Product Sense: Building a Product to Flag and Stop Fraudulent Card Payments
Company: Square
Role: Product Manager
Category: Product Design & Strategy
Difficulty: medium
Interview Round: Other
Design a system that detects fraudulent credit card transactions. Treat it as a product problem: decide whose losses the system protects and which users it serves, which kinds of fraud matter most, how it should respond when a payment looks suspicious, and how you would prioritize the work and measure whether the system succeeds.
```hint Every decline has a cost
Turning away a legitimate purchase loses a sale and a customer's trust, so think about how the system handles uncertainty instead of treating every payment as a simple fraud-or-not call.
```
### Clarifying Questions
- Whose losses are we protecting: the bank that issued the card, the cardholder, or the merchant and payment processor that absorb chargebacks?
- Are we covering in-person card payments, online payments where the card is not present, or both?
- Which fraud types matter most: stolen card details used online, counterfeit or skimmed cards, takeover of a customer's account, or merchants who are themselves fraudulent?
- Should the system decide on its own, or should merchants or cardholders have a say on borderline payments?
### What a Strong Answer Covers
- A clear choice of the protected party and payment channel, with the fraud types that matter most for that choice.
- What each affected user sees when a payment is approved, challenged, held for review or declined: the merchant, the buyer and the fraud analyst.
- How detection works at a product level: the signals used, a risk score, graded actions and a feedback loop from confirmed fraud.
- Prioritized solutions that weigh fraud caught against friction and false declines.
- Metrics for fraud losses, false declines and approval rate, with guardrails.
### Follow-up Questions
- A large merchant complains that the system declined many good orders during a holiday sale. How do you investigate, and what do you change?
- Fraudsters adapt to any new rule within weeks. How do you keep the system effective without steadily adding friction for real customers?
- When a merchant chooses to accept a payment the system flagged and it turns out to be fraud, who should bear the loss?
Overview: A product sense prompt for product manager interviews that asks for a system to detect fraudulent credit card transactions. It tests choosing whose losses to protect, prioritizing fraud types, designing graded responses to suspicious payments, and balancing fraud caught against false declines and friction.