Higher authorization rates, more revenue
What switching moves, on average
- +18%Approval-rate uplift with smart routing
- 12.4%Lift in failed-payment recovery
- 22%Reduction in processing fees
- <50msFailover when a processor declines or dies
The stakes
Why authorization rates matter more than you think
Every declined payment is lost revenue, not just a failed transaction
A decline isn’t neutral. For a one-time purchase it’s a lost sale; for a subscription it’s the first step toward involuntary churn, and the customer usually never finds out it happened. At 100,000 transactions a month, the difference between an 88% and a 93% authorization rate is five thousand payments, every month, forever.
Most businesses don’t know their true authorization rate by PSP
Ask a finance team for the blended rate and someone can usually find it. Ask for the rate on European debit cards through their secondary processor and the room goes quiet. Without per-processor, per-segment numbers, you can’t tell whether a dip is your fraud rules, your processor, or just seasonality.
A single-processor setup has a ceiling you can’t optimize past
Once you’ve cleaned up your fraud rules and your retry logic, a single processor gives you nothing left to pull. Its issuer relationships, its regional coverage, and its risk appetite set your ceiling. The only way past it is a second processor and somewhere intelligent to route between them.
Diagnosis
What causes low authorization rates
Three culprits show up in almost every account we look at.
Routing all transactions through one processor
No processor is best at everything. One is strong on US credit, weak on European debit; another is the reverse. Forcing every transaction down one pipe means accepting that processor’s worst segments along with its best.
Fraud rules that are too aggressive, blocking good customers
False declines are quiet. The rule blocks the charge, no fraud occurs, and the dashboard looks safe, while a paying customer bounces off checkout. Studies keep finding merchants lose more to false declines than to the fraud the rules were written for.
No visibility into which PSP is underperforming and why
When performance data lives in each processor’s own dashboard, in each processor’s own format, nobody compares them. Underperformance goes unnoticed for quarters because there’s no side-by-side view that would make it obvious in a minute.
Four levers, one platform
How PaymentKit improves authorization rates
Smart routing across multiple processors
Every transaction goes to the processor with the best approval record for that card type, currency, and region. Soft declines cascade to a second processor instead of hitting a dead end. Routing rules are yours to set; the defaults learn from your own traffic.
Configurable fraud prevention rules
One set of fraud rules that applies consistently across every processor, tuned by you. Tighten screening on segments where fraud actually happens, relax it where it doesn’t, and stop paying for safety you don’t need with false declines.
Adaptive retries for failed transactions
Declines that can be saved, get saved. Retry timing adapts to the failure code and the billing cycle, so an insufficient-funds decline waits for a smarter moment while a do-not-honor tries a different route entirely.
Network tokenization to reduce declines
Network tokens replace raw card numbers with credentials the card networks keep current. When a customer’s card is reissued or expires, the token updates behind the scenes: the renewal goes through and nobody has to type in a new number.
Visibility
See exactly where you’re losing approvals
Optimization without measurement is guesswork. These three views live in Revenue Metrics.
Payment intent success rate
Of everything you attempted to charge, how much got through, tracked over time so a change in fraud rules or routing shows up as a visible step in the line, not an anecdote.
Volume success tracking
The same question in dollars. Attempted volume against successful volume tells you what declines actually cost last month, which is the number that gets an optimization project funded.
PSP performance benchmarking, side by side
Every processor’s authorization rate in one table, split by card type and region. This is where underperformance stops hiding, and the evidence you bring to your next rate negotiation.
Single-processor setup vs. authorization optimization
One pipe, one risk appetite, one ceiling. Or multiple processors, one intelligent layer on top
- Approval oddsSingle processorEvery transaction takes that processor’s approval odds, strong segment or weak.PaymentKit orchestrationEach transaction routes to the processor with the best record for that exact profile.
- Declined chargesSingle processorA declined charge is retried on the same rails that just declined it, or not at all.PaymentKit orchestrationSoft declines retry on a second processor; recoverable revenue gets recovered.
- OutagesSingle processorAn outage means revenue stops until the processor fixes it.PaymentKit orchestrationFail over is automatic: an outage becomes a routing event, not an incident.
- Performance dataSingle processorPerformance data lives in the processor’s dashboard, graded by the processor.PaymentKit orchestrationIndependent, side-by-side benchmarking of every PSP you run.
- Your ceilingSingle processorYour ceiling is set by someone else’s issuer relationships.PaymentKit orchestrationThe ceiling moves: add a processor where your current mix is weakest.
PaymentKit vs. billing-only platforms
The numbers Chargebee and Recurly can't show you
Both are capable billing layers, and both ride on whatever your gateway approves. There's no routing layer underneath them, so there's no lift to report.
- Authorization liftChargebee / RecurlyYour gateway's approval rate is your approval rate. Billing rides on top; nothing lifts it.PaymentKit+18% average approval-rate uplift from smart routing and cascading across processors.
- Failed-payment recoveryChargebee / RecurlyRetry schedules re-run the charge on the same gateway that just declined it.PaymentKitAdaptive retries across processors lift failed-payment recovery by 12.4% on average.
- Processing feesChargebee / RecurlyOne gateway's pricing. The invoice is what it is.PaymentKitCost-aware routing cuts processing fees by 22% on average.
- Gateway outageChargebee / RecurlyBilling waits until the gateway comes back.PaymentKitFailover in under 50ms: renewals land on a healthy processor instead of failing.
Best fit
When authorization rate optimization makes the biggest difference
High-volume subscription businesses with recurring billing
Recurring charges hit the same optimization surface every cycle, so every point of improvement repeats monthly. This is also where network tokenization earns its keep: card churn is constant at volume, and tokens absorb most of it silently.
Teams processing across multiple geographies
Cross-border is where single-processor setups leak the most. Local acquiring through the right processor per region routinely beats one global pipe on both approval rates and fees. Routing by geography is often the single biggest lever available.
Businesses that have hit a ceiling on a single PSP
If you’ve already tuned retries and fraud rules and the number won’t move, you’re at the structural limit. Adding a second processor and routing between them is what moves it. Teams in this position tend to see the fastest, clearest lift.
