Do You Need a Payment Orchestration Platform or Just a Second Gateway?
For most merchants, a second acquirer fixes outage and coverage problems for far less money. These are the volume, market, payment-method and approval-rate thresholds where orchestration starts to earn its fee.
Redundancy or routing complexity: which problem you actually have
The question usually comes up after something breaks. Your only processor goes down on a peak trading day, or approval rates in one market slip and nobody can say why. Then two kinds of vendor get in touch. Orchestration vendors offer one layer over many processors. Acquirers offer to be your backup. Which one you need depends on whether your problem is redundancy or routing complexity.
A payment orchestration platform sits between your checkout and your processors. You get one integration, one token vault and one rules engine, connected to many acquirers, payment service providers (PSPs), fraud tools and local payment methods. A second acquirer is simpler. You keep your main provider and integrate one more directly, so traffic has somewhere to go when the first one fails or underperforms.
If what you need is a fallback when your main processor goes down, a direct second acquirer is usually cheaper and faster. Orchestration starts to pay for itself when you run many processors across many markets, each with its own payment methods, approval patterns and data rules.
What a direct second acquirer covers, and the stored-credential gap
If you process in one to three markets, a second acquirer covers most of the risk. You get failover during outages, a second route for borderline authorizations, and leverage on fees at renewal. Global acquirers and PSPs such as Adyen, Checkout.com, Worldpay, Stripe and Nuvei will compete for secondary volume, and so will the large bank-owned acquirers in each region.
The gap most teams find late is stored credentials. If your main processor tokenizes your customers' cards, those tokens generally work only with that processor. During an outage the second acquirer can take newly entered cards, but it can't charge returning customers, subscriptions or one-click checkouts. For a subscription business, that can be most of its volume.
You can close the gap without buying full orchestration:
- Network tokens. Visa, Mastercard and other schemes issue tokens that can move between acquirers, but only if you, or a neutral provider acting for you, are the token requestor. Tokens your processor requested in its own name may not move. Ask both acquirers exactly how they handle tokens provisioned by the other.
- A standalone third-party vault. Vendors such as VGS and Basis Theory store card data for you and forward it to whichever processor you choose. Some orchestration vendors, Spreedly among them, can also be used mainly for their vault. This setup also cuts down how much of your own environment falls under PCI DSS.
For most mid-market merchants, a second acquirer with portable tokens solves the outage problem. Your engineers build retry and failover logic once, and after that they maintain just two integrations.
Where orchestration earns its place beyond two processors
Orchestration pays off once every new market, processor or payment method means another integration project. It adds value in four places.
Local payment methods without a separate build for each
Brazil means Pix and boleto. India means UPI and RuPay. Add iDEAL in the Netherlands (now moving toward Wero), BLIK in Poland, M-Pesa in Kenya, GCash in the Philippines and konbini payments in Japan. A good orchestration layer gives you these as pre-built connectors. Without one, each method is its own build, and you have to update it every time the provider changes its API.
Routing by rule and by data
Routing means picking a processor for each transaction based on card BIN, issuing country, currency, amount, cost or recent approval performance. With two processors, a few if-then rules in your own code will do. With five processors across eight markets, you need a real rules engine, A/B testing, and reporting that compares approval rates across processors on the same basis.
Local acquiring and regional compliance
Issuers generally approve more when the acquirer is in the same country, and some markets practically require local processing. Brazil, Mexico, India and Indonesia all reward a local acquiring setup. In India, RBI tokenization rules limit who can store card data. Data-residency rules also differ by region: GDPR in the EU and UK, LGPD in Brazil, and localization requirements in several Middle Eastern and Asian markets. A platform can help you meet these rules, but buying one doesn't make you compliant. Confirm where data is stored and processed with qualified counsel in each market.
Negotiating leverage
If you can move volume between processors in a day, every processor knows it. That pricing pressure is hard to put a number on, but it's real.
Volume, market and approval-rate thresholds for orchestration
These are rules of thumb based on how the economics usually work out. They aren't industry benchmarks. If two or more apply to you, model the orchestration case. If none do, a second acquirer is almost certainly the right call.
| Signal | Second acquirer is enough | Orchestration worth evaluating |
|---|---|---|
| Annual card volume | Under roughly $50M | Above roughly $100M, or growing quickly toward it |
| Markets needing local acquiring | 1–3 | 4 or more |
| Local payment methods to support | Fewer than 4 | 5 or more, across regions |
| Processors connected | 1–2 | 3 or more, or plans to add |
| Approval-rate gap between markets or processors | Under about 2 points | 3+ points your PSP can't explain |
| Stored-credential share of volume | Low, or covered by portable network tokens | High, with tokens tied to one processor |
Worked example (hypothetical)
Say a digital subscription company processes $200M a year at an average ticket of $80. That's about 2.5 million transactions across nine countries, at a 20% gross margin. If better routing and local acquiring lift approvals by one percentage point, about $2M more volume gets approved. At a 20% margin, that's roughly $400,000 in gross profit, before you count lower churn from fewer failed renewals. Now assume orchestration costs $0.04 per transaction plus a $60,000 platform fee. That's about $160,000 a year, and the case works.
Compare a merchant with $20M in volume at a $40 ticket in two countries, or 500,000 transactions. The same one-point lift approves about $200,000 more volume, worth roughly $40,000 in margin. Many platforms set annual minimums above that. For this merchant, a second acquirer with portable tokens provides most of the resilience at a fraction of the cost.
Be realistic about the lift you assume. One point is a sensible planning figure for a business whose setup is already reasonable. Bigger gains usually come from fixing a specific problem: no local acquiring, weak 3-D Secure handling in strong customer authentication markets, or no retry logic. You can fix several of those without a new platform.
Pricing ranges and the contract terms that lock you in
Orchestration pricing varies widely from contract to contract. The figures below are typical 2026 market ranges. Local pricing and enterprise discounts will differ, so confirm with each vendor.
| Option | Typical cost structure | Typical time to live | Solves |
|---|---|---|---|
| Direct second acquirer | Processing fees only; integration effort is internal | 6–12 weeks | Outage failover for new cards, pricing leverage |
| Second acquirer + standalone vault | Vault fees roughly $0.01–$0.05 per token use, or about $1,000–$10,000/month | 8–16 weeks | Full failover including stored credentials; less PCI scope |
| Full orchestration platform | Roughly $0.02–$0.10 per transaction; platform fees from about $2,000 to $25,000+/month; annual minimums often in the tens to hundreds of thousands of dollars | 3–9 months | Multi-processor routing, local methods, unified reporting |
Established vendors include Spreedly, Primer, Gr4vy, Yuno, CellPoint Digital, IXOPAY, APEXX and Paydock. Some large PSPs now also offer routing to third-party processors. That can work, but test how neutral the routing really is.
Push back on these in contracts and demos:
- Per-attempt pricing. If you pay for every attempt, each retry and each cascade to a backup processor gets billed again. Ask to pay on successful transactions, or for a cap on attempts.
- Vault export limits. Get a written commitment to a PCI-compliant bulk export of card data and network tokens to a processor or vault you choose, with fees and timelines stated. If a vendor won't commit to that, you're locked in.
- Uplift claims without a control group. A vendor promising approval gains should be able to run a holdout test that compares routed traffic with your current setup on similar transactions.
- Connector surcharges. Some vendors charge per processor or payment-method connector, or for keeping each one up to date. Get the full list and the price of each connector.
- Routing owned by an acquirer. If an acquirer owns the orchestration layer, confirm that its rules engine can route volume away from that acquirer without penalty.
- A new single point of failure. Orchestration platforms go down too. Ask for uptime SLAs and details of regional hosting, and find out what happens to authorizations if the platform itself is unavailable.
The decision rule and a 60-day evaluation plan
If you process in three or fewer markets, use no more than two processors and need fewer than four local payment methods, add a direct second acquirer and make your tokens portable with network tokens or a standalone vault. Choose orchestration if you're already past those limits, or will be within about 18 months, and your approval data shows gaps your current setup can't close.
Gather your own data before you talk to vendors:
- Pull 90 days of authorization data broken down by issuing country, card brand, BIN range and processor. Flag markets where approval rates sit clearly below the rest.
- Classify your declines. Soft declines, including most "do not honor" responses, can often be recovered with retries or rerouting. Hard declines, such as stolen cards or closed accounts, can't. If most of your declines are hard, better routing won't help much.
- Measure your stored-credential exposure. Work out what share of volume relies on processor-specific tokens, and what failover looks like for that volume today.
- Ask your current PSP which local acquiring, network token and retry features you haven't switched on. Turn those on first so you have a fair baseline.
- Run one side-by-side comparison. Get quotes for a second acquirer plus a vault, and from two orchestration platforms. Model each against a realistic approval lift, using the worked example above as a template.
If the orchestration case only works when you assume a lift of more than two points, go with the second acquirer and review again in 12 months.
Frequently asked questions
Is a payment orchestration platform the same as a payment gateway?
No. A gateway connects your checkout to one processor or acquirer. An orchestration platform sits above several gateways and acquirers and sends each transaction to one of them based on rules you set. Many orchestration vendors connect to existing gateways rather than replacing them.
Can a second acquirer handle failover for subscriptions and saved cards?
Only if your stored card credentials are portable. If your main processor holds the tokens, the backup acquirer can't charge those cards during an outage. Network tokens requested in your name, or a third-party vault, fix this, usually for much less than full orchestration costs.
How long does it take to implement payment orchestration?
Typically three to nine months. It depends on how many processors and payment methods you connect and whether card data has to move out of an existing vault. A direct second acquirer usually takes six to twelve weeks. Token migration is the most common cause of delay on both, so ask your current provider about export timelines early.
Does orchestration reduce PCI DSS compliance scope?
It often does, because the vendor's vault holds raw card data and your systems handle tokens. You still have compliance obligations, and your scope depends on how your checkout is built, so confirm your setup with a Qualified Security Assessor. Data-protection and residency rules such as GDPR, Brazil's LGPD and India's RBI requirements apply on top of PCI DSS and vary by market.
What approval-rate improvement should we expect from smarter routing?
Nobody can give you a credible figure without looking at your data, so treat headline claims with caution. For planning, assume about one percentage point if your setup is already reasonable. Larger gains usually come from fixing specific problems, such as missing local acquiring or poor retry logic. Insist on a controlled test against your current setup before you sign a multi-year commitment.
Should we choose an independent orchestration vendor or one offered by our PSP?
An independent vendor usually gives you more neutral routing and an easier exit, and that matters most when negotiating leverage is part of the case. A PSP's own routing can be quicker to switch on if most of your volume already runs through that PSP. Either way, get written terms on routing volume to competing processors and on exporting your vault.
Written by the Adviser Hub Cloud editorial team. Prices, features and rules change often - confirm details with the vendor or a licensed professional before you decide.