Google·Software Engineer·Technical Phone Screen·Senior

Senior

Jul 2026

Summary

Data science interview at Google with a product metrics question framed around Lyft. Just the one question from what I can tell, but it's the kind that spirals fast if you don't have a structure ready.

Questions Asked (1)

If you were the PM for Lyft, what dashboard would you build to monitor the overall health of the app?

Product Analytics & MetricsProduct Sense & Ideation

Author's notes

I blanked for a second on where to even start.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Approach this by first defining Lyft's core business goals and user journeys (riders and drivers), then map metrics to each stage of those journeys using a structured framework like AARRR or a supply-demand lens. Organize the dashboard into logical sections that reflect both real-time operational health and longer-term business health, ensuring you cover both sides of the marketplace.

Pro tip: Demonstrate product maturity by explicitly calling out the dual-sided marketplace nature of Lyft — a healthy app requires balancing both rider demand and driver supply, and ignoring either side is a common mistake that signals shallow thinking.

1. Define the Goal & Audience

Clarify what 'overall health' means for Lyft — revenue, reliability, user satisfaction, or growth — and identify who will use this dashboard (executives, ops teams, engineers). This scoping prevents building a dashboard that tries to serve everyone and ends up serving no one.

2. Map the Core User Journeys

Outline the key flows for both riders (open app → request ride → match → trip → payment → rating) and drivers (go online → receive request → pickup → drop-off → earnings). Metrics should trace health at each stage of these journeys.

3. Define Metric Categories

Organize metrics into buckets: Acquisition (new users, driver sign-ups), Engagement (rides per user, driver utilization), Reliability (app crash rate, match success rate), Satisfaction (rider/driver ratings, cancellation rate), and Revenue (GMV, take rate). This structure ensures comprehensive coverage.

4. Prioritize North Star & Supporting Metrics

Identify a single North Star metric (e.g., completed rides per day) that best captures overall health, then select 5-8 supporting metrics that act as leading or lagging indicators. Avoid metric overload by ruthlessly prioritizing actionability.

5. Design for Actionability & Alerting

Structure the dashboard with real-time operational metrics (ETA accuracy, surge pricing, driver availability by geo) alongside weekly/monthly trend views, and include anomaly alerts so teams can act quickly when a metric degrades.

Key Points to Mention

Dual-sided marketplace dynamics: rider demand metrics (ride requests, conversion rate) AND driver supply metrics (active drivers, acceptance rate, utilization rate)
North Star Metric: completed rides per day/week as the single best proxy for overall app health
Reliability & technical health: app crash rate, API latency, ride-matching success rate, and payment failure rate
User satisfaction signals: average rider and driver ratings, cancellation rate (rider-initiated vs. driver-initiated), and support ticket volume
Financial health indicators: Gross Merchandise Value (GMV), average revenue per ride, take rate, and promo/incentive spend efficiency
Geographic and temporal segmentation: ability to drill down by city, time of day, and day of week to surface localized issues like driver shortages in specific markets

AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.