How to find hidden growth opportunities in your product | Albert Cheng (Duolingo, Grammarly, Chess.com)

Albert Cheng frames growth as helping people reach a product's value, rather than as an isolated effort to push metrics. That framing changes with the user's stage: a prospective user must understand the value proposition, while a long-tenured habitual user may need a different reason to remain engaged. He argues that local improvements to a metric through friction or paywalls can work in a narrow sense, but a durable company needs teams to begin with the user journey and then form specific hypotheses from it.

His early musical training supplies a useful analogy, not a claim that music makes someone a growth practitioner. Both domains combine repeated practice, quick feedback, and tolerance for mistakes with a formal structure and room for creative judgment. Cheng uses that comparison to explain why a growth organization needs both models, metrics, experiments, and channels, and the imaginative hypotheses that make those tools productive.

Cheng's explore-and-exploit model separates choosing a promising problem from concentrating resources on solving it. Too much exploration leaves a team scattered across unconnected ideas. Too much exploitation produces local optimization, saturation, and stagnation. He applies the cycle at the level of an insight rather than treating it only as a company-wide strategic choice: investigate behavior, articulate the finding, spread it to adjacent teams, and return to more divergent work once the opportunity has been exhausted.

Chess.com's game-review feature illustrates the method. The team expected people to review losses to find mistakes, yet observed that about 80% of reviews followed wins. It changed the loss experience from emphasizing blunders to highlighting strong moves and offering encouragement. Cheng reports that this lifted game reviews by 25%, subscriptions by 20%, and retention as well. The larger lesson is not to copy a UI treatment blindly. A result grounded in a human response to failure can prompt neighboring teams, such as the puzzles team, to test whether their own experiences feel unnecessarily cold.