PhonePe PM Interview — Insurance North-Star Metric

It tests whether you can choose a single north-star metric for the insurance distribution business that captures real user value, separate that output from the input metrics a PM actually moves with the roadmap, and name guardrails that stop the team gaming the number. The interviewer presses on whether your metric survives the Indian context of low penetration, the trust deficit, claim-rejection fear, and thin take-rate economics. Hand-wavy framework recitals get screened out fast.

State the value the user receives from insurance, then commit to one north-star metric that measures that value rather than a transaction. Decompose it into two or three input metrics you can influence through the roadmap. Name guardrails covering conduct and product health, such as complaint rate, free-look cancellation, and claims experience. Then stress-test your own metric for gaming before the interviewer does. Commit to one number and defend it instead of listing six.

Picking revenue or policies-sold as the north-star and being unable to explain the user value behind it. Naming the output you are measured on but not the inputs you can move. Proposing a north-star with no guardrail and ignoring mis-selling and lapse risk. Reciting a textbook metric framework without applying it to PhonePe insurance specifics. Forgetting the Indian-market realities. Jumping to features before defining what success means.

It behaves like a PhonePe insurance pod lead: it opens with the prompt, probes every metric for its exact denominator and timeframe, and introduces a live complication such as an engineering capacity cut to test whether your metric and roadmap survive. It never praises you mid-round and never teaches you the framework. The difference is that every probe and the final scorecard are transcript-backed, so you can review precisely where your reasoning broke.

Your transcript is scored against role-specific dimensions such as how well you tie the north-star to user value, whether you separate output from controllable inputs, the strength of your guardrails against gaming, and how clearly you reason from the Indian insurance context. Each dimension has observable signals from what you actually said. You receive a scorecard that quotes the moments that moved each dimension up or down, not a single opaque grade.

Do not start listing metrics immediately. Take the offered minute to state what value PhonePe Insurance delivers to a user and to whom, since that anchors every metric choice. Decide your single north-star before you speak so you can commit to it under pressure. Have one input metric and one guardrail ready in the first answer so the interviewer hears decomposition early rather than a single vanity number.

Treat it as a prioritization test, not a trick. Keep the same north-star, since the goal does not change when capacity shrinks, and re-sequence the input metrics by which roadmap bets still move the north-star most per unit of engineering effort. Say explicitly what you would stop doing and why, and confirm your guardrails stay active because conduct risk does not pause when the team is smaller.

One committed north-star phrased as retained, value-delivering cover rather than a transaction, with an explicit denominator and timeframe. Two or three input metrics tied to specific roadmap levers like quote-to-purchase conversion and first-renewal rate. Guardrails that name conduct and health risks such as grievance rate and free-look cancellation. Every choice is justified against low penetration, the trust deficit, and thin take-rate economics, and the candidate defends the single metric under push-back.

Because PhonePe Insurance is a distribution business with thin take-rate economics, and a revenue or policies-sold north-star rewards volume even when policies lapse early or were mis-sold. That erodes unit economics and invites conduct risk under IRDAI scrutiny. A strong answer uses revenue as a business constraint and an input, not as the north-star, and chooses a metric that only counts a sale when it became a real, retained cover.

Deeper than naming low penetration once. The interviewer expects you to connect specific market realities to specific metric choices: the trust deficit and claim-rejection fear make a claims-experience guardrail non-optional, price sensitivity and thin premiums make a revenue north-star fragile, and low-intent payments traffic means activation and attach are the inputs you can actually move rather than high-intent search conversion like Policybazaar relies on.