Fill-in-the-blanks
Solving the passive learning trap. Users felt stuck moving flashcards around. Two rounds of research invalidated a gamified concept before a single line of code, and the writing exercise that shipped instead moved engagement, revenue and churn.
- +4pppremium WAU/MAU, sustained
- +18%WAU/MAU peak with the campaign
- +7%revenue
- −18%monthly churn
The project
- Role
- Product Manager and Lead Researcher / Designer
- Timeline
- May–Aug 2024
- Company
- Early-stage B2C startup, ESL market
- Team
- Founder, 2 engineers, 1 junior designer, PM (me)
- Methods
- Opportunity Solution Tree, story mapping, assumption mapping, interviews, survey
Executive summary
I led the end-to-end discovery and delivery of a Writing / Fill-in-the-Blanks feature to address churn caused by a lack of perceived progress in a vocabulary-learning app. Through two structured rounds of user research, including invalidating an infinite XP concept before a single line of code was written, we shipped an MVP that delivered meaningful results within two months of launch: premium WAU/MAU +4pp (peaking at +18% during the campaign), revenue +7% and monthly churn −18%.
Absolute numbers reflect an early-stage company; the directional trends are the meaningful signal. RealLife English premium launched in August 2022.
PM signals
- Killed an idea before code. The infinite XP concept was invalidated in interviews with 3 segmented groups (power users, unsatisfied users, non-users), before any engineering time was spent.
- Pass/fail criteria set before research. Defined thresholds upfront, missed one (75% vs. 80%), and made an explicit, documented call to proceed instead of moving the goalposts.
- Structured discovery, not opinion. Opportunity Solution Tree, impact/feasibility prioritization, story mapping and leap-of-faith assumptions, with a hard gate: desirability first, then feasibility.
- Scope cut to ship faster. Traded a custom in-app keyboard for the native one; weeks of engineering saved with no impact on the core learning mechanic.
- Built for what comes next. Designed the mastery weighting so new exercise types plug in without refactoring, while respecting the existing SRS scheduling constraint.
- Measured across three layers. Engagement (WAU/MAU, practice completion), monetization (revenue, churn) and sentiment (N=733 survey), each with a named source.
Hands-on: product designer hybrid
I directed a junior designer on the first iteration. After the Daily Missions launch, I revamped the UI myself.