Most local heritage stays locked behind guided tours and weekend workshops — hard to scale, impossible to automate. Zhengbin Art Studio had the stories, the place, and the community. They needed a product. We built Zhengbin Memory Game — a self-guided mobile puzzle game set inside a working fishing harbor. Full ownership: growth strategy, UX, and engineering. Five rounds of live playtesting, because the only metric that mattered was whether players would tell their friends. They did. NPS hit 53.
Completion Rate +50%
Playability Score 74
NPS 53
Challenge
A game nobody finishes doesn't grow. A game players don't recommend doesn't scale.
Two things blocked the path to a viable product:
GPS killed immersion. Triggers fired too early, failed on arrival, and repeatedly yanked players out of the story.
"Fun" had no single definition. Story lovers wanted atmosphere. Puzzle players wanted challenge. One game had to earn both — without losing its identity.
This was a product positioning problem. UX was just where it showed up.
How We Worked
We don't prototype assumptions. We test them in the field.
Five rounds of playtesting at Zhengbin Harbor. Four to six players per round. Behavioral observation, post-play interviews, playability scoring, and NPS — every time. Twenty-plus testers total. Each round told us exactly where players stalled, where they lit up, and where the story lost them.
Data drove every decision. Not instinct.
Design Strategy
Every round asked the same question: how much can we change before we lose what makes it fun?
We treated each test as a calibration — drop-off points, confusion moments, and delight signals told us exactly what to fix and what to leave alone.
Design Decisions
Replaced GPS with visual landmark cues
Photo prompts replaced coordinate triggers — the single biggest source of frustration. Navigation backup handled edge cases. Immersion stopped depending on satellite accuracy.
Paced to hold momentum
Audio narration, exploration, and mini-game challenges rotate in calibrated sequence. Completion climbed once difficulty spikes stopped killing momentum.
Made failure a non-event
Win or lose, the story moves forward. Casual players finish. Puzzle players stay engaged. Neither group quits.
A 500-screen system built to stay consistent
Unified interaction logic and state transitions across every interface — cutting engineering back-and-forth and keeping quality uniform at scale.
We turned local history into something people actually want to play. And once they do, they go tell someone else.
0 → 1 MVP