Business Model, Market, and Competitive Landscape Questions

Change 1 — Accessibility mode: add rider settings for accessibility needs (assistive pick-up instructions, wheelchair-accessible vehicle priority, audio cues). Product mechanics: tag drivers with accessible-vehicle capability, route matchers surface these drivers, provide training micro-certifications. Social impact metrics: % fulfilled accessible rides, reported accessibility satisfaction, reduced wait times for riders with needs. Business metrics: retention among riders with accessibility needs, Net Promoter Score, incremental trips. Change 2 — Green trip option: allow riders to choose ‘Low-emissions’ mode that prioritizes EV/hybrid drivers or pools; offer small fare premium or credits. Measure social impact: CO2e avoided per trip, % trips green; business metrics: take rate, AOV changes, retention of eco-conscious riders, driver earnings effect. Implementation: pilot in dense zones with available EV supply, partner with utilities for charging incentives. Both changes must include equity safeguards (no discriminatory pricing), and track adoption across demographics to ensure inclusivity.

Design: three-tier loyalty for frequent business riders. Tiers: Silver (10–19 trips/mo), Gold (20–39), Platinum (40+). Benefits: Silver — waived booking fees on 2 rides/mo, priority pickup; Gold — 5% off rides, 4 waived fees/mo, monthly expense reports; Platinum — 10% off, airport lounge credits, dedicated account support, rollover credits. Cost assumptions: average fare $25, booking fee $2.50. If 20% of trips have fees waived, gross revenue drop is small vs. increase in trip frequency. Experiment: randomized controlled trial with matched cohorts of business riders for 90 days; primary metric: trips per user/month lift; secondary: churn, revenue per user, CAC. Sample size: power to detect a 5% lift with alpha 0.05 — estimate ~5k riders per arm (depends on baseline variance). Guardrails: cap waived fees per month, monitor margin; success threshold: statistically significant +5% trips and net revenue neutral or positive within 6 months.

Diagnostic plan:
Data to analyze:

• Hourly wait times by neighborhood, time-of-day, day-of-week
• Driver supply: active drivers, acceptance rates, cancellations
• Demand patterns: request volume, surge events, ride cancellations
• Trip-level telemetry: routing, ETA accuracy, traffic, incidents
• External: events, weather, transit outages, road closures
  Hypotheses to test:

1. Supply shortage (fewer drivers logged in or redeployed elsewhere)
2. Increased demand (events, weather, promotions)
3. Routing/ETA regression (algorithm or data error)
4. Local traffic or road closures
5. Driver behavior changes (longer rides, more cancellations)
   Immediate tactical responses:

• Short-term: surge pricing and targeted driver incentives for affected zones; push notifications to idle nearby drivers; temporarily lower acceptance thresholds for bonuses.
• Medium-term: repositioning drivers via ops hubs; targeted marketing to balance demand.
• Quick experiments: rollback recent matching/ETA changes in those zones; increase real-time routing transparency.
  Follow-up: prioritize root cause, run A/B tests for algorithmic fixes, and deploy localized operational playbook.

Two product-level integrations:

1. Transit Timetable & Fare Integration

• Product: show real-time transit options and combinated fares (Lyft + transit), enabling riders to plan multimodal trips
• Measurement: % of rides that start/end at transit stops, reduction in single-occupancy trip distance for those users, increased shared/pooled trips

2. First/Last-Mile Scheduled Shuttles & Guaranteed Connections

• Product: scheduling API that reserves short Lyft pooled rides synchronized to transit arrival/departure times; subsidized pooled rides for commuters
• Measurement: on-time connection rate, decrease in missed-transit events, rider satisfaction score for multimodal trips

Both require APIs with transit agencies for GTFS-rt, shared KPIs (mode-shift, reduced car trips), and pilot zones. Success measured by increased transit ridership, decreased reliance on car trips in pilot zones, and improved customer NPS.

Potential harms: misleading or coercive nudges that push drivers into unsafe or unprofitable behavior, geographic discrimination (over-serving high-income areas), privacy leaks from heatmaps, and incentivizing gameable behavior. Three safeguards: 1) Transparency & consent: clearly label predictive features, explain confidence and estimated earnings uplift, allow opt-out; log and audit AI suggestions. 2) Fairness constraints in algorithms: enforce minimum expected earnings per reposition suggestion, cap repeated suggestions that favor certain neighborhoods, add constraint to preserve equitable distribution of rides across areas and driver demographics. 3) Safety & privacy protections: do not provide suggestions while driving at high speed; anonymize and aggregate heatmaps with differential privacy for public surfaces; allow drivers to mark zones as off-limits for personal safety. Additionally implement monitoring, human-in-the-loop review for outliers, and impact audits with driver focus groups.