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

What if growth is less about tricks and more about taste?

Albert Chang's career feels like a study in marrying discipline with curiosity. He moves from piano practice to product experiments, and the through-line is surprising: repetition with a taste of creativity. I found myself nodding more than once while he described growth not as clickbait engineering but as the craft of connecting people to real value.

Explore and exploit — a practical compass

Here's what stood out: the explore-and-exploit framework isn't an abstract theory for Chang. Exploration is finding the right mountain to climb; exploitation is scaling the route once you find it. He illustrated the idea with a deceptively simple chess.com example: players review games mostly after wins, not losses. Flipping the coach messaging to highlight brilliant moves after a loss increased game reviews, subscriptions, and retention. That felt like a small product change with a surprising human insight behind it — people return to positive reinforcement.

Sampling over secrecy: a freemium lesson

Honestly, I didn't expect such a clear monetization hack to feel so human. At Grammarly, sampling paid suggestions into free users' flows — a reverse real-time trial — nearly doubled upgrades. The logic is counterintuitive only if you assume scarcity always drives conversion. Instead, giving users a curated taste of higher-value capabilities changed their perception of the product. It made free users see Grammarly as more than a spellchecker; they began to see a writing assistant worth paying for.

Run experiments like breathing

Chang wants chess.com to run 1,000 experiments a year. That number is provocative, sure, but the point is sharper: treat experimentation as a muscle, not a marketing stunt. He described an evolution from a handful of experiments to hundreds, and the real win was cultural — leadership commitment, observable wins, and sharing learnings across teams. When experiments become routine, insights compound faster than any single breakthrough.

AI as speed, not spectacle

One of my favorite moments was his practical take on AI. He uses text-to-SQL Slack bots to answer ad hoc data questions and accelerate discovery. He also builds AI prototypes for product screens so teams can discuss and test sooner. The refrain: use AI to shrink the feedback loop, but let product judgment decide which models to deploy. That balance — speed with discernment — comes through repeatedly.

Habit architecture beats viral fantasies

Duolingo's disciplined playbook and chess.com's fanatical community are different operating systems that both win. Chang argues that habit architecture — tightly-crafted core loops, metagames, and profiles — creates durable engagement. Virality and brand moments can be rocket fuel, but they work best when the product already creates reasons to return. I liked his practical advice: look for organic screenshot and sharing moments and staff them with delightful creative work.

Hiring for velocity

Another counterintuitive idea: hire for agency over narrow experience. He prefers people who learn fast, move fast, and discard old habits when the ground shifts. That resonated. Experience helps, but agile energy often beats experience-heavy inertia — especially when AI and product paradigms evolve quickly.

A few tactical takeaways

• Instrument before you iterate: proper tracking prevents embarrassing false positives and enables meaningful scaling.
• Share winning patterns: a single experiment insight should be exported across adjacent product areas for multiplier effects.
• Resurrect dormant users: design tailored re-entry experiences rather than chasing new-user-only growth.

What really caught my attention was how pragmatic everything felt. Chang is not romantic about experimentation or AI; he treats both as tools to speed discovery and amplify product taste. He refuses to be dazzled by technology for technology's sake and consistently returns to the north star: connect users to value.

Final reflection

Growth isn't a sequence of one-off wins. It's a mindset: test urgently, exploit thoughtfully, and keep the product honest. The most durable companies mix method and imagination — and they build systems that let small discoveries become company-wide habits. That felt like the clearest lesson of all.