Why Uber’s CPO delivers food on weekends | Sachin Kansal
Watch on YouTube ↗ Summary based on the YouTube transcript and episode description. Prompt input used 79979 of 81625 transcript characters.
Uber CPO Sachin Kansal has completed 700-800 driver/delivery trips personally and argues extreme dogfooding — not data alone — drove Uber’s biggest product bets.
- Kansal has personally completed 700–800 Uber driving and delivery trips; earns money then refunds it to customers.
- Each team (driver, rider, Eats) has a biannual OKR to fix 300 dogfooding-discovered issues; targets hit every half for 2–3 years.
- Three major gut-over-data bets: integrating taxis (now every NYC yellow cab is hailable via Uber), Uber for Teens, and safety sentiment features in 2017.
- Uber runs 33 million trips daily; saving 1 cent per trip compounds massively — efficiency grind funded all new growth bets.
- Mobility dropped 80% overnight during COVID; Uber divested self-driving unit and Uber Elevate to survive, became profitable ~2 years ago.
- Uber’s AV strategy is a hybrid marketplace — partner with Waymo, WeRide, Avride, May Mobility rather than own the stack.
- Kansal uses ChatGPT Deep Research as a driver-feature brainstorm tool, not an oracle; treats AI output as a starting point, not a decision.
2025-06-01 · Watch on YouTube