When Basic Fraud Scoring Stops Being Enough
Stripe Radar built its reputation by making fraud detection accessible. Any developer integrating Stripe’s payment stack could instantly inherit a machine learning layer trained on billions of transactions, with minimal configuration and no dedicated risk team required. For early-stage startups processing a few thousand transactions a month, that was more than enough. But fintech companies that mature – those running lending products, crypto on-ramps, neobanks, or cross-border remittance corridors – eventually hit a ceiling that Radar was never designed to break through.
Sardine is designed specifically for that ceiling.
Founded by former Coinbase and PayPal fraud executives, Sardine started with a premise that the fraud problem in fintech is not a payments problem – it is a behavioral problem. The company built its detection layer to analyze device signals, behavioral biometrics, and network graph relationships before a single dollar moves. That sequencing matters. Radar evaluates transactions after the payment intent is formed. Sardine intervenes at account creation, during onboarding, throughout session activity, and at the moment of funding initiation. The scope is different by design.
What Sardine Actually Catches That Radar Misses
The most persistent fraud vectors targeting fintechs right now are not simple card testing bots or stolen credit cards – those are table stakes. The real exposure sits in synthetic identity fraud, account takeover through social engineering, and first-party fraud where a real customer disputes a legitimate transaction. These attack types share a common feature: they look completely normal at the payment layer. A synthetic identity has a real bank account attached. An account takeover uses valid credentials. A first-party fraudster genuinely initiated the transaction. Radar scores these as low risk because, by payment-layer metrics, they are.
Sardine addresses this by building a behavioral baseline for every user session. The platform tracks how a person types, how they navigate a mobile interface, how long they hover before confirming a transfer, and whether those patterns match the device fingerprint history tied to that account. When a session suddenly shows robotic keystroke uniformity, or a user who always logs in from Chicago suddenly exhibits touch patterns consistent with a remote-access tool, Sardine flags it before the transaction clears. This is not a feature Radar has, because Radar is not positioned inside the session – it sits at the checkout node.
Device intelligence is another area where the gap widens. Sardine maintains a shared consortium of device and behavioral signals across its customer base, which means that a device associated with mule account activity at one neobank will carry that risk signal when it appears at another Sardine customer. That cross-network intelligence is structurally similar to how Stripe benefits from seeing payment card behavior across its merchant network, but Sardine applies it to identity and device layers rather than card numbers. For fintechs operating in high-risk corridors – ACH, instant payouts, crypto – that consortium effect is the actual product being sold.
The Market Sardine Is Eating Into
Stripe Radar’s installed base inside fintech is enormous, partly because the switching cost of moving off Stripe payments is high. But the fraud detection layer and the payment processing layer are separable decisions, and a growing number of fintech operators are making that separation deliberately. A company can stay on Stripe for payment routing while layering Sardine on top for identity verification, behavioral monitoring, and transaction risk scoring. That decoupling is exactly the wedge Sardine is exploiting – it does not ask anyone to leave Stripe, which removes the single biggest objection in any enterprise sales conversation.
The companies most likely to make this switch are those with compliance pressure that Radar cannot address. Bank Secrecy Act obligations, FinCEN reporting requirements, and state money transmitter licensing all demand explainable risk decisions. When a regulator asks why a specific transaction was approved or denied, “our ML model said so” is not an acceptable answer. Sardine’s platform generates audit trails, risk reason codes, and case management workflows that satisfy compliance teams and external auditors. Radar does not produce that output because it was not built for regulated entities – it was built for e-commerce merchants who need to minimize chargebacks, not document BSA compliance.
The pricing structure also shifts the competitive calculation. Radar charges a small per-transaction fee layered onto Stripe’s processing margin. Sardine operates on a platform subscription model that includes identity verification, fraud scoring, and compliance tooling under one contract. For fintechs processing at volume, the bundled cost of Sardine’s platform can land below what they would otherwise spend assembling those functions from separate vendors – a KYC provider here, a device fingerprinting tool there, a case management system somewhere else. That consolidation argument is straightforward to make to a CFO who is already managing five separate vendor relationships for functions Sardine rolls into one.
Why This Competitive Moment Is Harder for Stripe to Ignore
Stripe has expanded its compliance and identity tooling over the past two years – Stripe Identity, enhanced Radar rules, and improved dispute management are all part of that push. But Stripe’s fundamental incentive is to keep payment volume on its rails, which means its fraud product will always be optimized around transaction approval rates and chargeback ratios rather than around the broader identity and behavioral risk surface that regulated fintechs actually need to manage. Sardine does not have that constraint. Its entire business model is built on the assumption that the fraud problem extends well beyond the payment moment, and that fintechs will eventually pay for a tool that reflects that reality.
The telling detail is where Sardine’s sales team is focusing. They are not going after e-commerce. They are not targeting SaaS companies with Stripe Billing integrations. The pipeline is concentrated in crypto exchanges, earned wage access platforms, BNPL providers, and cross-border payment corridors – every category where ACH return rates, instant funding risk, and identity fraud are material business problems rather than edge cases. That targeting is deliberate, and it maps precisely onto the segment of Stripe’s fintech customer base that has the most acute unmet need.
