Lenny's PodcastBrian Tolkin: How twin turbine jets shape Uber and Opendoor
How product and ops fly as a twin turbine jet plane at Uber and Opendoor; calm leadership and product reviews as collaboration, not a firing squad.
CHAPTERS
- 0:00 – 2:14
Product + ops as a “twin-engine” advantage (Uber & Opendoor)
Lenny tees up the central theme: Brian has scaled two companies where product and operations are deeply intertwined. Brian introduces the “twin turbine jet” metaphor—each function can temporarily carry the business, but the system performs best when both work in harmony.
- •Why product/ops integration is a core scaling advantage
- •The “twin-engine plane” mental model for operating companies
- •How operating realities should shape product strategy
- •Why this topic is rare but critical in ops-heavy tech businesses
- 2:14 – 6:15
How ops experience makes better product leaders
Brian explains how starting in operations builds a deep understanding of how the business actually works and keeps you close to customers. That foundation helps leaders identify what to build in a scalable, technology-driven way later.
- •Direct customer exposure (onboarding, support, field work) as product training
- •Ops builds intuition for what truly moves metrics and outcomes
- •From hands-on workflows to scalable systems thinking
- •Why proximity to the customer matters early in a career
- 6:15 – 10:01
Bridging distributed ops with centralized product: the rise of Product Operations
They discuss why product teams sometimes undervalue ops, and how Uber built a feedback loop between HQ product teams and globally distributed ops teams. Brian describes the emergence and purpose of the Product Operations function as connective tissue.
- •Common product-vs-ops tension and how to reframe it as harmony
- •Local teams iterate faster and surface qualitative insights
- •The bidirectional gap: shipping features globally + receiving market feedback
- •What “Product Operations” meant at Uber and how it was formalized
- 10:01 – 12:17
Surge pricing origins: humans-in-the-loop before full automation
Brian recounts how surge pricing operated for years as a manual, GM-controlled system with guardrails. They unpack why: novelty/risk, local knowledge of events, and technical constraints of building real-time geospatial pricing.
- •Manual surge parameters: when surge could activate, caps, and geographies
- •Local context (events, timing) beating early algorithms
- •Risk management for a powerful pricing lever
- •The technical difficulty of always-on geospatial dynamic pricing
- 12:17 – 15:46
Scaling war stories: UberX naming + UberPOOL China launch crunch
Brian shares two scaling stories: UberX started as a placeholder name that stuck, and the all-night firefight to launch UberPOOL in Chengdu. The episode highlights how high-stress launches create enduring lessons and memories.
- •UberX was a placeholder (“X”) that became the permanent brand
- •Chengdu UberPOOL launch: matching algorithm failures hours before go-live
- •Coordinating with US teams overnight under deadline and press pressure
- •Post-launch relief and the ‘best meal’ effect after extreme stress
- 15:46 – 16:51
Opendoor’s COVID pivot: shutting down and virtualizing homebuying/selling
Brian describes how COVID disrupted Opendoor’s core model of entering homes, forcing the company to stop buying homes temporarily. They used the pause to virtualize major parts of the process and emerge with a new operating model.
- •Why the core business was paused (health + market uncertainty)
- •Real estate activity signals from China influenced decision-making
- •Rapidly virtualizing a physical, ops-heavy workflow
- •Stress in the moment vs. fondness in hindsight
- 16:51 – 17:56
What Opendoor does (and which customer it serves best)
Brian gives a clear explanation of Opendoor as a digital platform for buying and selling real estate, centered on a seller product providing certainty and simplicity via an all-cash offer. Lenny reflects on how painful traditional home selling is and why the value proposition resonates.
- •Seller journey: enter home info → receive all-cash offer
- •Core customer value: certainty, simplicity, speed, control of closing
- •Pain points in traditional selling (showings, pricing uncertainty)
- •Opendoor as an ops + tech system, not just software
- 17:56 – 20:52
Making product and ops work: respect, leverage, and designing for real-world messiness
Brian outlines key principles for ops-heavy tech companies: mutual respect, ruthless focus on where tech creates leverage, and acknowledging that real-world systems are messy and need fail-safes. He also explains how the balance between ops and centralized functions evolves as companies mature.
- •Mutual respect as a prerequisite for high-functioning product/ops collaboration
- •Prioritize tech investment where leverage is highest (e.g., dispatch + pricing)
- •Make explicit tradeoffs about what not to build yet
- •Real-world entropy: cancellations, GPS issues, scheduling mistakes require resilience
- •How organizations centralize and shift as scale/optimization overtakes expansion
- 20:52 – 25:34
From ‘do things that don’t scale’ to automation: Uber driver onboarding evolution
They walk through a concrete scaling example: early Uber driver onboarding was 1:1 and highly operational, then gradually shifted to group sessions, videos, and eventually automated credential validation. The lesson: ops iterates first, then product/engineering scales what works.
- •Evolution: 90-minute 1:1 onboarding → small groups → larger classes → video
- •Scaling breaks processes (credential checks) and forces productization
- •Using OCR/automation to validate driver IDs at scale
- •Automation frees ops capacity to tackle the next hardest bottleneck
- 25:34 – 31:29
Running great product reviews (without making them a firing squad)
Brian shares how he structures product reviews to balance accountability with the primary goal: making the product better. He emphasizes leader behavior—probing, offering ideas as non-mandates, sharing missing context, and keeping the discussion small while distributing artifacts broadly.
- •Two explicit goals: inform/accountability + improve the product
- •How leaders reduce fear: questions over mandates, ideas as suggestions
- •Breadth vs. depth: leaders bring cross-context, teams bring immersion
- •Keep attendees <10; share docs/recordings widely as durable artifacts
- •Cadence: two weekly signup slots + ‘volun-telling’ to ensure quarterly coverage
- 31:29 – 40:07
Jobs To Be Done at Opendoor: practical adoption over dogma
Brian explains why JTBD is especially useful at Opendoor since employees aren’t frequent customers of home selling. They embed JTBD thinking into templates and language while avoiding rigid adherence, focusing on cultural internalization and better customer empathy.
- •JTBD forces deeper customer perspective when you can’t ‘build for yourself’
- •Accounting for off-platform context in long, complex real estate journeys
- •Implementation: product review template includes problem + JTBD sections
- •Avoiding dogmatism—frameworks are tools, not rules
- •Example reframing: ‘get an offer’ vs. broader job like price discovery
- 40:07 – 47:09
Experimentation at low volume: A/B testing limits, alternatives, and conviction-building
Brian breaks down why A/B testing is harder with fewer, high-stakes transactions and urges teams to do power analyses before committing. He outlines alternative approaches (observational methods, diff-in-diff, confidence tradeoffs) and when intuition must fill the gap—paired with clear feedback loops.
- •Run power analysis first; don’t force tests that can’t reach significance
- •Accept long runtimes for high-value questions (e.g., 6-month tests)
- •Alternatives: observational data, diff-in-diff, sister/twin markets, segmentation
- •Adjusting standards: e.g., 80% vs 95% confidence when appropriate
- •When to trust intuition—and how to set feedback loops post-ship
- 47:09 – 52:53
Zillow: from competitor to partner, and why vertical integration is hard
They discuss Zillow’s attempt to replicate Opendoor’s model, its challenges, and the eventual partnership. Brian highlights the complexity of a vertically integrated real estate transaction business requiring excellence in pricing, ops, risk, and capital markets—not just software and traffic.
- •Partnership rationale: Zillow reach + Opendoor transaction capability
- •Why iBuying is complex: pricing, operations, risk discipline, capital markets
- •Vertical integration as Opendoor’s DNA from day one
- •Staying ‘competition-aware’ but not ‘competition-focused’
- •The real competitor is the traditional way of buying/selling homes
- 52:53 – 56:21
Staying calm under pressure: leadership behavior and building the stress muscle
Brian explains that reflecting stress onto teams tightens performance rather than improving outcomes. He shares mantras and learning methods—repeated exposure, reflection, and studying other founders’ stories—to build perspective and remain even-keeled in crises.
- •Stress contagion harms team execution; calm improves clarity
- •Mantras: ‘never as good as you think / never as bad as you think’
- •Experience-based perspective: crises pass; learn tools that worked before
- •Don’t run from stressful situations—reflect and build resilience
- •Learning from others: biographies/podcasts like Founders to normalize nonlinearity
- 56:21 – 1:00:19
Finding the ‘kernel of truth’ amid noisy signals (and capturing ideas responsibly)
Brian describes product management as filtering countless inputs to identify what truly matters for customers and the business. He emphasizes documenting ideas to show respect and preserve context, then focusing execution on the highest-leverage problems.
- •Signals come from everywhere—customers, ops, execs, field visits, media
- •The PM job: separate noise from what moves the customer/business
- •Early Uber focus: connect rider/driver, price correctly, collect payment
- •Write ideas down to acknowledge contributors and revisit later
- •Use a backlog system (sheet/Jira/etc.) to operationalize the intake
- 1:00:19 – 1:14:39
Failure Corner + Lightning Round: early UberPOOL misstep, then rapid-fire favorites
Brian recounts an early UberPOOL launch strategy that over-optimized for specific commuter corridors and company-based matching, discovering liquidity was the only thing that mattered. They close with quick hits: favorite books, media, products, mottos, a key mentor, and a wild July 4th Uber interview story.
- •Early UberPOOL bet: commuter corridors + company-based matching didn’t scale
- •Key learning: liquidity dominates everything in marketplace products
- •Using aggressive promos ($5 Workpool) to test the upper bound of demand
- •Lightning round: books (Shoe Dog, Black Swan, etc.), shows, products (Fi, Particle)
- •Career influence: Jeff Holden; July 4th interview during Uber SUV launch
