Smart Routing
What Is Smart Routing in Payments? Definition and How It Works
Definition
Smart routing is the automated process of selecting the optimal acquirer, processor, or payment path for each transaction in real time, using rule-based logic or machine learning to maximise approval rates and minimise processing costs.
How it works
Smart routing operates within a payment orchestration layer or switch. When a transaction is initiated, the routing engine evaluates a set of parameters, card type, BIN range, transaction amount, currency, merchant category, and geographic origin, against a configured rule set to select the best available processing path. The decision is made before the authorisation request is dispatched.
Rule-based routing uses static or semi-static logic: conditions and priority rankings set by the merchant or their payment provider. Examples include routing all Amex transactions to Acquirer A, routing cross-border transactions over a threshold amount to Acquirer B, and falling back to Acquirer C if real-time approval rate data shows Acquirer A performing below a set threshold.
ML-assisted routing adds a dynamic layer on top of rule-based logic. A model trained on historical authorisation data makes probabilistic predictions about which routing path is most likely to produce an approval for a given transaction profile. The model updates as new transaction data is processed, adapting to shifts in acquirer performance, card network behaviour, and fraud patterns.
Routing decisions can also incorporate cost optimisation: routing the same transaction to the acquirer with the lowest effective interchange or scheme fee for that specific card type and amount. Cost-based routing is typically applied within the constraints of approval rate thresholds, least cost routing that sacrifices approval rate typically produces a net-negative outcome.
Why it matters
Approval rate improvement is the primary routing benefit: routing each transaction to the acquirer statistically most likely to approve it recovers revenue that static single-acquirer setups leave on the table. Gains of 1-3 percentage points are achievable on cross-border and high-value transaction segments where acquirer performance varies most.
Cost reduction is achievable without sacrificing approval rate: routing to lower-cost acquirers for transaction types where multiple acquirers have equivalent approval rates reduces effective MDR. This requires granular cost and approval rate data per transaction segment.
BIN-level routing precision matters: routing based on card network alone is a blunt instrument. Smart routing that operates at the BIN level, identifying specific issuing banks, card programs, and product types, can fine-tune approval rate performance significantly versus scheme-level routing rules.
Routing rules require maintenance: acquirer performance changes over time as networks evolve, fraud patterns shift, and acquirer infrastructure changes. Static routing rules that were optimal at configuration time degrade without regular review and updating.
With PXP
PXP's smart routing engine evaluates approval rate history, cost, card type, and BIN data in real time for every transaction. Routing rules are configurable in the PXP dashboard and can be updated without a code change. ML-assisted routing is available for merchants with sufficient transaction volume to support model training.
Frequently asked questions
What's the difference between smart routing and payment orchestration?
Payment orchestration is the broader capability: it encompasses acquirer connectivity, transaction lifecycle management, retry logic, reporting, and routing. Smart routing is the specific decision engine within orchestration that selects which acquirer or processor handles each transaction. You can have orchestration without sophisticated routing logic; smart routing is a subset of a full orchestration capability.
How are routing rules configured?
Routing rules are typically configured in the payment provider's dashboard or via API. Rules can be as simple as card-type-to-acquirer mappings or as complex as conditional logic based on transaction amount, currency, geographic origin, and real-time acquirer approval rate. Enterprise merchants often work with their payment provider to design and test routing rule sets before deploying to production.
Can smart routing reduce processing costs without hurting approval rates?
Yes, through least-cost routing applied within approval rate constraints. For transaction segments where multiple acquirers show equivalent approval rates, routing can be directed to the lower-cost option without a conversion trade-off. The key is having per-segment approval rate data at sufficient granularity, cost routing applied across a blended pool of transactions without segment-level data tends to degrade overall approval rates.
How does ML-assisted routing differ from rule-based routing?
Rule-based routing uses manually configured conditions and priorities that remain static until changed. ML-assisted routing uses a model trained on historical transaction outcomes to make probabilistic routing decisions dynamically. ML routing adapts as acquirer performance changes without requiring manual rule updates, but requires sufficient transaction volume to train effectively. Most enterprise deployments use both in combination.
Revolutionize your business with PXP
Take complete control of your commerce and payments with one platform.
Get Started