Ecommerce Payment Routing: How to Route Transactions Across Providers
How ecommerce payment routing works: routing inputs, rule design, local acquiring, cost and approval trade-offs, cascading retries, failover, network rules and how to measure each route.
Quick answer
Payment routing decides which provider or acquirer processes each transaction. Rules use card data (issuer country, brand, debit or credit), transaction data (currency, amount, market, customer- or merchant-initiated) and provider data (fees, measured approval rates, health, capabilities). Start with simple rules such as local acquiring by issuer country, add failover for provider outages and cascade only soft declines within network rules. Measure approval rate, cost and fraud per route against a control, because routing that looks cheaper can lose more in declined sales.
Where This Fits
Routing is one part of payment orchestration, which provides the multi-provider layer routing depends on. Decline handling is in payment failure handling, and how authentication interacts with routing is in 3D Secure in ecommerce.
What Can Be Routed?
Routing choices exist at several levels. You might choose between payment providers, between acquirers behind one provider, between merchant accounts in different countries, or, for some debit cards, between networks. Each choice is only available if your contracts and integrations support it, so map what is actually possible before designing rules.
Routing Inputs
| Input | Example values | Why it matters |
|---|---|---|
| Issuer country (from BIN) | US, GB, DE, IN | Local acquiring and cross-border fees |
| Card brand and type | Visa credit, Mastercard debit | Pricing, network options |
| Currency and amount | EUR 85, USD 1,200 | Provider support, risk, fees |
| Customer- or merchant-initiated | Checkout vs renewal | Agreement references and SCA rules |
| Payment method | Card, wallet, bank transfer | Only some providers support each |
| Provider health | Error rate, latency | Failover decisions |
| Measured approval rate | By route and segment | Performance-based routing |
Designing Routing Rules
Write rules as an ordered list of conditions and targets, evaluated top to bottom, with a default route at the end. Keep them in configuration rather than code so payments and finance teams can review changes, and version every change so you can correlate it with performance.
rules:
- when: { initiated_by: merchant }
route: original_agreement_provider # renewals stay put
- when: { method: local_bank_transfer, market: NL }
route: provider_b
- when: { issuer_country: US, currency: USD }
route: us_acquirer
- when: { issuer_country: [GB, IE] }
route: uk_acquirer
- default: provider_a
failover:
provider_a: provider_b
us_acquirer: provider_aLocal Acquiring
The most common routing gain is processing cards through an acquirer in the issuer's country. Issuers can be more willing to approve domestic transactions, and cross-border fees may be lower. The effect varies a great deal by market, so treat it as a hypothesis to test rather than a guaranteed improvement.
Cost vs Approval Rate
Routing purely on fees is a common mistake. A route that costs slightly less but approves fewer payments loses far more in abandoned orders than it saves. Compare routes on cost per successful payment and revenue per attempt, not on the headline fee.
Want routing decisions based on your own payment data?
ZSpace Labs can set up route-level reporting and configurable rules so payment changes are measured rather than guessed.
Cascading Retries
Cascading sends a declined payment to another provider. It can recover some soft declines caused by a specific acquirer or temporary issues. It must not be used for hard declines (lost or stolen card, closed account, do-not-honour codes that indicate fraud), and card networks restrict excessive retries of declined transactions. Limit cascades to one alternative route, only for decline codes you have classified as retryable, with the same idempotency protections as any other payment call.
Remember the shopper. A cascade that requires a second 3DS challenge or adds several seconds of delay may cost more than it recovers. See payment failure handling for decline classification.
Failover for Provider Outages
Failover is routing by provider health. Track error rates, timeouts and latency per provider, and switch traffic when they cross thresholds. Use a circuit-breaker pattern: stop sending traffic to a failing provider, test it periodically and return gradually. Failover only helps if the alternative provider supports the same payment methods and has saved credentials it can use, which is why token portability matters.
Recurring and Merchant-Initiated Payments
Renewals and other merchant-initiated transactions reference the original customer-initiated agreement, including the network transaction identifier. Moving them to a different provider can lose that reference and lower approval rates. Keep them on the original route unless your providers and tokens explicitly support migration. See ecommerce recurring payments.
Measuring Routing Performance
- Approval rate by route, issuer country, card type and amount band
- Cost per successful transaction, including cross-border and scheme fees
- 3DS challenge and success rates by route
- Latency and error rates
- Fraud and dispute rates by route
- A control group or holdout so changes are compared fairly
Rules-Based vs Performance-Based Routing
There are two broad styles. Rules-based routing uses fixed conditions (issuer country, currency, method). Performance-based routing, often marketed as smart routing, adjusts choices using recent approval rates, costs and provider health, sometimes with machine learning.
| Rules-based | Performance-based | |
|---|---|---|
| How decisions are made | Ordered conditions set by your team | Recent outcome data, sometimes a model |
| Transparency | Easy to read and audit | Harder to explain individual decisions |
| Adapts to change | Only when someone edits rules | Automatically, within limits |
| Data needed | Little | Enough volume per segment to be meaningful |
| Best for | Most merchants, especially early | High-volume merchants with several providers |
Routing Advantages and Limitations
Routing can lift approvals in specific segments, lower costs and keep payments flowing during outages. It cannot fix poor data quality, fraud problems or checkout friction, and every extra route adds operational work: more settlement reports, more dispute portals and more contracts. Gains are often concentrated in a few segments (certain issuer countries or card types), so look for those rather than expecting a uniform improvement.
How to Implement Routing Step by Step
- 1. Inventory possible routes: providers, acquirers and merchant accounts you can actually use
- 2. Segment your data: approval rate and cost by issuer country, brand, type and amount
- 3. Write eligibility rules so unsupported methods and currencies never reach a route
- 4. Add one targeted rule, such as local acquiring for your largest foreign market
- 5. Run it against a holdout and measure approval, cost and fraud
- 6. Add failover with health thresholds
- 7. Classify decline codes before enabling any cascade; see failure handling
- 8. Keep renewals on their original route; see recurring payments
- 9. Review monthly and retire rules that do not pay their way
Worked Example
An illustrative scenario, not a client case: a UK electronics retailer adds a second provider and initially routes by lowest fee. Approval rates on high-value orders drop, and revenue per checkout falls despite lower fees. The team switches to routing by issuer country with a performance check, keeps high-value orders on the better-performing route and introduces a 10% holdout to measure future changes.
Common Mistakes
- Routing on fees alone
- Cascading hard declines
- Moving merchant-initiated renewals to providers without the original agreement
- Rules hidden in code with no change history
- No holdout, so improvements cannot be proven
- Failover to a provider that cannot use saved cards
Need routing that is measurable and safe?
Talk to ZSpace Labs about payment routing and orchestration development and payment analytics automation.
Conclusion
Good routing is simple rules, measured carefully. Start with eligibility and local acquiring, add failover, cascade only soft declines and judge every change by approval rate and revenue, not fees. Related: payment orchestration and recurring payments.
Common questions
Choosing which payment provider, acquirer or network path processes each transaction, based on rules about the card, the transaction and provider performance.