Favorite Products & Optimization

## How to Approach (Brief Framework) - Part A (Admiration & Comparison): For each product, state the job-to-be-done, what it does uniquely well, and 2–3 competitors with trade-offs. - Part B (Improvement Plan): Pick one product, define target users and JTBD, list pain points, define success metrics (North Star + guardrails), and propose a prioritized roadmap (using RICE or Impact/Effort). Close with experiment design and risks. --- ## Part A — Three Products I Admire 1) Google Maps (technology) - Why I like it: Best-in-class coverage, fast routing, strong multimodal support (driving, transit, walking, cycling), tight integration with local search and reviews, reliable offline maps. - Key competitors and comparisons: - Apple Maps: Strong iOS integration and design polish; improving data quality. Historically weaker POI depth and international coverage. - Waze: Superior incident reporting and real-time driver alerts; weaker place search and non-driving modes. - HERE WeGo: Solid offline routing and licensing; less consumer mindshare and fewer local discovery features. - Trade-off: Maps balances breadth (global coverage, POIs) with routing accuracy. Opportunity remains in explainability, first/last-mile, and reliability for EV and parking. 2) Notion (technology) - Why I like it: Flexible building blocks (blocks + databases) let users model workflows from notes to lightweight apps. Strong templates and community ecosystem. - Key competitors and comparisons: - Evernote: Great for note capture, simpler; less flexible for structured data and collaboration. - Coda: Powerful automations and packs; steeper learning curve and heavier spreadsheet paradigm. - Confluence: Enterprise documentation strength; less flexible for personal productivity and databases. - Trade-off: Notion wins on flexibility and usability at the cost of performance for very large workspaces and a learning curve for new users. 3) OXO Good Grips Swivel Peeler (non-technology) - Why I like it: Ergonomic handle reduces hand fatigue; sharp, replaceable blade; dishwasher safe; durable. - Key competitors and comparisons: - Generic Y-peelers: Cheap and fast but less ergonomic and dull faster. - Kuhn Rikon peeler: Very sharp and light; handle comfort varies, blades can chip. - Trade-off: OXO optimizes ergonomics and reliability over the absolute lowest price. --- ## Part B — Improving Google Maps I’ll focus on enhancing end-to-end trip confidence for non-driving and mixed-mode use cases (walking, transit, EV, and arrival/parking). This aligns with user trust, safety, and task completion—core to navigation products. ### 1) Target users and jobs-to-be-done (JTBD) - Urban commuters (transit + walking): “Help me get to work reliably, including first/last mile, without crowded trains or missed transfers.” - Pedestrians and travelers: “Help me feel safe and confident navigating unfamiliar areas and the last 50 meters to the entrance.” - Drivers (including EV): “Help me arrive on time with predictable parking/charging.” ### 2) Key pain points - First/last mile ambiguity: Confusing building entrances, campus/indoor transitions, and last 50–100 meters. - Transit crowding and reliability: Limited visibility into real-time crowd levels and platform-specific delays. - Pedestrian safety: Desire for safer routes at night (lighting, foot traffic, sidewalk quality) and accessibility needs (curb cuts, elevators). - Parking and arrival: Uncertain street/garage availability; time lost circling for spots. - EV charging reliability: Stations shown as available but broken, wrong plug type, or payment friction. - Route explainability: Users want to know why a route was chosen and to tune preferences (e.g., fewer left turns, avoid tolls, prioritize safety). - Offline robustness: Downloads can be stale; failures in low-signal environments. ### 3) Success metrics - North Star: Completed trips without confidence loss - Definition: A navigation session that reaches the intended entrance without app-switching or itinerary abandonment. - Proxy: “Successful trip rate” = completed trips / started trips. - Supporting metrics (by pain point): - ETA accuracy: Median absolute ETA error (|ETA − actual|). Target: improve by 20–30% for walking/transit legs. - Last-50m success: Share of sessions with no backtracking >50 m near destination. Target: +25% improvement. - Transit crowding accuracy: Correlation between predicted vs. observed crowd levels; reduce “overcrowded surprise” rate by 30%. - Parking success rate: Users find a spot within 5 minutes of arrival. Baseline 40% → 60%. - EV session reliability: Successful charging sessions on recommended stations. Baseline 70% → 85%. - Safety satisfaction: Thumbs-up rate on “safer route” option; incident reports per 1,000 km decrease by 20% on these routes. - Guardrails: App crash rate, battery consumption, data/privacy incidents, average route time inflation (avoid overly conservative routing). Small numeric example (ETA accuracy): - Suppose baseline median |ETA − actual| for walking = 4.0 minutes across 10,000 trips. Target 3.0 minutes represents a 25% reduction. If achieved, expect user-reported confidence to improve and lower app-switching by 10–15%. ### 4) Prioritized roadmap (12-month horizon) Prioritization uses RICE (Reach, Impact, Confidence, Effort) qualitatively; items chosen for high user impact and feasibility. Phase 1 (0–6 months) 1. Arrival & last-50m clarity (High impact, Medium effort) - Features: Entrance-level directions, AR-guided last steps, indoor/outdoor transition cues, photo cues for entrances. - Data: Fuse POI polygons, Street View imagery, venue maps, user-contributed entrance photos. - Success: +25% improvement in last-50m success; +10% increase in trip completion without backtracking. 2. Parking & curbside experience (High impact, Medium-High effort) - Features: Real-time street/garage availability, price/time overlays, “park and walk” time cost in ETA, save frequent garages. - Partnerships: Parking providers and municipalities; ML predictions from historical occupancy. - Success: Parking success 40% → 60%; ETA error reduction for driving trips with parking considered. 3. Explainable routing & preferences (Medium-High impact, Medium effort) - Features: “Why this route” summaries (traffic, safety, transfers), preference toggles (left turns, tolls, ferries), safe-route option for walks at night. - Success: +15% route acceptance; -20% re-route after start; improved trust NPS. Phase 2 (6–12 months) 4. Transit crowding and reliability (High impact, High effort) - Features: Real-time crowding and reliability predictions at the line/vehicle level, platform guidance. - Data: GTFS-RT, agency feeds, on-device signals (opt-in), historical patterns; privacy-preserving aggregation. - Success: -30% “overcrowded surprise” reports; +10% on-time arrival for transit commutes. 5. EV charging reliability (High impact for EV segment, Medium effort) - Features: Verified station status (heartbeat pings), plug-type matching, payment integration, user verification of successful sessions. - Success: EV session reliability 70% → 85%; -40% “arrived but broken” flags. 6. Offline robustness & delta updates (Medium impact, Medium effort) - Features: Auto-refresh for downloaded regions, compact delta updates, predictive prefetch based on calendar/trips. - Success: -30% navigation failures in low-signal areas; fewer map-staleness complaints. Example RICE scoring snippet (qualitative): - Arrival & last-50m: Reach High (all users), Impact High, Confidence High, Effort Medium → Do first. - Transit crowding: Reach High (urban users), Impact High, Confidence Medium, Effort High → Phase 2. ### 5) Experimentation and validation plan - Rollout: Start with 1–2 cities per feature (e.g., Tokyo and NYC for transit; SF and London for last-50m). Gradual ramp by platform. - A/B tests: Measure North Star and secondary metrics; guardrails on battery, crash rates, and time-to-first-fix (TTFF) for GPS. - Measurement details: - ETA median absolute error: median(|ETA − actual|) across trips in treatment vs. control. - Parking success: fraction of driving trips with a spot found within 5 minutes; survey validation for a subset. - Safety satisfaction: opt-in feedback post-trip on “safer route.” - Qual/quant blend: Diary studies for last-50m, intercept surveys at stations, heatmaps of abandonment. ### 6) Risks, assumptions, and mitigations - Data bias in “safer routes”: Risk of reinforcing inequities. - Mitigation: Use multi-signal safety (lighting, sidewalks, traffic), allow user opt-in, publish model cards, and add transparency. - Privacy: Transit crowding and EV telemetry must be opt-in and aggregated. - Mitigation: Differential privacy, device-side processing where possible, clear consent flows. - Battery and performance: AR and real-time feeds can drain battery. - Mitigation: Adaptive sampling, on-demand AR, performance budgets in rollout. - Partnership reliability: Parking and EV operators vary in API quality. - Mitigation: Provider reputation scores, fallback heuristics, user verification signals. --- ## Summary - Focus the product on increasing trip confidence and reliability, especially for last-50m, parking/arrival, transit crowding, and EV charging. - Measure success with a clear North Star (completed trips without confidence loss) and targeted metrics per pain point. - Ship high-impact, medium-effort features first (arrival clarity, parking, explainability), then expand to data-intensive improvements (crowding, EV) with careful experimentation, privacy, and performance guardrails.