The Twenty Minute VCAdam Fishman: Takeaways from Lyft and Patreon; How to Build Viral Loops | 20VC #897
CHAPTERS
- 0:00 – 2:58
From quantitative marketing to product-led growth: Adam’s path into growth
Adam shares how he entered growth through early web analytics and experimentation, blending consumer psychology with an engineering mindset. He describes the step-function jump from public-company marketing into startups, culminating in the Zimride-to-Lyft pivot that shaped his approach to scaling.
- •Early career in web analytics and experimentation (pre–Google Analytics era)
- •Quantitative marketer roots: channel optimization, conversion work, living in Excel
- •Transition from public company to Zimride, then rapid pivot to Lyft
- •Lyft as a “graduate degree” in company-building and broad responsibility
- •Evolution toward product-led growth across the full customer lifecycle
- 2:58 – 4:39
Key lessons from Lyft and Patreon: speed, failure tolerance, and customer reverence
Adam contrasts what he learned at Lyft versus Patreon. Lyft taught him the compounding advantage of speed and a healthy relationship with reversible mistakes, while Patreon taught deep respect for customers—especially creators with public megaphones.
- •Speed enables reversible decisions and raises an org’s risk tolerance
- •Growth requires near-immunity to failure and frequent wrong bets
- •Patreon creators provide fast, loud feedback when you miss the mark
- •Quant-only decision-making can backfire when you skip customer input
- •Growth must balance outcomes with trust and long-term relationships
- 4:39 – 6:16
How growth teams learn from failure: structured experimentation and documentation
The conversation turns to how to avoid repeating mistakes when moving fast. Adam outlines an experimentation discipline that treats failed tests as learning assets, emphasized through strong documentation, sharing, and postmortems.
- •Failed experiments are acceptable only if they produce actionable learning
- •Use structured shots-on-goal, not “spaghetti cannon” tactics
- •Document experiments rigorously and broadcast learnings across the org
- •Postmortems and written records prevent repeating the same failures
- •Thick skin is a prerequisite because the hit-rate declines over time
- 6:16 – 7:55
Data vs intuition: blending quantitative signals with qualitative customer insight
Adam explains why he rarely relies on gut, but also warns against over-rotating on metrics alone. Patreon’s missteps often came from trusting numbers while failing to ask creators how changes would feel.
- •Avoid “gut-only” decision-making, but don’t worship dashboards either
- •Quantitative data can miss sentiment, trust, and perceived fairness
- •Qualitative inputs: customer conversations, prototyping, concept testing
- •Numbers can predict outcomes while still producing backlash
- •Best practice is a deliberate blend of quant + qual evidence
- 7:55 – 9:56
Defining “growth” and the modern Head of Growth role
Adam offers a clear definition of growth as understanding the company’s growth mechanics, identifying levers, and testing them to separate fact from fiction. He distinguishes the growth function (acquisition, retention, monetization) from the leadership role that steers strategy and teams.
- •Growth connects more people to existing product value (not creating value from scratch)
- •Core domains: acquisition, retention/activation, monetization
- •Growth orgs focus on levers and disciplined testing to validate hypotheses
- •Head/VP of Growth should typically lead a team, not be a solo practitioner
- •Role scope varies by company stage and title inflation realities
- 9:56 – 12:15
Hiring your first growth leader: sweet spot between too junior and too senior
Adam explains how to decide between hiring senior leaders versus junior doers, starting from whether the company understands its growth model. He advocates for a “player-coach” director-level profile when founders want progress without premature empire-building.
- •Start by clarifying whether you truly understand your growth model
- •Define which skills/functions are needed to iterate on that model
- •Founder involvement level influences the right hire and org structure
- •Too junior: order-takers who can’t prioritize; too senior: hires who only want to hire
- •Ideal early hire: director-ish, quantitative generalist, strong communicator
- 12:15 – 15:33
Where growth should live in the org: standalone team vs embedded in product/marketing
Adam breaks down why the right org design depends on your growth model and constraints. He explains the trade-offs between focus/accountability in standalone teams versus reduced friction (but often weaker focus) when growth is embedded.
- •Don’t blindly replicate the Facebook growth org model
- •Org placement should match the primary problem: acquisition vs retention vs monetization
- •Viral/collaboration-driven products often fit best inside Product
- •Paid or channel-heavy growth often fits best inside Marketing
- •Org design is dynamic; standalone teams often fold into other orgs over time
- 15:33 – 18:23
Growth loops 101: how loops power growth models (with Lyft’s referral loop)
Adam defines growth loops as closed circuits where outputs become new inputs, creating compounding cycles. He illustrates with Lyft’s give-get referrals and explains how multiple loops compose the product experience, later forming qualitative and quantitative growth models.
- •Loop definition: a self-contained cycle where outputs feed back as inputs
- •Lyft example: invite → signup → ride → inviter rewarded → repeat
- •Products are made of multiple loops that can compound over time
- •Qualitative growth model: visual mapping of loops end-to-end
- •Quantitative growth model: adding conversion rates/step changes to the loop map
- 18:23 – 21:13
Designing effective loops: simplicity, levers, plateaus, and sequencing new loops
Adam advises keeping loops simple (typically 3–6 steps) and using them to identify concrete optimization levers. He also covers why loops plateau (max scope/horizon), why user-generated loops scale further, and how to sequence new loops over time.
- •Keep loops simple; overly granular maps don’t help prioritization
- •Within a simple loop, levers include incentives, step conversion, invite rate, volume
- •Loops plateau when you reach max scope/horizon
- •User-generated loops tend to have higher ceilings than company-generated/paid loops
- •Start with 1–2 core loops, forecast plateaus, then expand eligibility or add new loops
- 21:13 – 25:21
Where companies go wrong with loops—and a referral loop flop in tutoring
Adam highlights two common loop mistakes: mislabeling slow processes as loops and trying to run too many loops at once. He shares a personal failure at Wyzant, where a referral program failed due to stigma around tutoring in the US—an example of psychology overpowering incentives.
- •Not everything is a loop; true loops must be closed circuits
- •Slow partnership/BD motions often aren’t self-propelling loops in practice
- •Too many concurrent loops create distraction and kill focus
- •Wyzant referral program failed despite incentives due to parental stigma
- •Lesson: customer psychology and social norms can break “pattern-matched” loops
- 25:21 – 28:38
Hiring process for growth leaders: prework, competency design, and “interviewing on the job”
Adam proposes a more realistic approach to hiring leaders: simulate the job as much as possible and even work side-by-side when feasible. He then outlines the formal process starting with founder prework, defining competencies, and building interviews that mirror real growth work.
- •Traditional exec interviews are often broken; real collaboration reveals more
- •Prework: know what you need based on your growth model (product vs marketing hire)
- •Ideal early profile: quantitative generalists, strong communicators; even failed founders
- •Define core competencies and how you’ll test each
- •Use exercises that approximate real work, not abstract trivia
- 28:38 – 36:14
The interview gauntlet: quantitative & prioritization exercises plus behavioral signals
Adam gets specific on interview components: a dummy dataset analysis to test insight generation, a prioritization exercise with evolving constraints, and behavioral interviewing to assess humility, feedback reception, and influence. He shares example questions that expose ego, blame, and learning ability.
- •Quant exercise: analyze dataset, build cohorts, identify drop-offs, propose next digs
- •Prioritization: rank bets with frameworks (impact/confidence/reach), adapt to new info
- •Behavioral interview probes risk tolerance, grit, humility, and feedback handling
- •Sample prompts: times you were wrong, influenced executives, responded to criticism
- •Strong leaders take accountability; chronic blamers are red flags
- 36:14 – 42:50
Onboarding and making hires successful: 30/60/90 plans, visibility, and early momentum
Adam reframes onboarding as employee activation that drives long-term retention. He recommends a clear 30/60/90-day plan shared broadly, heavy founder time investment early, and fast feedback loops to detect lack of momentum before the 90-day mark.
- •Onboarding is “activation” for employees; early months decide trajectory
- •Use The First 90 Days approach with clear milestones and deliverables
- •Month 1: learning/relationships; Month 2: iterate or create growth strategy; Month 3: longer-term alignment
- •Publish plans widely (exec team, peers, team) to set expectations and accountability
- •If there’s no hunger/momentum by ~45–60 days, have a candid reset conversation
- 42:50 – 48:02
World-class cross-functional communication: predictable updates, tailoring, and prototypes
Adam outlines how to make growth welcome across functions through overcommunication, stakeholder-specific approaches, and reliable cadences. He shares his ‘Ship It Friday’ email format at Patreon, emphasizes visual communication via prototypes, and advocates transparency with kindness to invite feedback.
- •Overcommunicate until others truly understand; assume your context isn’t shared
- •Tailor channel and depth to stakeholders (1:1s for CEO/CFO vs async for others)
- •Create predictable streams like weekly updates with consistent sections
- •Ship It Friday email: metrics, highlights/lowlights, learnings, upcoming work; include humanity/fun
- •Use prototypes/visuals to compress alignment time; practice candid, kind transparency
- 48:02 – 52:32
Patreon rebrand: when positioning is the growth constraint—and how to do it safely
Adam explains why Patreon needed repositioning and how growth leaders can drive brand work responsibly. He stresses getting expert help, building internal alignment (especially with founders), using qualitative ‘brand-blind’ testing, pairing brand promises with product changes, and preparing for short-term backlash.
- •Rebranding can be the primary growth lever when positioning caps adoption
- •Don’t assume branding expertise—bring in coaches and frameworks
- •Bring skeptics (CEO, board, functional leaders) along; it’s emotionally hard for founders
- •Test positioning qualitatively (brand-blind statements) rather than simplistic A/B tests
- •Brand is a promise; product must deliver—launch features alongside the new identity and ignore social hot takes for ~2 weeks
- 52:32 – 1:01:06
Quickfire: timeless principles, dying tactics, hiring mistakes, and standout recent growth
In rapid-fire, Adam compares management to parenting, highlights enduring psychology as the foundation of growth, and notes which tactics have become harder or commoditized. He also shares common hiring misconceptions, a major Patreon ‘fee-asco’ mistake, and a recent product-led strategy he admires (Locket).
- •Management vs parenting: growth mindset, choices, empathy, goals, rewards—nearly identical
- •Timeless growth: psychology and biases (loss aversion, scarcity, status)
- •Harder/commoditized: social graph/address book hacks, incentivized referrals, content marketing, SEO (higher bar now)
- •Big hiring mistakes: expecting a quick fix; hiring growth before product–market fit
- •Worst experiment: Patreon fee shift rolled out without creator input; Impressed by Locket’s widget-driven, invite/scarcity-based viral PLG
