The decision most fintechs make without realizing they're making it
Every company building embedded payments eventually chooses between three fundamental program models: Banking-as-a-Service (BaaS), Payment Facilitation (PayFac), and direct Sponsor Bank. Most of them make the choice by defaulting to whichever vendor they talked to first.
That default is expensive. The model you choose determines your margin ceiling, your compliance exposure, your capital requirements, and how hard or easy it becomes to change direction at $50M in volume versus $5M. Making it deliberately — before any vendor conversation — is the single highest-leverage decision in the embedded payments playbook.
Here's how each model actually works, from someone who has operated all three.
Banking-as-a-Service (BaaS)
What it actually is: You access banking infrastructure — ledgers, card issuance, ACH rails, compliance infrastructure — through a middleware provider (Synctera, Treasury Prime, Unit, Infinant) that manages the relationship with an underlying sponsor bank. You build on top of their stack.
The economics: BaaS providers charge a percentage of transaction volume plus monthly platform fees. At early volume, the unit economics are acceptable. At scale, they become untenable. The take rate is flat — your cost scales at exactly the same rate as your revenue, permanently. There is no margin expansion from volume.
The compliance reality: The bank holds the charter. You operate under their BSA/AML program. This feels like an advantage early — you don't have to build compliance infrastructure. It becomes a constraint at scale when the bank's risk appetite limits your product roadmap.
The scaling limit: Most BaaS programs hit an economics ceiling at $20–50M in monthly volume. Below that, the simplicity is worth the margin cost. Above it, you're paying infrastructure costs that belong to a model you've outgrown.
Right for: Early-stage programs, MVP validation, teams without in-house compliance expertise, programs where speed-to-market is the primary constraint.
Payment Facilitation (PayFac)
What it actually is: You become a sub-merchant aggregator. You board your customers as sub-merchants under your master merchant account. You control the payment experience, own the customer relationship, and capture the margin between what you pay the acquiring bank and what you charge your customers.
The economics: Better unit economics than BaaS at scale — you're capturing interchange margin directly rather than sharing it with a middleware layer. But the margin requires volume to materialize. Below $10M in monthly processing, the compliance and operational overhead often exceeds the economics gain.
The compliance reality: You own sub-merchant underwriting. KYC, KYB, chargeback liability, fraud management — these live with you. This is a real operational commitment, not an abstraction.
The scaling limit: Chargeback liability grows with volume. Sub-merchant underwriting becomes a significant operational function. The model scales well economically but requires operational infrastructure that many product-led teams underestimate.
Right for: SaaS platforms with established merchant relationships, companies with existing compliance infrastructure, programs where the economics case at scale justifies the upfront investment.
Direct Sponsor Bank Relationship
What it actually is: You work directly with a bank that sponsors your program. You hold your own program agreements, negotiate your own economics, and operate under the bank's charter while maintaining significantly more control over your program model.
The economics: The best unit economics of the three models at scale. You negotiate interchange sharing, float economics, and fee structure directly. There is no middleware margin extraction. The ceiling is the highest.
The compliance reality: The highest compliance burden. The bank will examine your BSA/AML program, your KYC/KYB process, and your operational controls. You need to be able to pass a bank-grade compliance review — or work with someone who can prepare you for it.
The scaling limit: The bank's risk appetite is the constraint. A conservative sponsor bank limits your product roadmap regardless of your own capabilities. Matching your program to the right bank — one whose risk profile aligns with your product — is the most important selection decision in this model.
Right for: Programs with $20M+ in monthly volume, companies with compliance infrastructure or the ability to build it, programs where long-term economics justify the setup investment.
The comparison that actually matters
The right model isn't the one with the best brochure. It's the one that fits your volume trajectory, your operational capability, and your economics requirements at the scale you're trying to reach — not where you are today.
Most programs that end up in the wrong model got there by optimizing for speed-to-market at the expense of economics design. BaaS is faster to launch. It's also the model you'll spend the most money migrating away from if you grow.
The cleanest path: define the economics model you need at your target scale first. Then select the infrastructure that gets you there — starting with the model that makes sense for your current stage but with a clear view of when and how you transition.
The question to ask before any vendor call
Before you talk to a BaaS provider, a PayFac platform, or a sponsor bank, answer these five questions:
- What is your target monthly processing volume at 24 months? At 36?
- What take rate do you need to make the unit economics work at that volume?
- Do you have — or can you build — in-house compliance infrastructure?
- How much operational overhead can your team absorb on day one?
- What is your exit strategy from this model if it stops working?
If you can't answer these before your first vendor conversation, the vendor's defaults will answer them for you. That's how programs end up with economics that made sense at $5M and broke at $50M.