Summary — How to find hidden growth opportunities in your product

Author/Host: Lenny Rachitsky
Guest: Albert Cheng (Duolingo, Grammarly, Chess.com)

Overview

This episode explores how product-led growth teams find and scale impactful opportunities inside consumer subscription products. Albert Cheng shares mental models, an actionable framework (explore vs. exploit), real product examples (notably from Chess.com), and growth philosophies drawn from his experience at Duolingo, Grammarly, and Chess.com. He emphasizes retention, experimentation, human psychology, and treating growth as connecting users to product value — not just “metrics hacking.”

Key points & main takeaways

• Growth = connecting users to product value. It’s not just A/B testing or hacking metrics.
• User retention is the most valuable lever for consumer subscription businesses. If retention is weak, monetization becomes forced to day-one conversions.
• Explore vs. exploit framework:
  • Explore = finding the right opportunities (finding the right mountain).
  • Exploit = focusing resources to climb that mountain effectively.
  • Danger: too much exploration = scattershot, no pattern matching; too much exploitation = local maxima and stagnation.
  • Apply this framework both at the strategic/macroscale and at the micro/insight level for product features.
• Test behavioral assumptions with data; real user behavior may be counterintuitive.
  • Example from Chess.com: it was assumed users would review games mainly after losses. Data showed 80% review after wins.
  • Product change: when users lost, the game review emphasized “best/brilliant moves” and encouraging coaching messages instead of bluntly surfacing blunders.
  • Outcome: that change materially increased engagement and monetization (game reviews up ~25%, subscriptions up ~20%).
• Growth work benefits from a mix of structure (models, metrics, experiments) and creative hypothesis generation—parallels with music training: repetition, tight feedback loops, resilience to mistakes.
• Run a high volume of experiments to surface learnings quickly; Albert's stated goal has been to run very large numbers of experiments (e.g., targeting ~1,000 experiments/year in prior roles) to accelerate learning.

Notable quotes / insights

• “User retention is gold for consumer subscription companies.”
• “When you're in exploratory mode, think of it as finding the right mountain to climb. In exploitation mode, you're focusing your resources on climbing that mountain.”
• “Growth is the job to connect users to the value of your product.”
• On parallels with music: both music and growth require consistent repetition, tight feedback loops, and a blend of structure plus creativity.

Topics discussed

• Albert’s background (piano, perfect pitch) and how musical practice shaped his growth approach.
• Explore vs. exploit framework and how to apply it both macro and micro.
• Real growth experiments and wins at Chess.com (game review example).
• The importance of psychological framing in product experiences (positive reinforcement vs. negative feedback).
• Experimentation practices and high-volume testing as a strategy for learning and scaling.
• Brief mentions (not expanded in this transcript) of growth wins at Duolingo and Grammarly, use of AI to accelerate growth work, and the role of brand/community in growth.

Action items & recommendations (practical next steps)

• Use the explore/exploit mental model:
  • Allocate deliberate time/space to explore hypotheses and to exploit validated wins.
  • Avoid extremes — maintain a balance between discovery and scaling.
• Validate behavioral assumptions with data before building/monetizing:
  • Instrument events and measure real user flows (you may find counterintuitive behaviors).
• Reframe product messaging to match user psychology:
  • If users respond better to positive reinforcement (e.g., reviewing wins), tailor the experience accordingly.
• Run many small, rapid experiments:
  • Prioritize learnings and pattern-match across experiments rather than treating each as isolated.
• Focus first on retention improvements; stronger retention makes monetization easier and more sustainable.
• Mix structured growth processes (models, metrics) with creative hypothesis generation — encourage both analytical and creative work in your team.

If you want, I can:

• Extract a short checklist you can apply to your own product (questions to ask, metrics to monitor).
• Convert the Chess.com example into a template experiment you can adapt.