BaaS vs. Direct Sponsor Bank
The most consequential program model decision in embedded finance. BaaS gets you to market faster. A direct sponsor bank relationship lets you own the economics. The tradeoffs compound at scale.
BaaS is a middleware model. Direct bank is a principal model. The economics, compliance, and product flexibility are structurally different.
| Dimension | BaaS (Banking as a Service) | Direct Sponsor Bank |
|---|---|---|
| Time to launch | 60–120 days | 6–12 months |
| Interchange capture | 0–25 bps (after middleware share) | 80–160 bps (full capture) |
| Float economics | Retained by bank/BaaS layer | Negotiable — 3–5% yield available |
| Compliance ownership | Shared — BaaS handles significant layer | Full — you own BSA/AML, KYB/KYC |
| Product flexibility | Limited to BaaS provider's product set | Full — limited only by bank appetite |
| Operational complexity | Low to medium at launch | High — requires compliance ops team |
| Economics at $10M/mo | ~$150K–$300K annually | ~$960K–$1.9M annually |
| Right for | Early-stage, MVP, under $5M/mo volume | Scale programs, card issuing, full economics |
The economics gap at scale: At $10M monthly processing volume, the annual difference between BaaS and a well-structured direct program is $600K–$1.6M. That gap compounds every year you stay on BaaS after the volume justifies a direct relationship. Most programs wait too long to make the transition.
When BaaS is the right choice
BaaS is the correct starting point when speed to market is genuinely more valuable than economics optimization. Pre-product-market fit, when you are testing whether the financial product works at all, BaaS removes infrastructure complexity and lets you focus on product.
BaaS is also appropriate when your volume trajectory makes the economics gap immaterial. Under $2–3M monthly volume, the absolute dollar difference between BaaS and direct is small relative to the operational overhead of running a direct program.
The mistake is not choosing BaaS. The mistake is choosing BaaS without a defined migration trigger — a volume threshold, an economics target, or a product requirement — at which you commit to moving to a direct model.
When to move to a direct sponsor bank relationship
Three triggers typically drive the migration decision: volume ($3–5M monthly is the common inflection point), product requirements (card issuing, float economics, or compliance programs that BaaS providers don't support), and economics visibility (when the annual leakage becomes a line item leadership can no longer ignore).
The migration itself takes 6–18 months depending on how the BaaS program was designed. Programs built with migration in mind — clear contract termination provisions, integration architecture that doesn't tightly couple to the BaaS layer, compliance programs that operate independently — move in 6–9 months. Programs that weren't designed for migration often take 12–18 months and significant re-architecture.