Uber India PM Interview — Rider Retention Sequencing

It tests whether you can rank three rider-retention initiatives for the India market and defend the order under challenge. The interviewer checks that you state an explicit prioritization criterion before ranking, attach each initiative to a specific retention or north-star metric, name the guardrail metric you would protect, and recognize the driver-side cost of every rider-side win. The hardest part is re-sequencing cleanly when a constraint such as halved engineering capacity is introduced, without abandoning the plan. It is a working-session style round, conversational but data-heavy, mirroring how Uber actually interviews PMs.

Clarify scope and the single metric you are optimizing before proposing anything. State the criterion you will rank on, for example impact on 90-day retention, effort, and confidence, and say why those dimensions. Then present three initiatives, each tied to a measurable retention metric, with an explicit reason one comes before another, usually a dependency or a learning that de-risks the next bet. For every rider-side move, name the driver-side or unit-economics cost and the guardrail metric you would watch. Close by stating what you would cut first if forced.

The frequent failures are jumping into solutions before clarifying scope or the metric being optimized, optimizing the rider side while ignoring driver supply and surge economics, presenting a ranked list with no stated criterion, naming no north-star or guardrail metric, and collapsing the entire plan when the interviewer flips a constraint. Reciting a framework name like RICE without any Uber India numbers behind it also reads as weak. Asserting impact and effort figures with no stated assumption is a recurring trap that the interviewer will probe immediately.

The behavior is modeled closely on the real round but it is consistent and never tired or distracted. It interrupts rambling, restates your logic in one sentence to test it, and never praises an answer mid-session. It probes every claim for the baseline and attribution rather than accepting the headline metric. It will flip a constraint on you once you are doing well, exactly as a real loop interviewer pressure-tests a plan before taking it to leadership. The difference is uniform rigor and a transcript you can review afterward.

Your transcript is scored against role-specific dimensions such as prioritization-criterion clarity, marketplace tradeoff reasoning, retention-metric grounding, constraint recalibration, and personal product-decision ownership. Each dimension has observable anchors, so two evaluators would land close. You receive a scorecard that names the specific sequencing decision you could not defend under challenge and the moment a claimed number lacked a baseline. Scoring rewards applied judgment under pressure over reciting frameworks or naming generic best practices.

Do not solution yet. Restate the problem in your own words, confirm the scope, which cities, which rider segment, what timeframe, and name the single retention metric you will optimize for, for example 90-day cohort retention in the top five cities. State the criterion you will rank initiatives on and why. This signals structure before content, which is what the interviewer is listening for in the opening, and it earns you more real context about the situation rather than less.

Treat a flipped constraint, for example engineering capacity halved or a budget cut, as a re-sequencing problem, not a reason to discard your plan. Restate your ranking criterion, recompute which initiative still clears the bar under the new constraint, and explicitly name what you would cut or defer and why. Keep the optimized metric fixed. Candidates who keep the goal stable and re-derive the order under the new reality score well. Candidates who abandon the whole plan or freeze are the ones who lose this round.

A strong answer opens with scope and the metric being optimized, names an explicit ranking criterion, and presents three initiatives each tied to a retention metric with a dependency reason for the order. It volunteers the driver-side or margin cost of each rider-side move and the guardrail metric to watch, for example protecting driver earnings per hour while improving rider price predictability. It quantifies impact and effort with stated assumptions, and when a constraint is flipped it re-sequences without losing the goal and says exactly what it would cut.

Uber India is a two-sided marketplace, so a change that improves a rider metric can quietly worsen driver earnings, supply, or surge economics. The interviewer probes the driver side because recognizing that tension is one of the strongest signals Uber evaluates. Strong candidates name the specific driver-side cost of each rider-side initiative and the guardrail metric they would monitor to keep the system balanced. Ignoring the driver side, even with a clever rider feature, is one of the most common reasons candidates are downgraded in this round.

Yes. The jam session evaluates structure, recognition of marketplace dynamics, and how you take critical feedback while reasoning, which is exactly what this round drills. Defending a sequence under pushback, re-sequencing under a flipped constraint, and naming guardrail metrics are the same muscles the jam tests. The main difference is the jam is collaborative with multiple Uber employees, so practice integrating challenge without abandoning a clear point of view, which this scenario deliberately pressure-tests.