Growth Update
My user base has grown from 600+ to 800+ signups since my last update, with most new users coming from my Reddit post about the English tutor launch.
What I Accomplished
Planned goals
- Launch English tutor
- Launch new welcome chain
- Launch AI prompts handler and study resources
- Start mother tongue tutor (Chinese and Tamil)
- Beef up blog with English/GP guides (failed)
Unplanned wins
- Launch 5 character.ai type personas for students
I shipped the English tutor, revamped the welcome chain, and integrated Chinese (though not yet rolled out). I also created 5 character.ai type personas for students to interact with, which wasn’t in my original plan.
Overall grade: A-
Key Learnings from User Analysis
After reviewing over a thousand chat messages, I’ve identified several important patterns:
Subject-Specific Trends
Math
- Users get frustrated when the solver shows unclear steps or provides incorrect answers
- Students primarily use the math solver for practice papers and tutorials
English
- Fewer repeat users compared to Math (possibly due to less frequent practice needs)
- Users tend to ask fewer questions overall
Hangout Mode
- Polarized reactions: some love it, others find it “lame” to talk to AI
- Engaged users spend all free messages on a single conversation
User Behavior
- Users rarely switch between subjects, sticking to what they initially came for
- Most active users are junior college students
- Some students repurpose the Math tutor for Physics and Chemistry questions
- Students often seek urgent help (GP essay due tomorrow, math problems, etc.)
- Frustration rises when urgent needs aren’t met
- Mode confusion exists (e.g., students using hangout personas to solve academic questions)
Power User Profiles
- Struggling student retaking O levels
- JC students using AI to complete math tutorials
Remaining Questions
- Why don’t many users activate their account or send their first message?
- What prevents users from returning after initial use?
Self-Assessment
What I Did Well
I maintained rapid experimentation (English tutor, hangout personas, welcome chain). Shipping produces valuable data. My guiding question remains:
How can I ship this feature in 10x less time?
Mistakes I Made
I should have analyzed user behavior much earlier. I need to be an expert at understanding my users to build something they want.
Strategic Pivot
Adding more subjects isn’t enough. My reflections reveal deeper issues:
- Wrong answers create significant user frustration
- Users lack a clear improvement path and can’t track progress
- The system struggles to identify specific weaknesses
- Being just a “homework solver” isn’t a compelling value proposition
- Parents pay for grade improvement, not homework shortcuts
- JC students show highest demand and engagement
- Regular return visits require structured revision and practice systems
The Ideal Product Would:
- Deliver consistently correct answers
- Create daily return motivation
- Make learning enjoyable
- Track progress meaningfully
- Identify weak areas precisely
- Demonstrably improve grades
- Elicit enthusiastic approval from both students and parents
Next Steps
Rather than expanding breadth, I’m focusing on depth with a single subject tutor.
I’m leaning toward English/GP because:
- It doesn’t require a rigid syllabus
- It’s broadly applicable
- AI can provide genuinely valuable writing feedback
Before building, I need to:
- Validate if there’s demand
- Develop marketing materials to clarify my value proposition
- Validate if the product will make parents and students go “hell yeah”
- Set clear boundaries on features and scope
Target price point: SGD7 per week
This Week’s Goals
- Finalize product concept (subject, core features, deliberate limitations)
- Create compelling marketing materials
- Complete working prototype