infrastructure on a custom-built site typically costs thousands in developer time before any actual revenue is collected — particularly when the requirements include subscription billing, multiple processors, or African payment methods. The same infrastructure as part of the standard small business website builder with payments integration at $12.50 a month brings professional-grade payment operations within reach of any small business. Choose the right processor for your business model and geography; let the website handle the integration; focus your energy on the revenue the payments represent rather than on the infrastructure that processes it.